# Base Reality Group — full text Generated from the site HTML. 201 pages. --- # About — Dave Saunders | Base Reality Group Source: https://www.baserealitygroup.com/about About Dave Saunders # I've spent thirty-five years getting new technology out of the lab _and into the world._ Wi-Fi, surgical robotics, and a fair amount in between. The industries change. The problem of getting something complex out of the lab and into a customer's hands does not. How I got here ## From internet standards to Bell Labs to the OR. I started in tech in the late 1980s at the first company to build commercial internet software for Macs and PCs. From there I got pulled into the standards bodies, the IETF and the ITU, where I ended up as technical and marketing chair for a communications protocol I co-invented. At Ascend Communications I ran product management for TNT, the highest-density modem-based access concentrator on the market during the dial-up years of the dot-com boom. Then I took the 30 million lines of source code inside Ascend's access concentrator line, branded it as a monetizable operating system called TAOS, and grew that into roughly a $500 million per year business unit. When Lucent acquired Ascend for $24 billion, I moved into Bell Labs as a research manager, running a skunk works on next-generation telephony, internet access, and a fringe technology nobody had heard of yet: Ethernet over wireless. That work put my group in the room with Steve Jobs. He had just returned to Apple and was designing what would become the first personal computer with built-in wireless networking. We built the original consumer Wi-Fi hotspot. Apple shipped it as the AirPort. I kept moving. Cloud platforms. Early SaaS, including InSite, an ITIL service-management platform built on Salesforce at Bluehawk Networks. Forty-plus products across a wide range of industries, most of them software, before surgical robotics ever found me. When surgical robotics found me ## In 2011, surgical robotics found me. I joined a startup commercializing a handheld robotic system for knee replacement surgery. That was the beginning of a fifteen-year run in the hardest version of this problem. In 2016, Johns Hopkins called. They had technology that needed a commercial path. My co-founder and I licensed it, built Galen Robotics from the ground up, recruited most of the team, and set up operations, manufacturing, and engineering. In July 2023, we received FDA De Novo clearance for the world's first cooperatively controlled microsurgical robotic assistant. A 35-patent portfolio, licensed from Johns Hopkins and fully granted by clearance. $20 million Series A. Eight years of work. FDA De Novo Clearance · July 2023 $20M Series A 35 Patents Licensed from Johns Hopkins Since then ## The stage I know best. Since clearance I've embedded with four more surgical robotics teams and about a half-dozen other hard-tech founders. All at the same stage: promising prototype, unclear path forward. That stage is the one I know best, and it's where the decisions made in the next six months will either protect everything the team has built or quietly undermine it. Forty-plus products commercialized over thirty-five years, and the through-line is the same one: my interest has always been the little gears that bring technology together, and the work on the other side of the fence to get it into customers' hands. * * * ### Why this work Commercializing hard-tech is the hardest version of this problem. Regulated products like medical devices sit at the most demanding end of it: the consequences of getting it wrong are real, and the regulatory environment doesn't forgive a sloppy program. Most founders are building something genuinely important and navigating it largely alone. The same shape shows up at every scale. A SaaS product, a connected device, a five-person software shop: the discipline that gets a surgical robot through FDA is the same discipline, sized to fit, that gets any product into a customer's hands and keeps the business underneath it standing. That combination is what keeps me in this work: hard problems that matter, with people who care about getting it right. Companies I've worked with Galen Robotics $20M Series A raised  ·  FDA De Novo clearance granted July 2023  ·  35-patent portfolio, all granted by clearance AIM Medical Devices Trak Surgical Corosant Medical Agilis Robotics Pharyvac In partnership with ongoing research at **Johns Hopkins University Laboratory for Computational Sensory Robotics.** A few specifics ## The record, plainly stated. Career span 35+ years of product commercialization across software, telecom, cloud, and medical devices Products commercialized **40+** Bell Labs Research manager, post-Lucent acquisition of Ascend. Led a skunk works on next-generation telephony, internet access, and Ethernet over wireless. Apple AirPort My group at Bell Labs built the original consumer Wi-Fi hotspot. Apple shipped it as the AirPort, the first personal computer with built-in wireless networking. Ascend Communications Product line manager for **TNT**, the highest-density modem-based access concentrator on the market. Branded and productized **TAOS**, the operating system inside the Ascend access concentrator line: 30 million lines of source code into a **~$500M/year business unit**. Lucent acquired Ascend for **$24 billion**. Surgical robotics 15 years, starting 2011 Galen Robotics **Co-founder and CTO, 2016–2023** FDA pathway **De Novo clearance July 2023**: world's first cooperatively controlled microsurgical robotic assistant IP portfolio Managed a 35-patent portfolio licensed from Johns Hopkins, all granted by FDA clearance. Not an inventor; a commercialization architect. Funding (Galen) **$20M Series A** Post-clearance 4 additional surgical robotics engagements, approximately 6 other hard-tech founders Academic partnership Johns Hopkins University Laboratory for Computational Sensory Robotics Book Author of _Founders Who Finish_, available at [davesaunders.net/book](https://davesaunders.net/book) Newsletter _The Build_, monthly printed and mailed to subscribers, at [davesaunders.net/free-trial](https://davesaunders.net/free-trial) Work with Dave ## If this sounds like the kind of help your company needs, let's talk. There are two ways to engage, from a single paid hour to an embedded fractional role. Start wherever makes sense. [Work With Me](/work-with-me) [Apply to Work With Dave](/contact?interest=fractional) [My Values for Helping Good Clients Win](/values) --- # Books — Dave Saunders | Base Reality Group Source: https://www.baserealitygroup.com/books Books # Some problems are too expensive _to learn by trial and error._ This is especially true when the product is complex or the market is unforgiving, where a wrong turn at the wrong moment can cost a funding round, a submission, or the company itself. The books I've written come from the same place the work does. Thirty-five years of building products across software, telecom, cloud, and medical devices, including fifteen years in surgical robotics. The lessons inside come from real programs, some that worked and some that didn't, rather than from frameworks borrowed from other industries. Featured title ![Cover of Founders Who Finish, the book by Dave Saunders](/brand-assets/opt/fwf-book-cover-3d.webp) ## _Founders Who Finish_ Most founders who start a company don't finish. The ones who do share a particular set of habits the rest never develop, and this book is about what those habits actually look like in practice. It's written for founders who are already in it: a real product, investors who wrote checks, and the kind of pressure that comes with both. It's for the ones who want to understand what separates the companies that get their product to market from the ones that run out of runway before they get there. [Buy Now](https://davesaunders.net/book) Available at davesaunders.net/book More coming ## Additional titles in progress. I have additional books in progress. They go deeper on specific domains: the regulatory process, risk management, building a product program that survives diligence. When they're available, they'll be here. Coming soon Title to be announced Coming soon Title to be announced Bundle ## All titles at a discount. When additional titles are available, all books will be offered as a bundle at a discount. Check back when new titles launch. Bundle pricing Available when additional titles launch * * * ### Newsletter If you buy the book at davesaunders.net/book, you'll have the option to join my newsletter. That's where I share the shorter version of what's in these pages: specific scenarios, real decisions, the kind of thing that's hard to put in a book but easy to put in an email. --- # Contact / Apply — Dave Saunders | Base Reality Group Source: https://www.baserealitygroup.com/contact Contact / Apply # I work with a small number _of companies at a time._ If you think there's a fit, the fastest way to find out is to fill out the application below. I read every one. If it looks like the right match, I'll be in touch to set up a conversation. If you'd prefer to reach out directly: [dave@baserealitygroup.com](mailto:dave@baserealitygroup.com) Application ## Seven questions. Answer them plainly. A founder who can answer them clearly is ready to do this work. Prefer direct email? [dave@baserealitygroup.com](mailto:dave@baserealitygroup.com) Full Name Email Address 01 What company are you building and what does the product do? * * * 02 What stage are you at? Select stage Pre-seed Seed Series A Series B Other * * * 03 What does your current regulatory pathway look like (FDA or your industry's equivalent)? Select status Defined and documented In progress Unclear We haven't started Not applicable * * * 04 Where does your product documentation stand: risk management, requirements, V&V? * * * 05 What is your monthly burn rate? Select range Under $100K $100K–$300K $300K–$1M Over $1M Prefer not to say * * * 06 What are you trying to solve, and what does success look like in six months? * * * 07 How did you find me? Submit Application Thank you. I'll review your application and be in touch if there's a fit. --- # The Surgical Robotics Market Is No Longer a One-Platform Decision | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-04-20-surgical-robotics-competition-intensifies-2026 [← All Field Notes](/field-notes) April 20, 2026 Industry Roundup The competitive landscape in surgical robotics changed materially in the first quarter of 2026. Johnson & Johnson submitted a new robotic system to the FDA. CMR Surgical made its formal U.S. debut. Medtronic followed an earlier clearance with plans to expand into new procedure categories. For founders building surgical devices or adjacent enabling technology, the procurement environment that hospitals are operating in looks different today than it did eighteen months ago. ## J&J Submits Ottava for FDA De Novo Classification In January 2026, Johnson & Johnson MedTech submitted the Ottava robotic surgical system for FDA de novo classification, targeting soft tissue procedures in the upper abdomen, including gastric bypass, sleeve gastrectomy, small bowel resection, and hiatal hernia repair. The submission is supported by data from an investigational device exemption study, with a separate IDE approval received in late 2025 for inguinal hernia trials, according to reporting from MedTech Dive. The submission puts two of the world's largest medical device companies in active FDA review simultaneously. J&J and Medtronic are both competing for soft tissue robotic procedure volume that Intuitive Surgical has held almost entirely for the past decade. ## CMR Surgical Formally Enters the U.S. Market In late March 2026, CMR Surgical introduced its Versius robotic system to U.S. surgeons at the Society of American Gastrointestinal and Endoscopic Surgeons meeting in Tampa. CMR reports more than 45,000 patients worldwide have been treated on Versius robots across 30-plus countries. The company received 510(k) clearance in December 2025 for its next-generation Versius Plus system in gallbladder removal, as reported by MedTech Dive. CMR is positioning Versius as complementary to existing platforms rather than a direct replacement decision. CEO Massimiliano Colella has described the approach as fitting into what hospitals already have in place. For hospitals that have resisted robotics investment due to capital or commitment concerns, a modular, lower-disruption platform changes the evaluation calculus. ## Medtronic's Hugo Expands Beyond Urology Medtronic received FDA clearance for its Hugo robotic system in urology in December 2025 and is now pursuing expansion into general surgery and gynecology. The initial U.S. launch is concentrated at select leading institutions, with broader rollout planned as clinical evidence accumulates, according to MedTech Dive. Industry analysts expect meaningful but limited market impact through 2027. A robotic platform that justifies capital across three clinical departments is a materially different procurement conversation than a single-indication system. Medtronic's multi-specialty ambition changes the return-on-investment calculation for hospital procurement teams, and it changes the competitive surface area that all adjacent technologies have to navigate. ## Intuitive Adds Cardiac Indications, Projects Slower Growth Intuitive Surgical received FDA clearance for nine cardiac surgical indications, including mitral valve repair and left atrial appendage closure, and is beginning a limited U.S. rollout with specialized training programs. The company projects worldwide da Vinci procedure growth of 13 to 15 percent in 2026, below last year's approximately 18 percent growth rate, according to MedTech Dive. The slower growth projection reflects the base effect of a large installed base and the time it takes for new procedure indications to translate into volume. For founders building cardiac-adjacent enabling technology, the clearance signals where Intuitive sees its next growth frontier. ## New Entrants Target the Uncontested Hospital Segment A set of smaller developers including Distalmotion, Moon Surgical, and Virtual Incision are pursuing ambulatory surgery centers and hospitals that have never operated a robotic surgery program, according to MedTech Dive analysis. These institutions represent the market segment where Intuitive has the least penetration and where capital constraints are most acute. Two tiers are emerging in robotic surgery: large multinationals competing for established hospital accounts, and purpose-built platforms targeting greenfield institutions. The implications for how you position enabling technology, instruments, or software differ materially depending on which tier your device is built for. Dave's take What I watch most closely in this shift is not the competitive dynamics themselves but what they are doing to the hospital procurement process. Buyers who spent the last decade with a single option are now running genuine platform evaluations, and that changes the evidence requirements, the sales cycle, and the risk calculus for everything that touches the robotic workflow. If your product depends on hospital adoption of a specific platform, you need a clear thesis on which platform wins in your target accounts, and why. From Dave’s video library Dave on selling as clarifying the buyer's math instead of pitching, which is exactly what changes when hospitals go from one robotic platform to a genuine evaluation. When the buyer runs a real comparison, the sale turns on economics and evidence, not the demo. I'm here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to The Build newsletter: [The Build →](/field-notes/the-build) --- # CMR Surgical US Launch: When Execution Discipline Defines Market Entry | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-04-21-cmr-surgical-us-launch-execution-discipline [← All Field Notes](/field-notes) April 21, 2026 Industry Roundup CMR Surgical's CEO Massimiliano Colella made a decision that most investors would have resisted: in his first week on the job, he stopped the company's US launch rather than let it proceed with the original Versius system. His reason was direct. In a market where hospitals make decade-long technology commitments, there is no room for a disappointing first impression. The decision to wait for Versius Plus, and the principles behind it, are worth examining for any founder planning a market entry in a high-stakes regulated sector. ## CMR Surgical Delays US Entry to Launch the Right Product According to reporting from MedTech Dive, CMR Surgical CEO Massimiliano Colella made the decision to halt the planned US launch within days of taking the role, choosing to wait until the Versius Plus system was ready. He described the market environment as one where the company could not afford a single mistake. The Versius Plus system, which received FDA clearance in late 2025, brings new surgical instruments, enhanced visualization, improved data tracking, and greater system efficiencies over the original generation. Most of those improvements can be deployed to existing units through software updates. The delay was not a regulatory setback. It was a strategic choice to control the narrative of first contact with US hospital systems. CMR has treated more than 45,000 patients worldwide across more than 30 countries on Versius robots, and the company has raised nearly $1.5 billion in venture funding. It had the resources to wait. What Colella assessed was whether the original product was the right one to launch with in a market that would remember the first impression. ## Why the US Surgical Robotics Market Punishes Weak Launches The US hospital technology procurement cycle is structured in ways that make first impressions durable. Purchasing decisions involve clinical champions, OR leadership, supply chain, and finance. Once a hospital commits capital to a platform, the evaluation calculus shifts: the burden of proof moves to any competitor attempting to displace it. Relationships with Intuitive Surgical's training, service, and clinical support organizations run deep at institutions with installed da Vinci systems. CMR's positioning strategy leans into this reality rather than against it. Colella has described Versius as complementary to what hospitals already have, not as a displacement play at the top of the market. The modular, portable design allows hospitals to rotate a single unit across multiple operating rooms. For an institution that has resisted a multi-million-dollar fixed platform commitment, this is a materially different evaluation conversation. The MassDevice Surgical Robotics Special Report for 2026 identifies new sites of care, specifically ambulatory surgery centers and hospitals without any existing robotics program, as the biggest area of market differentiation this year. CMR, alongside Distalmotion, Moon Surgical, and Virtual Incision, is targeting these greenfield accounts rather than attempting to dislodge Intuitive at its core accounts. ## GE HealthCare and Medtronic Integrate Intraoperative Imaging with Navigation GE HealthCare and Medtronic announced integration of the bkActiv intraoperative ultrasound system with the Stealth AXiS surgical navigation and robotic platform, according to MassDevice reporting. The integration connects real-time imaging with surgical navigation for procedures performed on the Stealth AXiS system, which received back-to-back FDA clearances in early 2026 for spine, cranial, and ENT procedures. For founders building imaging, navigation, or data infrastructure that touches robotic workflows, partnerships of this kind signal where the integration surface area is expanding. Large platform developers are actively seeking imaging and sensing capabilities to differentiate their systems. That creates an acquisition and partnership environment that rewards companies with demonstrated clinical evidence and production-ready integrations. ## Two Tiers of Competition Are Solidifying The surgical robotics competitive landscape in 2026 has effectively split into two tiers. Johnson & Johnson and Medtronic are pursuing soft tissue procedure volume at established hospital accounts, going directly after Intuitive's installed base with comparable large-platform systems. CMR, Distalmotion, Moon Surgical, and Virtual Incision are building access at institutions and care settings Intuitive has not reached. These tiers have different implications for founders building adjacent technology. An enabling device or software platform built for the top-tier hospital accounts needs to navigate a procurement process that involves major capital commitments and entrenched relationships. A product built for ASCs and greenfield hospitals is operating in a lower-inertia environment where first-mover positioning and clinical evidence can move faster. Understanding which tier your target customers operate in is part of the strategy. Dave’s take The CMR story is really a story about what it costs to be wrong at launch in a market that does not give you a second first impression. Colella walked in and stopped a launch that the organization had momentum behind. That is a hard decision to make under investor pressure and org-wide expectation. The fact that he made it tells you something about how seriously the company is taking long-term account relationships over short-term revenue timing. I see founders face this same pressure constantly: ship now and fix it later, or hold until the product is actually ready. In regulated markets with long sales cycles, the cost of shipping early is not just one bad customer. It is a reference that follows you into the next ten conversations. From Dave’s video library Dave on why the build clock is not the real clock, and why the only proof that counts is external. CMR stopping its own US launch to wait for a better product is the same discipline: ready is negotiable forever, proof is not. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to The Build newsletter: [The Build →](/field-notes/the-build) --- # Robotic Surgery Expands Into Endovascular Practice as Microbot Launches Liberty at SIR | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-04-22-endovascular-robotics-commercial-launch-sir-2026 [← All Field Notes](/field-notes) April 22, 2026 Industry Roundup Microbot Medical is making the commercial debut of its Liberty robot at the Society of Interventional Radiology annual conference this week, marking the first FDA-cleared single-use, remotely operated system for peripheral endovascular procedures to enter clinical practice in the United States. The launch lands during a period of broad expansion across surgical robotics: Medtronic’s Hugo is building its first US commercial footprint, J&J’s Ottava is in FDA review, and Intuitive Surgical has cleared nine new cardiac surgery indications. For founders building surgical devices or adjacent enabling technology, the market is moving across multiple clinical specialties at once. ## Microbot Medical Brings the First Endovascular Robot Into Clinical Practice Microbot Medical received FDA clearance for the Liberty system in September 2025, making it the first single-use, remotely operated robot cleared for peripheral endovascular procedures in the United States. The system targets procedures below the aortic bifurcation, a segment of vascular surgery where the skills gap between procedure demand and available trained operators has become a genuine bottleneck at hospital systems around the country. The commercial launch at the Society of Interventional Radiology conference in April 2026 represents the company’s full market entry after a controlled pre-commercial phase. To support the rollout, Microbot exercised investment options with a potential value of up to $92.2 million, according to reporting from MassDevice. That capital is going toward the commercial infrastructure needed to support hospital accounts beyond early adopters, including field support and training programs. For founders building in the interventional radiology adjacency, the Liberty launch matters beyond Microbot’s own market position. It establishes that the endovascular robotics category is moving from development to commercial practice, which changes the landscape for integration partners, imaging system developers, and capital equipment vendors working in the same clinical environment. The first cleared system in a category is always the procurement signal that hospital accounts were waiting for before engaging seriously. ## Medtronic’s Hugo Builds Early Commercial Footing at US Academic Centers Medtronic completed its first US procedure with the Hugo robotic-assisted surgery system in early February 2026, with a prostatectomy performed at Cleveland Clinic, according to MedTech Dive. Early commercial sites include Duke University Hospital and Atrium Health Wake Forest Baptist High Point Medical Center, with Atrium Health being the first non-study institution to install the system. Hugo received FDA clearance in December 2025 for urologic procedures after completing the largest US study for robotic-assisted urological surgery, with a 98.5 percent surgical success rate against an 85 percent performance goal. Gynecologic and general surgery indications are planned as next expansions. The company’s rollout strategy is deliberately focused on leading academic and health system institutions rather than broad hospital adoption, building reference accounts and trained surgeon communities before moving into the broader market. The conservative early-stage approach reflects what competing against Intuitive’s installed base actually requires. Credible references at recognizable institutions change the procurement conversation at every subsequent account. Companies that skip this step tend to find the sales cycle longer and harder than it needed to be. ## J&J’s Ottava Is in FDA Review and the Market Has Not Priced In a Third Platform Yet Johnson & Johnson submitted its Ottava robotic surgery system for FDA de novo classification in January 2026, targeting general surgery applications. A clearance decision would introduce a third credible large-platform competitor into a market that has operated as a one-company environment for soft tissue surgery for more than two decades. Hospital procurement teams are already beginning to hold conversations that were not possible two years ago. Accounts that deferred robotics investment while waiting for competitive alternatives now have the prospect of evaluating three platforms with different clinical footprints, pricing structures, and ecosystem strategies. For founders building surgical instruments, imaging systems, or workflow software that runs alongside these platforms, the platform selection question now has real implications for integration roadmaps and partnership sequencing. ## Intuitive Surgical Extends Da Vinci Into Cardiac Surgery Intuitive Surgical added nine new FDA-cleared indications for cardiac surgery in early 2026, including mitral and tricuspid valve repair, according to MedTech Dive’s analysis of 2026 robotic surgery trends. The company is projecting worldwide da Vinci procedure growth of 13 to 15 percent for the year, down from 18 percent in the prior period but still representing meaningful expansion across a large installed base. The cardiac expansion is significant because it extends robotic assistance into procedures where open surgery has remained the standard despite its complexity and recovery burden. The mitral valve repair indication addresses a procedure with a known skills concentration problem in the cardiothoracic surgery community, the same dynamic that drove robotic adoption in urology and then gynecology. For founders building in cardiac adjacencies, the cleared indication set creates a reference pathway for clinical trial design and regulatory submissions targeting the same procedure category. Dave’s take What I keep coming back to is how much is happening at once. Soft tissue, spine, endovascular, cardiac, ophthalmology, ENT — the market is not expanding sequentially. It is expanding in parallel, and each specialty has its own procurement dynamics, evidence requirements, and clinical champion profile. The founders I work with who are building enabling devices or workflow software for robotic procedures need to be very specific about which clinical environment they are actually building for. That answer changes what evidence you need, who makes the purchase decision, and what timeline you are working against. Building for a category is not a strategy. Building for a specific clinical problem in a specific care setting with a specific evidence plan is. From Dave’s video library Dave walks through the four-question filter for cutting a pile of ideas down to the one worth building. A market expanding into six specialties at once is exactly when a founder needs that filter most. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to The Build newsletter: [The Build →](/field-notes/the-build) --- # AI-Integrated Surgical Navigation Reaches Neurosurgery as Medtronic Clears Stealth AXiS for Cranial and ENT Procedures | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-04-23-ai-surgical-navigation-neurosurgery-cranial-ent [← All Field Notes](/field-notes) April 23, 2026 Industry Roundup Medtronic received FDA 510(k) clearances for its Stealth AXiS surgical system for cranial and ENT procedures in March 2026, following an earlier clearance for spine procedures the same month, according to MedTech Dive. The system combines AI-enabled surgical planning, real-time navigation, and robotic assistance, and integrates with GE HealthCare’s bkActiv for intraoperative ultrasound imaging. For founders building in neurosurgical adjacencies, imaging, or AI-assisted planning tools, this is a credible commercial entry point in a specialty that has operated largely without robotic assistance at this level of integration. ## Medtronic Stealth AXiS Brings AI Navigation Into the Brain and the Sinuses The Stealth AXiS system received 510(k) clearance for ENT procedures on March 16, 2026 and for cranial applications on March 26, 2026, per MedTech Dive. For cranial surgery, the system generates patient-specific brain maps using AI-enabled tractography, providing surgeons with a three-dimensional model of neural architecture before the procedure begins. For ENT, it delivers navigation and visualization adapted to sinus and skull-base anatomy, where surgical margins between functional tissue and the operative target are measured in millimeters. The system also integrates with GE HealthCare’s bkActiv ultrasound platform, adding real-time intraoperative imaging to a workflow that traditionally relies on preoperative imaging alone. Medtronic has positioned this platform within a market it estimates at $15 billion for cranial and spinal technologies, per MedTech Dive. This matters for founders working in neurosurgical devices or imaging software for two distinct reasons. First, the cleared system establishes what the AI-assisted planning workflow looks like in commercial practice, which becomes the integration target for any companion device or software layer. Second, the clinical champion who is now using AI tractography for brain tumor surgery has a working reference point for what AI-assisted navigation can do. That changes what they will evaluate when a startup comes in with a new device or tool. ## Ambulatory Surgery Centers Are the Next Competitive Frontier for Surgical Robotics While large-platform robotic systems pursue academic medical center footholds, a distinct tier of smaller robotic platforms is building specifically for ambulatory surgery centers where Intuitive Surgical has limited presence, according to MedTech Dive’s analysis of 2026 competitive dynamics. Virtual Incision’s MIRA robot is positioning a miniaturized form factor as the differentiator for outpatient soft tissue procedures. Moon Surgical’s Maestro system, backed by J&J venture capital and a partnership with Nvidia, is targeting the same ASC soft tissue segment. CMR Surgical’s Versius Plus is explicitly targeting both hospital and ASC accounts as part of its 2026 US commercial strategy. For founders building capital equipment, disposables, or workflow software, the ASC market operates on different procurement dynamics than academic medical centers. Purchase decisions move faster, involve fewer stakeholders, and clear at lower total capital thresholds. The evidence requirements are real, but the cycle from credible clinical data to signed first account is shorter. The challenge is that the first cleared system in a procedure category at an ASC sets the standard for every platform that follows in that setting. ## International Capital Continues to Flow Into Surgical Robotics Beijing-based Surgerii Robotics closed a $100 million Series D financing round led by Loyal Valley Capital in 2026, according to MedTech Dive. The company is deploying the capital toward global commercialization and next-generation platform development. Surgerii is part of a broader pattern of international surgical robotics developers scaling capital alongside their geographic reach. Shanghai-based MicroPort MedBot and Japan’s Medicaroid are both expanding global distribution, per MedTech Dive’s coverage of the 2026 competitive landscape. The practical implication for US-based founders is that the hospital systems they are approaching are not choosing between domestic platforms. They are evaluating a global field. When a procurement committee at a regional medical center considers a robotic surgery program, the set of options they are reviewing in 2026 is substantially larger than it was three years ago, and international competitors are arriving with clinical data from markets where they have been operating for years. ## Cardiac Valve Robotics Moves Toward a TAVR Alternative Corcym and CardioPrecision have entered a joint development collaboration targeting robotic aortic valve replacement, positioning their combined approach as a less-invasive alternative that preserves the durability advantages of open surgical repair, per MedTech Dive. The collaboration enters a structural heart market where Intuitive Surgical also received FDA clearances in early 2026 for mitral and tricuspid valve repair. For founders building in structural heart adjacencies, the opening of robotic-assisted valve procedures creates an important question about evidence path. Cardiac surgery and interventional cardiology have distinct clinical champion profiles, reimbursement structures, and evidence requirements. The device you are building determines which clinical community you need to earn, and that community’s expectations are set by the cleared platforms already operating in their procedure space. Dave’s take The Stealth AXiS clearances are worth paying close attention to beyond Medtronic’s own market position. AI-assisted surgical planning is becoming a procurement expectation in specialties that previously ran on static preoperative imaging alone. If you are building a neurosurgical device or a sinus procedure tool, your clinical champion now has a cleared AI navigation system as a reference point, and that changes the baseline you are designing against. The second thing worth noting is the ASC competition story. The platforms chasing ambulatory surgery centers are not doing it because hospitals are saturated. They are doing it because the procurement dynamics are faster and the competitive position against Intuitive is more defensible. If you are building anything that lives in that care setting, those dynamics matter for your sales cycle and your funding story. From Dave’s video library Dave on what AI actually does inside a working business and where it breaks, including the FDA's own list of more than a thousand cleared machine-learning devices. Useful context now that AI navigation is becoming a procurement expectation in the OR. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to The Build newsletter: [The Build →](/field-notes/the-build) --- # Intuitive Surgical Raises 2026 Outlook as Robotic Surgery Expands Into General Surgery and Gynecology | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-04-24-robotic-surgery-procedure-growth-q1-2026-market-expansion [← All Field Notes](/field-notes) April 24, 2026 Industry Roundup Intuitive Surgical reported 16% worldwide da Vinci procedure growth in the first quarter of 2026 and lifted its full-year guidance to 13.5–15.5%, driven by expansion in general surgery and a 31% rise in after-hours procedures in the U.S., according to MedTech Dive. The quarter came with revenue of $2.77 billion and shares climbed nearly 8% on the news. For founders building surgical devices or adjacent enabling technology, the Q1 numbers carry a more specific signal than general market momentum. ## What the Q1 Procedure Numbers Tell Founders About Where the Market Is Going Intuitive’s Q1 breakdown shows 14% U.S. procedure growth and 19% international, per MedTech Dive. The U.S. driver was general surgery and gynecology, not urology, combined with a 31% rise in after-hours utilization as hospitals worked to extract more throughput from their existing da Vinci installations. International growth was partly offset by competitive pricing pressure in China and persistently below-historical performance in Japan. The specialty expansion matters for founders because it changes who the clinical champion is at the accounts now converting to robotics. Urology established the da Vinci’s initial commercial footprint in the U.S. General surgery and gynecology champions carry different evidence requirements, different procedural cadence, and different relationships with hospital administration. The accounts coming online in 2026 are not the same programs that adopted early. Founders building for any of these specialties are dealing with a different procurement environment than the one that existed three years ago, and that changes what your first reference site needs to look like and what evidence it takes to move the next ten. The raised guidance also implies Intuitive expects this conversion rate to continue. That creates a specific window for founders building complementary devices and tools at exactly the moment when the installed base is expanding fastest. Whether that window is an advantage for you depends on whether your evidence plan is calibrated to this cohort of new accounts rather than to the early-adopter programs that have been running for a decade. ## Medtronic Closes $585 Million Acquisition of CathWorks Medtronic closed its $585 million acquisition of CathWorks on April 20, 2026, per MedTech Dive. CathWorks makes the FFRangio system, which uses AI and computational modeling to derive fractional flow reserve values from standard coronary angiograms, eliminating the need for a pressure wire or adenosine administration. Medtronic had been co-promoting the system since 2022 before the deal, announced in February, closed this month. The deal confirms a pattern worth noting for founders building AI-enabled diagnostic tools in interventional specialties. By the time Medtronic closed, CathWorks was not a startup with a compelling technology thesis. It was a commercial asset with an established reimbursement position, peer-reviewed clinical data, and an integrated workflow position at cath labs in multiple markets. The acquisition premium was paid for that finish, not for the potential. For founders in AI diagnostics categories, the CathWorks outcome is a more instructive data point about what acquirers are actually paying for than any earlier-stage deal. ## Stereotaxis Announces Acquisition of Robocath for Endovascular Robotics Stereotaxis announced in April 2026 that it would acquire Robocath for up to $45 million, per MedTech Dive. The deal combines Stereotaxis’s magnetic robotic navigation technology with Robocath’s mechanically actuated robotic catheter platform for endovascular surgery. CEO David Fischel described the combination as uniting two complementary navigation modalities. Stereotaxis had received FDA approval for its Magic ablation catheter in January 2026, entering the disposables market for its robotic EP systems for the first time. For founders building in endovascular robotics or catheter navigation, the acquisition reflects consolidation of technical approaches in a market that has developed largely in parallel across magnetic and mechanical modalities. It also signals that Stereotaxis is moving into a more aggressive commercial posture as it builds out a complete platform rather than a navigation-only story. ## Medtronic Leads $100 Million Investment in Pulnovo Medical Medtronic led a $100 million investment in Pulnovo Medical on April 21, 2026, alongside a commercial agreement, per MedTech Dive. Pulnovo is developing pulsed field ablation technology for cardiac arrhythmia treatment, entering a market where Boston Scientific’s Farapulse system has established the current cleared standard. The investment follows Medtronic’s CathWorks close and reflects continued portfolio expansion in cardiac electrophysiology and structural heart. The PFA market has moved from a single cleared system to a multi-vendor competitive environment faster than most procedural device categories do. For founders building tools adjacent to EP labs, including mapping, imaging, or workflow software, the strategic relationships available in that setting are changing, and the companies that established early positions with the Farapulse accounts may find those relationships revalued as new platforms arrive. Dave’s take The Intuitive Q1 numbers confirm that the market is expanding, and the new accounts coming online in general surgery and gynecology are not the same programs that ran first-generation robotic procedures. Your first reference site in 2026 is a different conversation than it would have been three years ago, because the baseline of what cleared robotic surgery looks like in those specialties has changed. The CathWorks deal is the one I would spend the most time with if I were building an AI diagnostic tool in interventional cardiology or any adjacent field. Medtronic did not buy potential. It bought a finished commercial asset with a reimbursement position. That is the actual finish line to design toward. From Dave’s video library Dave on getting your first customers when nobody knows you exist. New robotic programs coming online in general surgery and gynecology are new first-reference conversations, and winning them is the same ground game. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to The Build newsletter: [The Build →](/field-notes/the-build) --- # SS Innovations Showcases Next-Generation Surgical Robotics Platforms at SMRSC 2026 | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-04-25-next-gen-surgical-robotics-platforms-smrsc-2026 [← All Field Notes](/field-notes) April 25, 2026 Industry Roundup The Global Multi-Specialty Robotic Surgery Conference in New Delhi this week drew more than 1,600 attendees from 19 nations and became the platform for SS Innovations to unveil a range of next-generation concepts: drone-deployed battlefield surgery, a zero-footprint mobile operating room, and an AI-integrated humanoid surgical robot still in early conceptual development. Separately, AcuityMD closed an $80 million Series C at a $955 million valuation, and the European Medicines Agency launched a new Breakthrough Device pilot that mirrors the FDA designation pathway for the first time. ## What SS Innovations' SMRSC 2026 Showcase Tells Founders About the Platform Roadmap SS Innovations used the Global Multi-Specialty Robotic Surgery Conference in New Delhi to present four distinct platform concepts alongside its commercially operating SSi Mantra system, which ran ten live telesurgeries during the event, according to a press release from SS Innovations. The concepts ranged from the SSi Vimana, a miniature drone-mounted robotic system designed for battlefield trauma surgery, to the SSi Operion, a fully mobile operating room with overhead-integrated robotics designed for deployment in hospitals, remote locations, and military settings. The SSi Avtara humanoid surgical robot is the furthest from market, described by the company as a conceptual platform for AI-integrated mobility and real-time sensing across healthcare and defense environments. The single-arm robotic assist carts for endoscopy and ultrasound positioning are the most grounded in near-term clinical reality, currently in early validation in India. What matters for founders is not the distance between any of these concepts and commercial reality. It is what the range tells you about where the form factor question is heading. The surgical robotics market has operated for most of its commercial history on the assumption that the reference architecture is a fixed, multi-arm system in a hospital OR. SMRSC 2026 is one more data point in a growing body of evidence that this assumption is unwinding. Mobile ORs, single-use endovascular robots, ASC-focused platforms, and drone-deployed systems are all being developed concurrently, and the hospital OR is no longer the only integration environment that matters. Founders building enabling technology, instruments, or workflow software who have architected their integration around a single platform in a single setting are carrying more dependency risk than the market as it currently stands actually requires. ## AcuityMD Raises $80 Million for AI-Powered Commercial Intelligence in MedTech AcuityMD closed an $80 million Series C led by StepStone Group, with participation from Benchmark, Redpoint Ventures, ICONIQ, and Atreides Management, according to a company announcement on April 21, 2026, via Business Wire. The round brings total funding past $160 million and values the company at $955 million. AcuityMD consolidates fragmented claims data, FDA filings, and government records into a unified knowledge graph that commercial teams at MedTech companies use to identify market opportunities and target accounts. The platform now serves 16 of the top 20 MedTech companies and has identified more than $34 billion in pipeline opportunity for its customers. The $955 million valuation on $160 million total raised is a signal worth examining. The data problem in MedTech commercial operations is real and persistent. Clinical outcomes data, account-level procedure volume, physician specialization, and FDA clearance histories exist in fragmented, hard-to-query formats across dozens of different government and commercial databases. AcuityMD has spent years building the infrastructure to stitch those sources together into something a commercial team can actually use. The round suggests that solving that data problem at scale has real strategic value to the large companies that would otherwise be building it themselves. For early-stage founders, the relevant implication is less about competitive threat and more about baseline expectations. The large companies your sales team will be negotiating against at the account level are using AI-powered commercial intelligence to understand your target accounts as well as or better than you do. Your commercial readiness plan needs to account for that asymmetry. ## EMA Launches Breakthrough Device Pilot as EU Regulatory Pathways Evolve The European Medicines Agency held an information session on April 24, 2026, announcing a new pilot program for Breakthrough Medical Devices that mirrors the FDA's Breakthrough Device Designation, according to Healthcare Digital. The program will offer manufacturers of highly innovative devices priority scientific advice and a streamlined regulatory route. Categories named specifically include neuro-implants and robotic platforms. The FDA Breakthrough pathway has measurably compressed timelines for qualifying devices in the US. If the EMA pilot develops with comparable effect, it changes the sequencing calculation for founders who have historically treated CE marking as the slower but more predictable path. The detail to watch is what the EMA defines as qualifying criteria for the designation and whether the scientific advice process translates into the kind of pre-submission engagement that has made the FDA Breakthrough pathway valuable in practice. The pilot announcement tells you the intent; the outcomes over the next twelve to eighteen months will tell you the execution. ## J&J MedTech Launches Next-Generation Pulsed Field Ablation System in Europe Johnson & Johnson announced the European launch of its VARIPULSE Pro3 pulsed field ablation platform at the Heart Rhythm Society meeting in late April 2026, per Healthcare Digital. The system is designed for atrial fibrillation treatment and offers claimed workflow improvements over existing thermal ablation methods. The launch follows Medtronic’s $100 million investment in Pulnovo Medical’s competing PFA technology announced on April 21, 2026. The PFA market has now crossed into multi-platform competitive territory faster than most procedural device categories do. For founders building tools adjacent to EP labs, the practical consequence is that the account relationships available in the cardiac electrophysiology setting are becoming more complex. Early-adopter accounts that built their program around the first cleared PFA system are now being approached by multiple platform vendors. The ecosystem relationships those accounts formed with early-stage companies have real value and real fragility at the moment when the platform choices are actively being re-evaluated. Dave’s take The SMRSC conference is worth paying attention to because it shows what the platform roadmap looks like several generations out. The companies showing drone-deployed surgery and mobile ORs alongside their commercial systems are telling you something about where they expect the care delivery environment to go, and that changes which integration points are stable across form factors and which ones are specific to the current hospital OR architecture. The AcuityMD round is a separate signal I would not ignore: when a platform serving the majority of the top MedTech companies raises near a $1 billion valuation, the commercial data problem has graduated from niche tool to competitive infrastructure, and founders building commercial plans without that layer are at a structural disadvantage. From Dave’s video library Dave on Barnum's playbook and how to tell showmanship from substance, with intent as the divider. Drone surgery and humanoid robots on a conference stage are exactly the kind of claims that test was built for. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to The Build newsletter: [The Build →](/field-notes/the-build) --- # Boston Scientific Cuts 2026 Guidance as Watchman and EP Categories Restructure | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-04-26-boston-scientific-2026-guidance-cut-watchman-ep-warning [← All Field Notes](/field-notes) April 26, 2026 Industry Roundup Boston Scientific posted Q1 2026 sales of $5.2 billion and adjusted earnings near the top of its range, then cut full-year sales growth guidance from 10.5%–11.5% down to 7%–8.5% and trimmed adjusted EPS, citing weakness in three named businesses: Watchman, electrophysiology, and urology. The same week, Johnson & Johnson opened a head-to-head pulsed field ablation trial of Varipulse Pro against Boston Scientific’s Farapulse, Abbott released late-breaking PFA and conduction system pacing data at HRS 2026, and the Heart Rhythm Society published its first joint scientific statement on PFA with the European Heart Rhythm Association. The week is a concentrated picture of what happens to a category when a near-monopoly leader meets credible commercial competition. ## Why Boston Scientific’s Watchman and EP Warning Resets the 2026 Outlook Boston Scientific reported Q1 2026 revenue of $5.2 billion, an 11.6% reported gain, and net income of roughly $1.34 billion, according to MedTech Dive. CEO Mike Mahoney told investors the company was not proud of the guide-down but believed it was the right thing to do. The numbers behind that decision were that Watchman delivered $506 million in the quarter at 19.2% growth, but standalone Watchman procedures began declining in mid-February as practice patterns shifted, hospital capacity tightened, and EP physicians took on more of the procedure mix. Mid-teens growth is now the planning expectation for the franchise in 2026. Electrophysiology grew 24% to $905 million in Q1, but the 2026 outlook for EP was reduced to roughly 10% growth as Boston Scientific acknowledged competitive share losses to Medtronic, Johnson & Johnson, and Abbott. Urology grew only 2%, with the company citing China value-based procurement, gaps in the core stone portfolio, and continued sacral neuromodulation disruption. For founders, the relevant pattern is not specific to Boston Scientific. It is the speed at which a category-defining product loses pricing and volume durability once two or three credible commercial competitors clear regulatory and start publishing comparable outcomes. Watchman and Farapulse are both products that built a category. Both are now meeting the structural pressure that arrives with a multi-platform competitive set, and both are doing so on a timeline that internal planning models almost always underestimate. ## J&J’s PERSIGMA Trial Pits Varipulse Pro Directly Against Farapulse At HRS 2026 in Chicago, Johnson & Johnson’s Devi Nair featured one of the first cases from the PERSIGMA trial, a head-to-head comparison of the company’s Varipulse Pro PFA platform against Boston Scientific’s Farapulse, according to MassDevice. Direct head-to-head comparative trials between commercial competitors in EP have been rare. PERSIGMA signals that J&J is willing to commit to generating comparative data rather than relying on workflow differentiation. For founders building in any device category where one company holds the majority share, the strategic implication is concrete. The playbook for taking share has moved from positioning around clinical workflow advantages to publishing the comparative evidence that makes the customer’s choice quantitative. If your competitive plan does not have a head-to-head evidence pathway in it, the absence is becoming visible faster. ## Abbott’s HRS 2026 Late-Breaking Data Confirms a Four-Way PFA Race Abbott presented four late-breaking presentations at HRS 2026, including six-month results from the FlexPulse IDE study of the TactiFlex Duo ablation catheter, the Volt CE Mark Extension Cohort data on posterior wall ablation, and first-in-human evaluations of the AVEIR conduction system pacing leadless pacemaker. The FlexPulse data reported 87% freedom from documented arrhythmias, a 98.3% safety profile with no major safety events, and 93.9% of patients not requiring a second ablation, according to Abbott’s announcement. With Boston Scientific’s Farapulse, Medtronic’s PulseSelect, J&J’s Varipulse Pro, and Abbott’s Volt all now commercial or near-commercial, the PFA market has graduated from a one-leader category to a four-way platform race. Founders building enabling technology, mapping software, or workflow tools designed around a single platform partnership are now carrying a higher dependency risk than the same architecture carried twelve months ago. ## HRS Publishes First Joint Scientific Statement on Pulsed Field Ablation The Heart Rhythm Society and the European Heart Rhythm Association jointly released the 2026 HRS/EHRA Scientific Statement on Pulsed Field Ablation for Cardiac Arrhythmias. Scientific statements from these societies are how professional medicine signals that a technology has graduated from emerging innovation to clinical standard. The statement carries two implications for founders. The first is that the PFA category now has formal clinical legitimacy across both regulatory regions, which compresses adoption timelines for credible entrants. The second is that the same legitimacy invites every credible competitor with an approved system into accounts that previously made a single-platform bet. The window for category-defining incumbency in PFA is closing on the schedule the joint statement just put on the calendar. Dave’s take The Boston Scientific guide-down is the kind of signal I tell every founder in a regulated category to study carefully. You can build a category and still lose volume and pricing durability faster than your operating plan assumes once two or three credible competitors clear regulatory and start generating comparable outcomes data. The question to ask about your own business is not whether you are the leader today, but whether the structural advantages you are counting on, regulatory position, evidence base, account relationships, training infrastructure, will compound or erode when the second and third commercial competitors arrive. If you do not know which of those advantages travel and which do not, you do not yet have a defensible position. You have a head start. From Dave’s video library Dave on pricing from the value you deliver and why underpricing shows. Boston Scientific just learned how fast pricing durability erodes once credible competitors arrive with comparable data, and that pressure reaches every founder eventually. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to The Build newsletter: [The Build →](/field-notes/the-build) --- # J&J Wins CE Mark for Ottava-Compatible Ethicon Stapler as Surgical Robotics Becomes a Platform Moat | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-04-27-jj-ottava-ethicon-stapler-ce-mark-platform-moat [← All Field Notes](/field-notes) April 27, 2026 Industry Roundup Johnson & Johnson received a CE mark on April 24 for the Ethicon 4000 surgical stapler, an instrument designed to run on the Ottava robotic surgery system in a market where Ottava itself does not yet have European authorization. The same week, Edwards Lifesciences raised its 2026 sales and TAVR guidance after a third consecutive quarter of double-digit transcatheter aortic valve replacement growth, contrasting cleanly with Boston Scientific’s guide-down two days earlier. Conavi Medical and Protaryx Medical each cleared FDA 510(k) milestones in interventional imaging and left-heart access. Read together, the week is a clean lesson in where the durable moat in surgical robotics and structural heart now lives. It is not the platform. It is the ecosystem and evidence the platform travels with. ## J&J Cleared an Ottava-Compatible Stapler in Europe Before the Robot Johnson & Johnson announced on April 24 that the Ethicon 4000 surgical stapler had received a CE mark and was available immediately in Europe for open and laparoscopic procedures, according to MedTech Dive. The instrument is designed to integrate with the Ottava robotic surgery system. Ottava itself does not yet have European authorization. Reading that timing carefully is the point. J&J’s strategic position has always been that surgeon familiarity with Ethicon instruments is a differentiator the company can carry into the robot market currently dominated by Intuitive Surgical. The Ethicon 4000 is the visible execution of that thesis. By landing the stapler in open and laparoscopic surgery first, J&J creates a base of surgeons who are already trained on the instrument, already comfortable with its 3D stapling tissue compression behavior, and already running it inside European hospital procurement pathways. When Ottava clears CE Mark, the instrument that runs on it is not a new ask. It is an instrument the surgeon has been using for months and now has a robotic version of. For founders building surgical robotic systems or enabling technology around them, the implication is structural. The competitive moat in surgical robotics has stopped being the robot itself. Intuitive built two decades of platform leadership on a moat that combined the platform, the instrument ecosystem, the surgeon training base, and the published evidence at every stage. Medtronic’s Hugo and J&J’s Ottava are now both pursuing that same multi-element moat in parallel. If your business plan assumes the differentiation is the platform mechanics or the surgeon experience, it is competing on the wrong axis. The differentiation in 2026 is everything that travels with the platform: the instrument set, the training infrastructure, the integration with hospital workflow, and the evidence base that makes adoption defensible to a procurement committee. ## Edwards Raised 2026 Guidance as TAVR Outpaced Boston Scientific’s Cut Edwards Lifesciences reported Q1 2026 sales of $1.65 billion on April 24, a 16.7% year-over-year gain, with TAVR sales of $1.2 billion at 14.4% growth, according to MedTech Dive. The company raised full-year sales growth guidance from 8%–10% to 9%–11% and lifted TAVR growth guidance from 6%–8% to 7%–9%. CEO Bernard Zovighian attributed the trajectory to a shift toward proactive evaluation of severe aortic stenosis patients. It was the third consecutive quarter of double-digit TAVR growth. The week-over-week contrast is the lesson. Boston Scientific cut 2026 guidance citing weakness in Watchman, electrophysiology, and urology two days earlier. Edwards raised guidance in the same window, in a structural heart category where competitive pressure was supposed to be intensifying given Medtronic’s Evolut platform exposure. The difference is not market sentiment. It is that Edwards has continued generating clinical evidence in aortic stenosis populations and pushing earlier intervention, while Medtronic’s Evolut faced unfavorable reintervention data that softened its competitive position. For founders building in any category where the leader and the challenger are both running comparable platforms, the thesis Edwards is validating is that the durable advantage comes from continuing to expand the indication, not from defending the current one. ## Conavi Medical Cleared a Combined IVUS-OCT System for Coronary Imaging Conavi Medical announced on April 20 that the FDA had granted 510(k) clearance for its next-generation hybrid imaging system, which combines intravascular ultrasound and optical coherence tomography in a single platform, per MassDevice. The integrated system lets interventional cardiologists visualize both deep vessel structures with IVUS and high-resolution surface detail with OCT in real time during the same procedure. Conavi expects to initiate a limited market release in select U.S. centers in the second half of calendar 2026. The convergence pattern is what matters here. IVUS and OCT have been separate categories with separate console hardware, separate workflows, and separate purchase decisions for hospitals. A single combined system collapses two equipment line items into one and removes the workflow penalty of switching modalities mid-procedure. Founders building enabling devices in interventional cardiology should read this as a signal that the procurement environment is moving toward integrated systems that reduce capital footprint and case-time variance, even at the cost of the specialty advantages a best-of-breed single-modality system might offer. The buyer that decides is increasingly the hospital, not the cardiologist alone. ## Protaryx Medical Cleared a Transseptal Puncture Device That Targets Workflow Time Protaryx Medical, a Baltimore-based startup, secured FDA 510(k) clearance for its next-generation transseptal puncture device used for left-heart access during minimally invasive cardiac procedures, according to BioPharma APAC. The system is designed to support zero-exchange delivery, with an atraumatic design intended to reduce procedural steps and remove the need for rewiring. The clearance positions the company to enter a transseptal access market that supports the full range of structural heart and electrophysiology procedures growing in volume across the U.S. The thesis for a startup competing in a category like transseptal access is workflow efficiency in a procedure that already exists, not a new clinical indication. That positioning is harder than founders generally model. The benefit Protaryx is selling, fewer steps and less rewiring, is exactly the kind of advantage a hospital operations team can quantify in seconds saved and disposables removed, but only if the company has the clinical-economic data to attach a number to those savings. If the data is on a planning slide, the procurement conversation never starts. If the data is in a peer-reviewed publication or a real-world registry, it does. The clearance is the entry ticket. The economic story is the ride. Dave’s take The J&J Ottava stapler clearance is the cleanest signal this week of where surgical robotics is actually heading. The companies that win this market will not win on platform mechanics. They will win on everything that travels with the platform, and the work to build that ecosystem starts before the robot ships, not after. If you are building enabling technology, instruments, training infrastructure, or analytics around a robotic platform, the question I would push you on is whether you have a credible position that survives the platform you are betting on losing share, because in a four-way platform race that is the base case for any single bet. The moat is the ecosystem. Build for the ecosystem. From Dave’s video library In this one Dave digs into why customers who sign up still vanish, and lands on the rule that people buy better working, not better looking. A platform moat built from instruments, training, and ecosystem is that rule at medtech scale. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to The Build newsletter: [The Build →](/field-notes/the-build) --- # CMS and FDA Launch RAPID Pathway as Medicare Coverage Compresses to Months for Breakthrough Devices | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-04-28-cms-fda-rapid-medicare-coverage-breakthrough-devices [← All Field Notes](/field-notes) April 28, 2026 Industry Roundup CMS and the FDA jointly announced the Regulatory Alignment for Predictable and Immediate Device pathway on April 23, a Medicare coverage program that allows a proposed national coverage determination to be issued the same day a breakthrough device receives FDA marketing authorization. According to MedTech Dive, current Medicare coverage decisions take roughly a year. Under RAPID, that timeline can compress to as little as two months. At HRS 2026 the same week, Boston Scientific reported pivotal first-line persistent AFib data for Farapulse, and Medtronic’s Sphere-9 catheter for ventricular tachycardia received FDA breakthrough designation. SquareMind closed $18 million for an AI-powered skin-imaging robot, with Sonder Capital, the firm co-founded by Intuitive Surgical’s Fred Moll, leading the round. Read together, the week reset the timing math for any breakthrough-eligible device company in the U.S. market. ## CMS and FDA Launched RAPID for Same-Day Medicare Coverage Decisions The Centers for Medicare and Medicaid Services and the Food and Drug Administration unveiled the Regulatory Alignment for Predictable and Immediate Device pathway on April 23, according to MedTech Dive. Under RAPID, evidence generated for FDA review can support Medicare coverage decisions, and CMS can issue a proposed national coverage determination on the same day the device receives FDA marketing authorization. Eligibility requires the device to be FDA-designated breakthrough, addressing an unmet clinical need for Medicare beneficiaries, and either a Class III device or a Class II device participating in FDA’s Total Product Life Cycle Advisory Program. CMS estimates roughly 40 devices currently qualify with another 20 potentially eligible. The Transitional Coverage for Emerging Technologies pathway will pause for new candidates as CMS focuses on RAPID implementation. The strategic implication for founders building breakthrough-eligible devices is meaningful and underappreciated. Under the prior model, the gap between FDA clearance and Medicare coverage was a known capital trap. A device cleared without coverage in hand had to fund a year of commercial activity, evidence generation, and reimbursement work against revenue that hospitals could not yet recover from CMS at scale. The capital plan, the hiring plan, and the salesforce plan all assumed that gap. RAPID does not eliminate the work, but it changes the time horizon over which the work has to be funded. A founder running a breakthrough-eligible device program now has a credible path to commercial revenue inside 90 days of FDA marketing authorization rather than 12 to 18 months. The financing implications are significant for any device that fits the eligibility criteria. ## Boston Scientific’s AVANT GUARD Trial Cleared a First-Line PFA Position in Persistent AFib Boston Scientific reported AVANT GUARD trial results at HRS 2026 showing pulsed field ablation outperformed antiarrhythmic drug therapy as a first-line treatment for persistent atrial fibrillation, per MedTech Dive. The 440-patient study compared the Farawave catheter to drug therapy and reported 12-month effectiveness of 56% for PFA versus 30% for drugs, with a major adverse event rate of 5.1%. The data was published in the New England Journal of Medicine. Boston Scientific framed the result as supporting an expansion of the Farapulse approval beyond drug-resistant cases, opening a market the company estimates at roughly $1 billion for persistent AFib treatment by 2028. The structural lesson for founders building in any contested category is what Boston Scientific just demonstrated about evidence sequencing. Farapulse won initial approval in paroxysmal AFib. The persistent AFib indication was a separate clinical population with separate evidence requirements, and the company invested in a randomized trial against the standard-of-care alternative. The AVANT GUARD result is now both a label-expansion driver and a reimbursement-grade evidence asset that travels into procurement conversations with cardiology programs nationally. Founders whose evidence strategy stops at the initial indication leave the most valuable expansion move on the table. The category leaders are running the next-indication trial in parallel with launch, not after launch. ## Medtronic’s Sphere-9 in Ventricular Tachycardia Earned FDA Breakthrough Designation Medtronic presented late-breaking early feasibility data for the Affera mapping and ablation system with the Sphere-9 catheter in post-MI ventricular tachycardia at HRS 2026, with 65.5% of patients remaining VT-free at six months, according to MedTech Dive. Concurrent with the data, the FDA granted Sphere-9 breakthrough device designation for VT treatment, and a U.S. pivotal trial has been authorized. The combination of a breakthrough designation and post-MI VT data positions Sphere-9 directly inside the eligibility window for the new RAPID coverage pathway when the device eventually receives marketing authorization. For founders building enabling devices in interventional electrophysiology, the read-through is that the breakthrough designation is now an even more strategically valuable regulatory artifact than it was a week ago. The designation has always pulled forward FDA review timelines and signaled clinical importance to investors. Under the new framework, it also unlocks a potential same-day Medicare coverage decision, which substantially compresses the path from clearance to first commercial revenue. If your regulatory strategy is not actively pursuing breakthrough designation for an eligible indication, you are leaving a year of revenue runway on the table. ## SquareMind Raised $18 Million for an AI-Powered Skin-Imaging Robot SquareMind, a Paris-based medical robotics startup, announced an $18 million round to support U.S. and European launches of its AI-powered skin-imaging robot, according to coverage from Tech Startups. The round was led by Sonder Capital, the venture firm co-founded by Intuitive Surgical’s Fred Moll, with participation from Bpifrance’s Deeptech 2030 fund, Adamed Technology, Calm/Storm Ventures, and Teampact Ventures. The product applies robotic imaging to dermatologic screening and is positioned to enter U.S. and European markets in the near term. Two structural points are worth pulling out for founders thinking about adjacent categories. First, the lead investor signal matters. Sonder Capital is one of a small set of funds whose pattern recognition in surgical and interventional robotics is genuinely deep, and the firm continues to back robotic imaging companies operating in clinical adjacencies to surgery. The investment thesis is that imaging robotics is a category that travels with the surgical robotics ecosystem rather than competing with it. Second, the geographic strategy is a U.S. and European simultaneous launch, not a European-first pattern that would have been more typical for a Paris-based startup five years ago. The capital and regulatory pathways for robotic imaging in the U.S. have moved enough to make a parallel launch feasible at this stage of company maturity. Founders building in robotic imaging or robotic-adjacent diagnostic devices should read that as a market signal about timing. Dave’s take The RAPID pathway is the most structurally important medtech regulatory news in years for any company building a breakthrough-eligible device for Medicare beneficiaries. Most founders I talk to treat the FDA pathway as the financing milestone and Medicare coverage as a downstream commercial issue, which is exactly the model that just stopped being correct. If your device fits the breakthrough criteria, Medicare coverage is now a clearance-week event, and your pricing strategy, salesforce plan, and capital plan need to be rebuilt around that compressed timeline. If your device does not fit the criteria, the question I would push you on is whether the indication you are pursuing could be redesigned to fit, because the gap between RAPID-eligible companies and non-eligible companies in commercial timing is now a year of runway you cannot otherwise buy. From Dave’s video library Dave on selling by walking the buyer through their own economics. The RAPID pathway collapses the gap between clearance and Medicare coverage, which means the buyer's math is now part of the product plan from day one. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to The Build newsletter: [The Build →](/field-notes/the-build) --- # Adagio vCLAS VT Pivotal Anchors a Week of Specialist Interventional Wins | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-04-30-adagio-vclas-vt-pivotal-specialist-interventional-wins [← All Field Notes](/field-notes) April 30, 2026 Industry Roundup While the platform giants have spent the last quarter trading guidance updates and indication breadth, the most interesting interventional news this week came from specialist companies clearing pivotal-grade milestones. Adagio Medical reported six-month results from the FULCRUM-VT pivotal trial of its vCLAS cryoablation system on April 27, hitting strong endpoints in a ventricular tachycardia population that has resisted catheter ablation for years. Imricor began U.S. commercial sales of NorthStar, the first cardiac mapping system designed for MR-guided ablation, on April 23. BrioHealth received conditional FDA approval on April 24 to begin a pediatric IDE trial of its BrioVAD ventricular assist device. Abbott posted six-month FlexPulse data on its TactiFlex Duo dual-energy catheter at HRS 2026. Read together, the week shows what specialist interventional companies look like when their pivotal evidence and U.S. commercial readiness arrive at the same time. ## Adagio Medical’s vCLAS Cryoablation Pivotal Cleared Six-Month VT Marks That No One Else Has Adagio Medical announced positive six-month results from FULCRUM-VT, a prospective multicenter IDE pivotal trial of the vCLAS Cryoablation System for ventricular tachycardia, on April 27 at the Heart Rhythm Society 2026 late-breaking session. The 209-patient trial was conducted at 20 U.S. and Canadian sites in patients with structural heart disease, both ischemic and non-ischemic cardiomyopathy. According to the company release and Medical Device Network coverage, the device achieved 98% acute non-inducibility of targeted VTs, 84% freedom from ICD shocks, 59% freedom from VT recurrence, and a 2.4% major adverse event rate at six months. Mean ablation time was 54 minutes per patient. The trial was the first to show equivalent effectiveness across both ischemic and non-ischemic VT substrates, which the company is positioning as a regulatory and commercial differentiator. The structural read for founders is that VT ablation has been a durable white space underneath the larger AFib market, and a pure-play specialist just produced pivotal-grade data inside that space while the platform companies kept their attention on dual-energy AFib catheters. Adagio is a small public company without the salesforce, market cap, or installed base of the EP incumbents. The pivotal-grade endpoint set, the subsetted ischemic and non-ischemic effectiveness, and the safety profile constitute exactly the package a focused specialist needs to convert pivotal data into reimbursement-grade evidence. The lesson for founders building in EP-adjacent or other underserved interventional spaces is that the platform companies will continue to optimize for indication breadth, and that creates persistent room for specialists who finish the evidence work on a single hard substrate. ## Imricor Began U.S. Commercial Sales of NorthStar, the First MR-Guided EP Mapping System Imricor Medical Systems began U.S. commercial sales of its NorthStar Mapping System on April 23, with the system showcased at HRS 2026 in Chicago. According to the company announcement, NorthStar is the first cardiac mapping and guidance system designed specifically to operate inside the MRI environment. The system gives physicians a real-time soft-tissue visualization of cardiac structure and tissue characteristics during interventional procedures, with no fluoroscopy required. Imricor reported that the U.S. commercial team is generating early demand and building a pipeline focused on early-adopter centers. The implication for founders building in interventional imaging or hybrid procedural environments is that the MR-guided EP suite has finally cleared the regulatory and infrastructure thresholds that have stalled it for a decade. Imricor is a small-cap U.S.-listed company, and the U.S. commercial launch is a meaningful operational milestone after years of development and European-first deployment. Founders developing devices that depend on imaging or environmental innovation should read the NorthStar launch as a proof point that institutional demand for radiation-free EP procedures is now real enough to support a focused commercial ramp. ## BrioHealth Received Conditional FDA Approval to Run a Pediatric VAD IDE Trial BrioHealth Solutions announced on April 24 that it received conditional FDA approval to launch the Brio4Kids Trial, a pediatric clinical evaluation of the BrioVAD System under the existing INNOVATE IDE. According to the company press release, the trial will study the BrioVAD in pediatric patients with advanced heart failure, with enrollment expected to start in mid-2026 and initial data anticipated in Q4 2026. The BrioVAD device features a fully magnetically suspended pump and a thin driveline, designed to address gaps in pediatric heart failure treatment. For founders building in advanced heart failure or other pediatric-indication device categories, BrioHealth’s conditional approval is a useful reference for how the FDA is handling parallel adult and pediatric IDE designs. Building the pediatric extension under the existing IDE rather than starting a separate program preserves capital and timeline. The structural lesson is that pediatric extensions can be designed into the original IDE strategy from the outset, rather than treated as a follow-on program once the adult device is commercially established. ## Abbott’s TactiFlex Duo Posted Six-Month Dual-Energy AFib Data at HRS 2026 Abbott reported six-month results from its FlexPulse trial of the TactiFlex Duo dual-energy ablation catheter at HRS 2026, per MedTech Dive. The 188-patient study in intermittent atrial fibrillation reported 87% arrhythmia-free survival at six months and a 98.3% safety profile with no major safety events. The TactiFlex Duo combines pulsed field and radiofrequency energy in a single contact-force-sensing catheter, positioning Abbott as the third major EP company with a credible dual-energy portfolio entry behind Boston Scientific and Medtronic. The data lands in the same conference cycle where Boston Scientific reported its AVANT GUARD first-line PFA results in persistent AFib and Medtronic showed early Sphere-9 VT data. The competitive read is that dual-energy is now table stakes for EP platform players, and the next round of differentiation will be on indication-specific evidence and reimbursement positioning rather than on energy modality alone. Founders building niche or specialist EP devices should treat dual-energy AFib as a settled platform story and plan their evidence and commercial strategy against substrates and indications where the dual-energy approach does not yet have differentiated data. Dave’s take The specialist pattern this week is the one I would push founders to study most carefully. The platform companies are running indication-breadth plays because that is what their cost structure requires. Adagio, Imricor, and BrioHealth are running focused-substrate plays because that is what their cost structure can sustain. If you are building a single-substrate or single-indication interventional device, your evidence work has to land at pivotal-grade depth on the substrate the platform companies have not yet contested, and your commercial readiness has to arrive at the same time. The companies that finish that combination capture the durable white space; the companies that finish only one half end up acquired into the platform play they were trying to differentiate from. From Dave’s video library Dave's four-question filter for choosing one focused bet over five diluted ones. Adagio, Imricor, and BrioHealth are running exactly that play: a single substrate, a cost structure that can sustain it, and nothing else. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to The Build newsletter: [The Build →](/field-notes/the-build) --- # CMR Surgical’s Versius Plus Gynecology 510(k) Anchors a Week of Indication Expansion | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-05-01-cmr-versius-plus-gynecology-510k-indication-expansion [← All Field Notes](/field-notes) May 1, 2026 Industry Roundup The last week of April closed with a set of announcements that had less to do with platform competition and more to do with how serious operators are widening the reach of devices that already exist. CMR Surgical filed a 510(k) on April 29 to add benign gynecology indications to Versius Plus, ten days after laying out the U.S. commercial program for cholecystectomy. enVVeno Medical received what it called the first-ever FDA IDE approval for a U.S. pivotal study of a non-surgical replacement venous valve on April 29. The EMA, the European Commission, and the MDCG opened applications for a new EU Breakthrough Medical Device Pilot on April 28. The clock on EUDAMED’s mandatory first four modules ran down to four weeks before the May 28 deadline. Prodeon Medical disclosed Urocross Expander minimally invasive BPH outcomes at AUA 2026. The pattern across the week is that the next leg of value creation in surgical robotics and advanced interventional devices is going to come from indication-expansion programs and regulatory-readiness work, not from new platform launches. ## CMR Surgical Filed a 510(k) to Take Versius Plus Into U.S. Gynecology CMR Surgical announced on April 29 that it has submitted a 510(k) premarket notification to the FDA for Versius Plus in benign gynecology, including total hysterectomy, oophorectomy, salpingectomy, and other gynecologic procedures. The filing builds on the cholecystectomy clearance announced earlier in April and on a global Versius installed base where gynecologic procedures already account for a large share of recorded cases. The company press release noted that Versius is the second most utilized surgical robotic system globally outside the United States and that gynecologic adoption has been a leading driver of that growth. The strategic read for surgical robotics founders is that CMR is running an indication-expansion ramp on a U.S. timeline rather than a U.S. launch on a single indication. Cholecystectomy clears the channel, validates installations, and trains hospital teams. Gynecology then drops onto the same installed base and converts gynecologic surgical demand into Versius procedure share without requiring a new sales motion. CMR has the global procedure data to support the gynecology case, and the company is using that data to compress the U.S. evidence cycle. Founders building competing or adjacent platforms should treat the gynecology submission as the start of CMR’s U.S. revenue ramp, not as a routine indication add. The other read is operational. Submitting a second 510(k) within weeks of the first is a signal that the regulatory function and the clinical evidence packages were built in parallel on a multi-year timeline, not sequenced. Founders running a single-indication clearance program with the second indication still in evidence collection are giving up the cadence advantage CMR is now demonstrating. ## enVVeno Received the First-Ever FDA IDE for a Non-Surgical Replacement Venous Valve enVVeno Medical announced on April 29 that the FDA approved an IDE to begin a U.S. pivotal study of the enVVe System for severe deep chronic venous insufficiency, which the company described as the first-ever IDE for a non-surgical replacement venous valve. The Transcatheter Venous Valve Endoprosthesis pivotal study evaluates a minimally invasive device for a patient population that has historically had limited durable interventional options. The company framed the IDE clearance as the gateway to a multi-year pivotal program in a category where there is no incumbent device. For founders building in deep venous interventions or other underserved categories, the enVVe IDE is a useful regulatory data point. The FDA cleared a pivotal program in a device class where there is no precedent device, no predicate, and no established trial infrastructure. The implication is that founders building first-in-class interventional devices have a workable IDE pathway when the evidence package and the regulatory engagement are tight enough, and that deep CVI is now an active pivotal trial space rather than a research-only category. The downstream commercial economics will depend on payer coverage, but the regulatory door is now demonstrably open. ## EU Breakthrough Medical Device Pilot Opened Applications on April 28 The European Medicines Agency, the European Commission, and the Medical Device Coordination Group launched a Breakthrough Medical Device Pilot on April 28, with applications for the first phase open until May 22. According to the EMA announcement, the pilot covers highly innovative technologies, including neuro-implants and advanced robotics, with a starting focus on cardiovascular technologies. Designated devices receive enhanced regulatory support and priority scientific advice from EMA expert panels, and the pilot runs in three phases through 2027 with the goal of laying the groundwork for a permanent EU breakthrough pathway. For founders with a CE Mark on the table for a genuinely novel device, the pilot is a near-term tactical opportunity. The four-week window between launch and the first-phase application deadline is tight, and the pilot is expressly aimed at the kind of devices that have struggled to find a coherent path through MDR. Surgical robotics, neuromodulation, and cardiovascular interventional founders with regulatory work already in motion should be evaluating whether their device profile fits the pilot scope and whether the priority scientific advice is worth the application effort. The applications close May 22, which is also the same week that EUDAMED becomes mandatory. ## EUDAMED Mandatory Use Starts May 28 The first four modules of EUDAMED, covering actor registration, UDI/device registration, notified bodies and certificates, and market surveillance, become mandatory on May 28, 2026. New MDR and IVDR devices placed on the EU market on or after that date must be registered in the UDI/Device module before first-unit placement, and legacy devices already on the market that continue to be marketed beyond that date must be registered by November 28, 2026. Manufacturers and authorized representatives have to register their organizations in the Actor module and prepare device data to the Eudamed functional specifications before any post-deadline market launch. For founders selling into Europe, the EUDAMED activation is an operational milestone disguised as a registration deadline. Companies that are still building the actor registration and UDI submission processes in May are not going to land them clean by May 28, and the cost of missing the window is direct disruption to first-unit placement. The companies that have been treating EUDAMED as a project rather than as a side activity since the November 2025 commission decision are the ones that will land the deadline without operational drag. Surgical robotics and interventional device companies with active EU launches in summer 2026 should be confirming this week that their notified body relationships, UDI structures, and authorized representative registrations are ready for the actor and device modules to go live. ## Prodeon Disclosed Urocross Expander Outcomes at AUA 2026 Prodeon Medical announced on April 28 that three scientific abstracts featuring the Urocross Expander System would be presented at the American Urological Association 2026 Annual Meeting, including data from the Expander-2 Trial reporting that 97% of device retrievals were successfully accomplished using a flexible cystoscope and the Urocross Retrieval Sheath. The Urocross is a minimally invasive device for benign prostatic hyperplasia, in a category that has seen growing interventional activity from incumbents including Boston Scientific’s Rezum and Teleflex’s UroLift, and a new push toward office-based interventional BPH treatment. For founders building in office-based interventional categories, the Prodeon AUA disclosures are part of a broader pattern in which next-generation BPH devices are converting from surgical alternatives into office-based procedures with retrieval steps designed for outpatient workflow. The 97% retrieval figure is the kind of operational outcome that hospital and ASC procurement teams care about, and the abstracts reflect a clinical evidence cycle organized around procedural workflow rather than around isolated efficacy. Founders building in interventional BPH or adjacent office-based categories should read the AUA disclosures as the kind of evidence package that converts a procedural innovation into a payer-grade and procurement-grade story. Dave’s take Indication expansion is where the next year of value creation in surgical robotics and interventional devices is going to be made, and most founders I work with are not running it as the strategic priority it deserves. CMR’s second 510(k) inside the same month is the model: build the gynecology evidence package on a global installed base, file it on the cadence the U.S. ramp requires, and convert installations into multi-indication revenue without rebuilding the sales motion. The companies that treat indication expansion as a regulatory afterthought are leaving the compounding math of a growing U.S. installed base on the table, and the ones who put the indication-two evidence and submission on the same critical path as the first clearance are the ones who will set the pace for the rest of the field. From Dave’s video library Dave on the mistake founders make right after early traction, when the compounding work stops feeling urgent. Indication expansion is that work for a device company: the second 510(k) matters more than the party for the first. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to The Build newsletter: [The Build →](/field-notes/the-build) --- # Stryker’s Q1 Cyber Recovery Reframes Resilience as a Medtech Strategic Asset | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-05-02-stryker-cyber-recovery-medtech-resilience [← All Field Notes](/field-notes) May 2, 2026 Industry Roundup Stryker reported first quarter 2026 results on April 30 with the cyber outage that hit the company on March 11 fully visible in the numbers. Q1 sales of $6.02 billion missed analyst estimates by roughly $320 million, and adjusted EPS of $2.60 came in $0.38 below consensus, with the company attributing the gap to roughly three weeks of operational disruption that delayed shipments and depressed manufacturing absorption. Mako installations posted a record Q1 in both the U.S. and internationally despite the outage, and CEO Kevin Lobo held the company’s full-year organic growth guidance of 8% to 9.5%. One week earlier, Medtronic disclosed and contained a separate cyberattack on its corporate IT systems with no operational impact reported. The combined picture across the two reports is that connected medtech has moved into a sustained cyber-pressure environment, and the strategic question for surgical robotics and interventional device founders is no longer whether to invest in operational resilience but how visibly that resilience shows up when the next attack lands. ## Stryker Q1 Sales Took a $320 Million Hit From the March Cyber Outage Stryker confirmed on its April 30 earnings call that the March 11 cybersecurity event caused approximately three weeks of global operational disruption and reduced both shipment volume and manufacturing absorption in the quarter. Q1 net sales of $6.02 billion grew 2.4% organically, well below the 8% to 9.5% trajectory the company is guiding for the full year, and adjusted EPS of $2.60 missed consensus by $0.38. Lobo described a recovery process he characterized as fast and successful, citing rapid threat removal, recovery of backups, and what he called strong customer communication during the disruption. The company maintained its full-year organic growth and adjusted EPS guidance, signaling that the lost Q1 volume is expected to be recovered across the remaining quarters rather than written off. The Mako line was the most strategically important data point in the report. Mako had its best ever Q1 for installations both domestically and internationally, with strong utilization rates and continued momentum into the Mako Shoulder launch on Mako 4 mid-year. Lobo’s direct line on the call was that Mako would have been even higher had the company been able to keep shipping through the end of the quarter. The signal for surgical robotics founders is that capital-equipment buying decisions in orthopedic robotics did not pause when the supplier had a global IT outage. Hospitals continued to install systems, the salesforce continued to close, and the customer relationships absorbed the operational stress. The ability of the Mako franchise to compound through the disruption is the kind of operational resilience that translates directly into multi-quarter revenue. For founders building robotic platforms, the relevant takeaway is that cyber-driven operational disruption is now part of the planning environment for major medtech competitors, and the recovery quality is being scored explicitly by both customers and the capital markets. The Q1 stock recovery on the maintained guidance was a vote on the credibility of the recovery, not on the absence of the incident. ## Medtronic Disclosed and Contained a Separate Cyberattack One Week Earlier Medtronic announced on April 24 that it had identified and contained a cyberattack on its corporate IT network, with no impact on products, manufacturing, distribution, customer connections, or financial reporting systems. The company stated that its corporate IT environment is maintained separately from the networks that support its products and operational systems, which limited the blast radius of the incident. The data-extortion group ShinyHunters had earlier claimed responsibility and asserted exfiltration of more than 9 million records, a claim Medtronic has not verified. The disclosure followed an SEC filing on April 18. The contrast between the Medtronic and Stryker outcomes is the operational story of the past six weeks in medtech IT. Medtronic absorbed the attack with no disclosed operational or product impact, supported by the architectural separation of its corporate and operational networks. Stryker absorbed roughly three weeks of global operational disruption, primarily because the affected Microsoft environment was load-bearing across both corporate and operational workflows. The architectural choice that determines how much of the company is taken offline by an attack on the corporate IT environment is now a strategic-grade decision, and it is being graded in earnings reports. For founders building connected medtech and surgical robotics platforms, the architectural separation question is now on the same operational priority list as cybersecurity hygiene. A platform whose product, manufacturing, and customer-connection networks are downstream of a single corporate IT environment will see the next disruption translate into the same kind of operational outage Stryker absorbed. A platform whose operational and product networks are run as a separate environment can continue to operate through a corporate IT incident, which is the architectural posture that allowed Medtronic to disclose without an operational impact. ## The Sector Now Has a Sustained Pattern of Targeting The Stryker and Medtronic incidents come on top of a broader pattern of cyberattacks against U.S. medical device makers in 2025 and 2026, with separately reported intrusions affecting other major medtech players over the past 12 months. The threat-actor pattern is mixed across nation-state-linked groups and financially motivated extortion crews, with Iran-linked actors named in the Stryker incident and ShinyHunters claiming the Medtronic intrusion. The diversity of the attacker profiles is itself a signal that medtech is now a recognized target category, not an opportunistic one. The operational implications for founders building in this category are concrete. Hospital procurement teams are increasingly asking suppliers about the architectural separation of corporate IT from product and operational systems, the documented incident-response posture, and the resilience of the supply chain to single-environment outages. Investor diligence is asking about the cybersecurity-readiness posture of the engineering team and the planned spend ratio for security work as a percentage of R&D. The work is no longer optional for early-stage companies that intend to ship a connected device into a U.S. hospital network in the next 36 months. ## What Changes for Surgical Robotics and Interventional Device Founders Three operational shifts follow from this week’s reports. The first is that cybersecurity work has moved from a compliance line item into a category that competes for capital with clinical evidence, regulatory submissions, and commercial buildout, and it deserves the same operating cadence and senior leadership ownership. The second is that the architectural separation of corporate IT from product and operational systems is now a design decision with direct earnings consequences, and the right time to make that decision is during early product architecture rather than after the first major hospital deployment. The third is that the credibility of the incident-response posture, including the runbooks, the practice cadence, and the customer-communication script, is now a piece of the diligence story for both hospital procurement and downstream investors. Founders building connected surgical robotics and interventional devices should treat this week’s reports as the working baseline for what hospital and investor counterparties expect. The companies that have been treating cybersecurity as an IT-team problem are going to be visibly behind the companies that have been treating it as a strategic operating problem. Dave’s take The Stryker recovery was impressive and the Mako numbers prove that customer relationships can absorb a serious operational hit when the supplier handles the recovery well. What I want every surgical robotics and interventional device founder I work with to take from this week is that the architectural decision to separate operational networks from corporate IT is the kind of unglamorous infrastructure work that nobody asks about until it determines whether you ship through your next incident or stop shipping for a month. Make the decision early when it costs a few engineering quarters, not late when it costs a quarter of revenue and a chunk of your customer trust. From Dave’s video library Dave pressure-tests the claim that the hard part of building has been removed, and shows it just moved. Stryker's cyber recovery is a case study in the unglamorous execution work that never shows up on a pitch slide but decides who survives. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to The Build newsletter: [The Build →](/field-notes/the-build) --- # Stealth AXiS Goes Global as Integrated Platforms Redraw Surgical Robotics Competition | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-05-03-stealth-axis-ce-mark-platform-stack-surgical-robotics [← All Field Notes](/field-notes) May 3, 2026 Industry Roundup Medtronic announced CE Mark for the Stealth AXiS surgical system on April 28, clearing the platform for spine and cranial procedures across the European Union and the broader CE-accepting markets. Five days earlier, three spine surgeons at HCA Virginia’s Reston Hospital Center performed the first U.S. surgeries using the cleared Stealth AXiS Autopilot system, completing a staged direct lateral interbody fusion and a three-level posterior instrumented spinal fusion. The two milestones together close out the launch sequence for the first surgical platform that integrates AI-powered planning, intraoperative navigation, segmental tracking, and robotic guidance into one product, and they land in the same quarter that Intuitive Surgical lifted full-year da Vinci procedure growth guidance and CMR Surgical filed a 510(k) to add gynecology to Versius Plus. The pattern across the three is the same. Integrated platforms that bundle planning, navigation, AI, and robotic delivery are now the unit of competition in surgical robotics, and the standalone-robot frame is being retired. ## Stealth AXiS Closes Out the Global Launch Sequence in Six Weeks Medtronic received CE Mark for Stealth AXiS on April 28, with the certification covering both spine and cranial procedures and authorizing commercial launch into the European Union and other CE-accepting markets. The CE Mark sequenced behind FDA spine clearance in February and FDA cranial and ENT clearance in March, completing a six-week run from initial U.S. authorization to global commercial readiness. Medtronic disclosed a partnership with GE HealthCare to integrate the platform with intraoperative ultrasound, which extends the integrated-imaging story already built around StealthStation navigation and Mazor robotics. The system also includes LiveAlign segmental tracking, which Medtronic describes as the first capability to visualize anatomic motion and patient alignment in real time during spine procedures without repeat imaging. On April 23, three spine surgeons at HCA Virginia’s Reston Hospital Center performed the first surgeries using the cleared U.S. Autopilot system, treating a single patient through a staged direct lateral interbody fusion and a three-level posterior instrumented spinal fusion with bilateral SI joint fusion. The clinical milestone matters less for the procedural detail than for the operating-cadence signal it sends. The platform reached commercial readiness, scheduled the first surgeons, completed the first cases, and produced the case report inside ten weeks of FDA clearance. That is the operational tempo competitors are now being measured against. ## Intuitive Lifted 2026 Procedure Guidance on da Vinci 5 Mix Shift Intuitive Surgical reported Q1 2026 results on April 22 with worldwide revenue of $2.77 billion, up 23% year over year, and lifted full-year da Vinci procedure growth guidance to a range of 13.5% to 15.5%, up from 13% to 15% prior. Total da Vinci procedures grew 16% in Q1 and Ion procedures grew 39%. The company placed 431 da Vinci surgical systems in the quarter against 367 a year earlier, with da Vinci 5 placements at 232 against 147, indicating a continued mix shift toward the integrated next-generation platform. CEO Dave Rosa attributed U.S. momentum to a 31% rise in after-hours procedures and higher overall utilization, both of which depend on the platform-level workflow features that differentiate da Vinci 5 from its predecessor and from competing standalone systems. The strategic read for surgical robotics founders is that Intuitive’s pricing power and procedure growth in 2026 are being driven by the integrated platform, not by any single instrument or capability. The customers paying the premium for da Vinci 5 are paying for the integrated workflow, not for the robot in isolation. Competitors trying to displace Intuitive on price or on a single feature continue to lose ground, while competitors fielding their own integrated-platform answers are the ones building durable market positions. ## CMR Surgical Builds the Versius Plus Indication Stack Ahead of U.S. Launch CMR Surgical filed a 510(k) on April 29 to add benign gynecology indications to Versius Plus, ten days after the cholecystectomy clearance and two months after the U.S. debut at the Society of American Gastrointestinal and Endoscopic Surgeons annual meeting in Tampa. The company has crossed 45,000 procedures globally on the Versius platform across more than 30 countries, with the digital analytics suite rolling out across Europe ahead of the major U.S. push later in 2026. The Versius Plus story is the indication-stack version of the integrated-platform story. CMR is building the platform breadth that lets a hospital justify a system purchase against a multi-specialty caseload, not against a single procedure economics case. The European Commission, EMA, and Medical Device Coordination Group launched the Breakthrough Medical Device Pilot on April 28, providing manufacturers of highly innovative platforms with priority scientific advice and a streamlined path to clinical validation. The pilot is sized for the integrated-platform players that need to run multi-specialty clinical programs and indication stacks in parallel with launch, and it lands at the same moment that the integrated-platform competitive frame is consolidating across the U.S. and European markets. ## What the Integrated-Platform Frame Changes for Surgical Robotics Founders Three operational shifts follow from this week’s milestones. The first is that the unit of competition has become an integrated platform that bundles planning, intraoperative navigation, segmental tracking, AI-driven workflow, and robotic delivery into a single product. A surgical robotics startup that ships a strong robotic motion platform without the surrounding planning, navigation, and AI workflow story is competing into a market where the leaders are selling something fundamentally larger. The second is that integrated-imaging partnerships, of the kind Medtronic announced with GE HealthCare and Intuitive has been building inside da Vinci 5, are now part of the platform spec rather than a future enhancement. The third is that indication breadth, of the kind CMR is building on Versius Plus, is now what justifies the capital commitment a hospital makes to the platform, and a single-indication launch strategy without a clear indication-stack roadmap will struggle to land the first deployments. Founders building surgical robotics platforms in 2026 should be reading the Stealth AXiS launch sequence as the new reference for what global commercial readiness looks like, the Intuitive Q1 report as the evidence that integrated-platform pricing power is durable, and the CMR Versius Plus indication stack as the model for how a multi-specialty platform earns the hospital purchase decision. The strategic question for the founders I work with has become what the smallest credible integrated platform looks like for their specialty, and how fast they can get to it. Dave’s take The Stealth AXiS launch sequence is the cleanest example I have seen of what the integrated-platform competitive frame looks like in practice, and the pace from FDA clearance to first surgeries to CE Mark is what every surgical robotics founder should be planning against. If you are building a robot without a planning, navigation, and AI workflow story, you are building into a market where the leaders are selling something fundamentally larger and the customers are paying for the difference. Build the smallest credible integrated platform for your specialty, build it from initial product architecture, and stop pretending the differentiation can be added in a later release. From Dave’s video library Dave on the AI he actually runs his business on and where each workflow breaks. The Stealth AXiS launch sequence shows what it looks like when AI stops being a feature and becomes the workflow itself. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to The Build newsletter: [The Build →](/field-notes/the-build) --- # Surgical Robotics Moves to ASCs as the Site-of-Care Shift Resets the Platform Spec | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-05-04-surgical-robotics-asc-site-shift-platform-spec [← All Field Notes](/field-notes) May 4, 2026 Industry Roundup Distalmotion told MedTech Dive on April 29 that its DEXTER robotic surgery system has now been used in more than 3,000 procedures across Europe and the U.S., that Q1 2026 was the company’s strongest quarter to date, and that its first U.S. ambulatory surgery center installation went live in Wichita in March. The pitch is that surgeries are leaving hospital operating rooms for ASCs faster than the legacy surgical robotics platforms can follow, and that a 125-square-foot, four-minute-setup, single-use-instrument robot is the spec that fits where the volume is going. The same week brought J&J’s definitive agreement to acquire Atraverse Medical, the launch of the PERSIGMA head-to-head trial pitting Varipulse Pro against Boston Scientific’s Farapulse, and a $150 million Series E for Aidoc’s clinical AI foundation model. Read across the four, the next phase of surgical robotics and interventional medtech is being defined by where the procedures actually happen, not by which robot ships the most instruments. ## Distalmotion Crosses 3,000 Procedures and Plants the Flag in U.S. ASCs Distalmotion CEO Greg Roche told MedTech Dive on April 29 that DEXTER has surpassed 3,000 cumulative procedures in Europe and the U.S., that Q1 2026 was the company’s best quarter, and that Cypress Surgery Center in Wichita, Kansas became the first U.S. ASC to acquire a DEXTER system in March. The platform is purpose-built for the lower-acuity site of care: a 125-square-foot footprint that fits through a standard doorway, four-minute setup, single-use instruments that eliminate sterilization workflow, and an open architecture that accepts third-party visualization and instruments. FDA de novo authorization has stacked through inguinal hernia repair in 2024, gallbladder removal and hysterectomy in 2025, and additional gynecology indications and ventral hernia repair pending. Distalmotion has raised approximately $400 million to date, including a $150 million Series G and a strategic investment from J&J’s venture arm. The structural backdrop for the strategy is that Medicare-certified ASCs grew at an average annual rate of 2.2% over five years and now exceed 6,400 facilities, while hospital openings have lagged closures over a 14-year window through 2023. Procedural migration into the ASC channel is the largest site-of-care shift in surgery, and the surgical robotics platforms designed for hospital operating rooms struggle to meet the ASC spec on footprint, capital intensity, and sterilization economics. The Distalmotion bet is that the platform engineered for the site where volume is moving wins the next phase of competition, even as Intuitive continues to dominate the installed hospital base. ## J&J Adds Atraverse to the Cardiac Ablation Stack as PFA Heads to the Same Site Shift J&J announced a definitive agreement on April 27 to acquire Atraverse Medical, developer of the Hotwire transseptal access system, with the deal expected to close in Q2 2026 and financial terms undisclosed. The Hotwire radiofrequency guidewire received initial FDA clearance in May 2024 and has been used in nearly 3,000 procedures by close to 100 cardiac electrophysiologists and interventional cardiologists since limited launch. J&J highlighted the impedance-sensing automatic energy shutoff, multi-sheath compatibility, and intracardiac echocardiography tip visibility as the differentiators it is integrating into the broader cardiovascular portfolio. J&J’s cardiovascular unit grew 13% in Q1 2026 to $2.38 billion, driven by Abiomed and Shockwave assets that the company has been weaving into a single ablation, mapping, and access stack. The day after the Atraverse announcement, J&J disclosed the PERSIGMA randomized trial enrolling up to 466 patients across 50 sites to compare Varipulse Pro against Boston Scientific’s Farapulse in persistent atrial fibrillation. Endpoints are seven-day adverse events and freedom from atrial tachyarrhythmia from day 61, with first cases at HRS 2026 in Chicago. Pulsed field ablation is the procedural innovation accelerating the migration of complex AFib cases out of the academic medical center and into community electrophysiology suites, and a head-to-head trial against the category leader is the evidence J&J needs to displace Farapulse from the procedural workflow that follows the site shift. ## Aidoc Raises $150M for a Clinical AI Foundation Model Already Live in 2,000 Hospitals Aidoc closed a $150 million Series E on April 29, led by Goldman Sachs Alternatives Growth Equity with participation from General Catalyst, SoftBank Vision Fund 2, and NVIDIA’s NVentures. The capital backs CARE, the company’s clinical foundation model, which received what the company described as the first FDA clearance for a comprehensive double-digit foundation-model-based triage system in clinical imaging earlier this year. CARE is deployed through the aiOS enterprise platform across nearly 2,000 hospitals and analyzes more than 60 million patient cases annually. Funding will extend additional clinical indications and add automated imaging draft report creation to the workflow. The connection to the site-of-care frame is that clinical AI infrastructure is the only practical way to maintain consistent imaging triage and decision support across a network that spans hospital ORs, hospital outpatient departments, ASCs, and community imaging centers. A radiology read or a pre-procedural imaging triage pipeline that runs identically across the multi-site enterprise is what makes the procedural migration economically viable for the hospital systems and platform vendors operating across both sides of the shift. The Aidoc round is enterprise-scale clinical AI for the network where surgical and interventional procedures are now distributed, not concentrated. ## What the Site-of-Care Shift Changes for Surgical Robotics and Interventional Founders Three operational shifts follow from this week. The first is that the platform spec for surgical robotics is being rewritten around the ASC environment, where small footprint, fast setup, single-use instrumentation, and capital-light commercial models are gating procurement decisions that the hospital-OR generation of platforms was never designed to win. The second is that the strategic acquirers are assembling cardiovascular and electrophysiology portfolios that follow the procedural migration, and J&J’s combination of Varipulse Pro, Carto 3 mapping, the Atraverse transseptal access system, and the PERSIGMA evidence campaign is the integrated answer to where AFib cases are now being treated. The third is that the clinical AI infrastructure layer is consolidating around a small number of foundation-model platforms with the regulatory clearances and the multi-thousand-hospital deployment footprint to operate across the entire site network, and the venture-backed AI tools that do not get into that infrastructure layer will struggle to land enterprise contracts in the post-shift world. Founders building surgical robotics platforms in 2026 should be reading the Distalmotion ASC entry as the new commercial frontier, the J&J cardiac ablation stack-build as the model for following the procedural migration into the lower-acuity site of care, and the Aidoc Series E as the signal that clinical AI infrastructure for the multi-site enterprise is now the layer the strategic acquirers and the hospital systems are paying for. Dave’s take The site-of-care shift is the structural change that decides which surgical robotics and interventional platforms compound through the rest of the decade, and the founders I work with should be designing for the ASC and the community electrophysiology suite from initial product architecture, not retrofitting after first hospital clearance. The platform that wins the next phase is the one that fits the room where the procedures are moving, with the workflow that the lower-acuity site can actually run, and the commercial model that matches the smaller capital budget. Build for where volume is going, not for where the legacy installed base sits today. From Dave’s video library Dave on finding the single riskiest assumption before you build. If the ASC is where surgical robotics is heading, the platform spec is an assumption to test at the architecture stage, not a retrofit after first hospital clearance. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to The Build newsletter: [The Build →](/field-notes/the-build) --- # Olympus-EndoRobotics Deal Reframes Distribution as a Surgical Robotics Channel Strategy | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-05-05-olympus-endorobotics-distribution-channel-strategy [← All Field Notes](/field-notes) May 5, 2026 Industry Roundup Olympus announced on May 4 a global distribution agreement with South Korea-based EndoRobotics, taking the Robopera console, articulated grippers, scope-mounted traction device, and EndoCubot ESD training simulator into Olympus’ EndoTherapy commercial channel and beginning the rollout in the U.S. The deal arrives ten months after Olympus and Revival Healthcare Capital launched Swan EndoSurgical with up to $458 million in committed capital and an Olympus acquisition option, which means the company is now running a parallel build-and-buy strategy for robotic GI alongside the new license-and-distribute path. The same week brought late-breaking pulsed field ablation data at HRS 2026 in Chicago that shifted the share question between Boston Scientific, Medtronic, and Abbott, an Abbott AI-enabled imaging clearance that drove 9.5% comparable sales growth in vascular technologies, and a Medtronic Mosaic Neo mitral replacement clearance with the first robotic-assisted implantations already complete. Read across the four, the surgical robotics and interventional commercialization map is widening, and OEM distribution is now a real third path alongside acquisition and direct field force. ## Olympus Signs EndoRobotics for Robopera ESD Platform Distribution Olympus and EndoRobotics announced an exclusive global distribution agreement covering the Robopera robot-assisted system, the articulated grippers, the scope-mounted traction device, and the EndoCubot training simulator, with the rollout beginning in the United States as part of Olympus’ EndoTherapy portfolio. The clinical target is endoscopic submucosal dissection for early-stage GI cancers and precancerous lesions, where Olympus already owns the dominant flexible endoscopy installed base and the third-space endoscopy training pipeline. EndoRobotics is a South Korea-based developer of robotic devices for the GI tract, and Mike Callaghan, general manager of EndoTherapy at Olympus, said the partnership is intended to help clinicians perform complex therapeutic procedures with greater confidence. The deal is a meaningful structural signal because Olympus is running it in parallel with the Swan EndoSurgical platform that Olympus and Revival Healthcare Capital launched in July 2025, with up to $458 million committed and an Olympus acquisition option attached to the spinout’s development milestones. The combined message to the surgical robotics and interventional ecosystem is that the largest flexible endoscopy OEM in the world is willing to distribute external robotic platforms while it continues building its own, and that the access pathway for a startup with a credible robotic GI product is now an exclusive commercial relationship with the channel that already owns the procedural environment, rather than a head-to-head sales war against the incumbent. ## HRS 2026 Late-Breakers Reset the Pulsed Field Ablation Share Question The Heart Rhythm Society annual meeting in Chicago produced new clinical evidence across the three competitors driving the $16 billion electrophysiology market into PFA-led growth. Boston Scientific’s AVANT GUARD trial, published in the New England Journal of Medicine and presented at the meeting, showed Farawave delivering 56% primary effectiveness in 440 persistent atrial fibrillation patients at 12 months versus 30% on drug therapy, with a 5.1% major adverse event rate and the deaths reported as procedure-unrelated. Boston Scientific is positioning the data as support for a first-line indication expansion that the company believes opens a $1 billion market by 2028. Abbott reported 87% arrhythmia-free survival at six months from its 188-patient FlexPulse trial of the TactiFlex Duo dual-energy catheter, with a 98.3% safety profile and 93.3% of patients treated entirely on the PFA component. Medtronic disclosed Sphere-9 early feasibility data showing 65.5% freedom from ventricular tachycardia recurrence at six months, with the FDA having granted breakthrough designation and the pivotal U.S. trial approved. The competitive read is that PFA is no longer a category in which any single platform owns the procedural workflow, and the share question is going to be settled across multiple indications and dual-energy combinations during 2026 and 2027. For founders building in the broader cardiac ablation, mapping, and access ecosystem, the implication is that the procedural workflow surrounding PFA is being rebuilt in real time, and the components that integrate cleanly into the new generation of ablation suites have a measurable commercial window that they did not have when the workflow was still organized around radiofrequency thermal ablation alone. ## Abbott AI-Enabled Imaging Clearance Drives Vascular Tech Growth Abbott obtained both FDA approval and a CE mark for an artificial intelligence-enabled imaging device, and the company linked the clearance to 9.5% comparable sales growth in its vascular technologies segment during the first quarter, according to MedTech Dive coverage of the May 4 disclosures. The clearance falls into the AI lifecycle management territory that the FDA device center has been working through advisory committee guidance to formalize during 2026, and it is one of a growing number of AI-enabled clearances the agency is now processing alongside the device backlog driven by FDA staffing changes earlier in the year. The signal for surgical robotics and interventional founders is that AI-enabled imaging is now contributing measurable top-line growth at the strategic-acquirer scale, and the platform companies are pricing the AI-imaging asset pipeline as a primary commercial driver rather than as a supplemental software layer. Founders building AI imaging components for image-guided intervention should be reading the Abbott disclosure as evidence that the strategic interest is real, that the FDA pathway is open for clearances of comparable scope, and that the strategic acquirers have an active purchase appetite for assets that integrate into their existing vascular and cardiology platforms. ## Medtronic Mosaic Neo Mitral Valve Clearance With Robotic-Assisted Cases Medtronic received FDA clearance on April 29 for an updated Mosaic Neo mitral replacement valve, and the company disclosed that initial robotic-assisted implantations had already been completed across U.S. centers as part of the launch. The clearance broadens the Medtronic structural heart portfolio in mitral surgery and confirms the company’s continued investment in the valve replacement category alongside the Hugo robotic-assisted surgery system rollout, which performed its first U.S. case at Cleveland Clinic in February. The robotic-assisted mitral implantation flag in the launch communications is the company drawing the connection between the structural heart asset base and the soft-tissue robotic platform that will eventually support more of those cases. For founders building in the structural heart, cardiac surgery, and surgical robotics adjacencies, the implication is that the strategic acquirers are now thinking about valve assets and robotic platforms as a single integrated commercial story, and the integration question is going to drive the next round of strategic interest in the cardiac robotic ecosystem. The Teleflex CEO appointment on April 30 of Jason Weidman, a 20-year Medtronic veteran, fits the same pattern, with the company restructuring its product portfolio under leadership that has the strategic-acquirer playbook in its operating background. ## What the Distribution Path Changes for Surgical Robotics Founders Three operational shifts follow from this week. The first is that strategic OEM distribution is now a real commercialization path for surgical robotics and interventional device startups, and the Olympus-EndoRobotics structure of an exclusive global distribution agreement with the channel that already owns the procedural environment is the model that founders should be evaluating against the cost and timeline of building their own field force. The second is that the procedural workflow around pulsed field ablation is being rebuilt across 2026 and 2027, and the components that fit the new generation of ablation suites have a measurable window that disappears once the new workflow is locked in. The third is that the strategic acquirers are integrating AI imaging, structural heart, and robotic-assisted surgical platforms as a single commercial story, and the founders building components that fit cleanly into that integrated story have a different conversation with the strategic-acquirer ecosystem than the founders building standalone point solutions. Dave’s take The Olympus-EndoRobotics deal is the clearest 2026 example of strategic distribution as a third commercialization path for surgical robotics, alongside the two paths most founders already model for. The companies I work with should be running the structural diligence on which incumbents already own the procedural environment they are trying to enter, and which of those incumbents have a known appetite for license-and-distribute relationships rather than a build-or-buy posture. That diligence work belongs in the commercialization plan from initial product architecture, not after the FDA clearance arrives and the field-force budget is the only path the company can fund. From Dave’s video library Dave on the distribution problem behind every good product nobody has heard of. The Olympus deal is the industrial-scale version: EndoRobotics built the robot, and Olympus already owns the rooms it needs to enter. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to The Build newsletter: [The Build →](/field-notes/the-build) --- # J&J Ottava FORTE Pivotal Resets the Soft Tissue Surgical Robotics Stack | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-05-06-jj-ottava-forte-pivotal-soft-tissue-robotics-stack [← All Field Notes](/field-notes) May 6, 2026 Industry Roundup Johnson & Johnson presented the FORTE pivotal trial results for its Ottava soft tissue surgical robot at the American Society for Metabolic and Bariatric Surgery annual meeting in San Antonio on May 5, with all primary safety and performance endpoints met across 30 Roux-en-Y gastric bypass patients at six U.S. sites and the 30-day follow-up window complete. The company filed for FDA De Novo classification on the basis of the FORTE pivotal package combined with the prior preclinical evidence, with the requested clearance covering general surgery indications including gastric bypass, sleeve-based procedures, small bowel resection, hiatal hernia repair, and upper abdominal surgery. The same week brought confirmation of Intuitive’s measured 2026 cardiac rollout following the January da Vinci 5 nine-indication clearance, the Boston Scientific-Penumbra shareholder vote scheduled for May 6 that moves the largest interventional acquisition of 2026 toward a second-half close, and Boston Scientific Q1 results that confirmed the strategic-acquirer balance sheets remain in active deployment mode for the platforms that fit the integrated commercial story. Read across the four, the soft tissue surgical robotics stack now has three credible cleared or near-cleared platforms, a measurably tighter pivotal trial design template than the field had been assuming, and a strategic-acquirer environment paying premiums for indication breadth and procedural workflow integration. ## J&J Ottava FORTE Pivotal Hits Across the 30-Patient Roux-en-Y Cohort Johnson & Johnson disclosed the Ottava FORTE pivotal results at ASMBS 2026 in San Antonio on May 5. The single-arm open-label prospective design enrolled 30 patients across six U.S. sites for Roux-en-Y gastric bypass and completed the 30-day post-procedure follow-up window, and the company reported that all primary safety and performance endpoints were met. The procedure cohort lost an average of roughly 13.6 kilograms in the first month, and Peter Schulam, J&J chief strategy officer for medtech, characterized the platform as integrating into existing hospital infrastructure in a way that could remove practical adoption barriers. J&J filed for FDA De Novo classification using the FORTE pivotal package combined with the prior preclinical evidence, with the requested indication coverage spanning gastric bypass, sleeve-based procedures, small bowel resection, hiatal hernia repair, and general upper abdominal surgery. The FORTE design is the structural data point that surgical robotics founders should be reading carefully. A 30-patient single-arm pivotal at six U.S. sites with a 30-day primary follow-up is a meaningfully tighter regulatory enrollment than the field has historically assumed for soft tissue robotic systems, and the De Novo route combined with that enrollment signals that the FDA device center is willing to accept compact pivotal designs from sponsors that can deliver clean preclinical packages and well-controlled safety endpoints. The implication for founders building soft tissue robotic platforms is that the regulatory entry path has measurable design optionality, that the cost of the pivotal trial is a real lever rather than a fixed cost, and that the founders who run regulatory strategy alongside engineering from initial product architecture have a different cost-and-timeline conversation with the FDA than the founders who treat the pivotal as a single-design problem to be solved at the end of development. ## Intuitive Cardiac Rollout Defines the Differentiation Lane Around Ottava The FORTE result lands in an environment Intuitive Surgical has spent the first quarter of 2026 redesigning. The FDA cleared the da Vinci 5 system on January 26 for nine cardiac indications, including mitral valve repair, tricuspid valve repair, mitral valve replacement, and left atrial appendage closure. Intuitive announced a measured 2026 rollout to a limited number of U.S. sites, with a dedicated team building the clinical evidence, training programs, and cardiac-specific instruments and accessories that the new indication base requires. The company is also pursuing cardiac approvals in Europe, with the platform already cleared for cardiac use in South Korea. The structural read is that the leader in soft tissue robotic surgery is moving the differentiation perimeter into a cardiac specialty where the procedural workflow is more complex, the training pipeline is longer, and the buyer environment values integrated platform credibility over the multipurpose surgical robotic positioning that defined the prior generation. Hugo and Ottava are positioned for general surgery and urology, and the cleared cardiac perimeter that Intuitive is now building does not directly intersect either platform’s first commercial cycle. The integrated commercial story that strategic acquirers are paying for, however, runs across all three platforms, and the founders building components that fit into general-surgery, urology, gynecology, or cardiac-robotic workflows have measurably different valuation conversations depending on which of the integrated stacks the component fits. ## Boston Scientific Penumbra Shareholder Vote Closes the Largest 2026 Interventional Deal The Penumbra shareholder vote on the $14.5 billion Boston Scientific acquisition is scheduled for May 6, with Boston Scientific guiding for a second-half 2026 close subject to the remaining regulatory clearances. The deal represents the largest interventional acquisition of 2026 to date and consolidates the venous thrombectomy and stroke-care interventional categories under a single platform alongside the prior Boston Scientific Q1 portfolio additions, including the closed Nalu Medical transaction with $588 million preliminary consideration, the completed Cortex acquisition, and the signed Scivita Medical agreement. Boston Scientific reported Q1 net sales of $5.203 billion, an 11.6% year-over-year increase with 9.4% organic growth, and electrophysiology and interventional cardiology and vascular therapies were the named demand drivers across the cardiovascular segment. The Penumbra close confirms that the strategic-acquirer balance sheets remain in active deployment mode for platforms that fit the integrated commercial story. Interventional founders building platforms in venous, peripheral arterial, neurovascular, and structural heart adjacencies should be reading the Boston Scientific deployment cadence as evidence that the buyer environment is paying premiums for the platforms that can demonstrate clean integration into the cardiovascular and neurovascular procedural workflows. The same pattern shows up at Zimmer Biomet, which delivered Q1 2026 organic constant-currency growth of 2.9%, raised the full-year EPS guidance to $8.40 to $8.55 from the prior $8.30 to $8.45, and increased the free cash flow growth guidance to 9% to 11% from the prior 8% to 10%. The sustained operating performance at the strategic-acquirer scale is the demand environment for the next round of medtech transaction activity. ## What the FORTE Design Changes for Surgical Robotics Founders Three operational shifts follow from the FORTE pivotal data and the broader week. The first is that the soft tissue surgical robotics regulatory pathway has measurable design optionality, and the founders running regulatory strategy from initial product architecture have a different cost-and-timeline conversation with the FDA than the founders who treat the pivotal as a fixed-design problem at the end of development. The second is that the integrated commercial story strategic acquirers are pricing across general surgery, cardiac, and interventional platforms is now the unit of valuation, and the components that fit cleanly into one of the integrated stacks have a different conversation with the buyer environment than standalone point solutions. The third is that the strategic-acquirer balance sheets remain in active deployment mode at the largest scale of the medtech ecosystem, and the founders whose platforms arrive at the strategic conversation in the second half of 2026 with credible commercial momentum will see the active deployment environment that the Boston Scientific Penumbra close confirms. Dave’s take The FORTE pivotal design is the most actionable data point of the week for the surgical robotics founders I work with. A 30-patient single-arm 30-day follow-up pivotal at six U.S. sites is real evidence that the FDA device center will accept compact pivotal designs in soft tissue robotic surgery when the preclinical package and the safety endpoints are clean. The regulatory strategy work that produces a compact design has to start at initial product architecture, alongside the engineering and the channel diligence, not after the pivotal trial protocol has already been frozen. From Dave’s video library Dave on making confident decisions with almost no data. A 30-patient pivotal that satisfied the FDA is proof that the right small evidence, designed against the right question, beats a mountain of the wrong kind. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to The Build newsletter: [The Build →](/field-notes/the-build) --- # BD Q2 Interventional Posts 7.3% Growth, Confirming the Surgical Consumables Layer | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-05-07-bd-q2-interventional-surgical-consumables-strength [← All Field Notes](/field-notes) May 7, 2026 Industry Roundup Becton Dickinson reported fiscal Q2 2026 results on May 7 with the BD Interventional segment generating $1.357 billion in revenue, up 7.3% reported and 5.3% on a foreign-currency-neutral basis, and broad strength across Surgery, Urology and Critical Care, and Peripheral Intervention. The result lands in the same week that Intuitive Surgical lifted full-year 2026 da Vinci procedure growth guidance to 13.5% to 15.5% on Q1 strength, that CMR Surgical continues the U.S. Versius Plus rollout following the December 2025 cholecystectomy clearance, and that Restore Robotics now holds FDA 510(k) clearance for four remanufactured da Vinci instruments. Read across the four developments, the surgical robotics platform competition is at the capital-sale layer, and the durable economic engine for the broader interventional and surgical device market sits in the recurring instrument and consumable layer that BD just confirmed is healthy across an entire focused MedSurg-Interventional portfolio. ## BD Q2 Interventional Posts 7.3% Reported Growth Across All Three Sub-Segments BD reported fiscal Q2 2026 (quarter ended March 31, 2026) on May 7 with continuing-operations revenue of $4.714 billion, up 5.2% reported and 2.6% FXN. The BD Interventional segment generated $1.357 billion at 7.3% reported and 5.3% FXN growth, with broad mid-single-digit performance across the three sub-segments. Peripheral Intervention delivered $515 million at 7.1% reported and 4.0% FXN growth. Urology and Critical Care delivered $430 million at 7.5% reported and 6.5% FXN growth. Surgery delivered $411 million at 7.4% reported and 5.5% FXN growth. CEO Tom Polen described execution as broad-based, with more than 90% of the business delivering mid-single-digit growth, and BD raised its full-year fiscal 2026 adjusted diluted EPS guidance to $12.52 to $12.72 from the prior $12.35 to $12.65 range while reaffirming the revenue growth outlook. The Q2 results follow the February 9, 2026 completion of BD’s separation of the Biosciences and Diagnostic Solutions business through combination with Waters Corporation, and the continuing-operations BD that posted the 7.3% Interventional growth is now a more focused MedSurg-Interventional company. The Q2 numbers are the first clean quarterly read on the residual portfolio after the spin. Recent product launches feeding the Interventional growth line include the BD CentroVena One central venous catheter insertion system, the BD Pyxis Pro automated dispensing solution in Europe, the Revello vascular covered stent (CE marked), and the Liverty TIPS stent graft (CE marked). The portfolio is built around recurring access, consumable, and instrument products that run against an installed clinical environment, and the residual MedSurg-Interventional profile is the one that surgical robotics and interventional founders should be reading for the durable demand picture in 2026. ## Intuitive Surgical Lifts 2026 da Vinci Procedure Growth Guidance to 13.5%-15.5% Intuitive Surgical reported Q1 2026 on April 22 with revenue of $2.77 billion, up 23% year over year, and net income of $822 million. Da Vinci procedures grew approximately 16% globally, with U.S. procedures up 14% and international procedures up 19%. The company placed 431 da Vinci systems in Q1 versus 367 a year earlier, and 232 of the Q1 placements were da Vinci 5 systems against 147 in the prior-year quarter. CEO Dave Rosa attributed the momentum to a 31% rise in after-hours procedures and higher overall utilization, with general surgery and newer procedures including appendectomy driving the U.S. growth. Intuitive lifted its 2026 worldwide da Vinci procedure growth guidance to 13.5% to 15.5% from the prior 13% to 15% range, and improved the non-GAAP gross profit margin guidance to 67.5% to 68.5%. The da Vinci installed base reached 11,395 systems as of March 31, up 12% year over year. The structural read is that the Q1 procedure growth across the 11,395-system installed base translates into a recurring instrument and accessory revenue base that delivered roughly 60% of Intuitive’s revenue in 2026, with capital sales contributing about 23% and services contributing the balance. The capital-sale layer where Medtronic Hugo, J&J Ottava, and CMR Versius Plus are now competing produces a fraction of the downstream economic flow that the cleared platforms generate over their installed lives. For the founders building component, instrument, and software platforms that fit into the cleared system architectures, the Intuitive recurring-revenue profile is the operating template for what durable medtech economics look like at scale. ## CMR Surgical Versius Plus Continues the Limited US Rollout CMR Surgical received FDA clearance for the Versius Plus second-generation soft-tissue robotic platform in December 2025 for cholecystectomy, with additional 510(k) submissions pending across general surgery, gynecology, urology, and gastrointestinal procedures. The company introduced the cleared platform to U.S. surgeons at the SAGES 2026 conference in Tampa in late March, and CEO Massimiliano Colella has described the U.S. debut as a watershed moment that follows more than 45,000 procedures conducted globally with the Versius platforms across 30-plus countries. The U.S. soft launch is concentrated on selective hospital and ambulatory surgery center partnerships during 2026, with broader expansion contingent on the additional indication clearances landing through the second half of the year. The Versius Plus rollout, the J&J Ottava De Novo filing on the FORTE pivotal package presented at ASMBS on May 5, and the Medtronic Hugo measured U.S. expansion together produce three credible cleared or near-cleared challengers to the da Vinci platform on the soft-tissue side. The competitive dynamic at the capital-sale layer is now the most active it has been in two decades. The instrument and consumable layer underneath each cleared platform is where the durable margin sits, and the founders building component, instrument, and software platforms that fit into the cleared system architectures are the ones positioned for the second-half 2026 strategic conversation. ## Restore Robotics Expands the da Vinci Instrument Remanufacturing Footprint Restore Robotics received FDA 510(k) clearance in early April for two additional da Vinci Xi instruments, the permanent cautery hook and the permanent cautery spatula, bringing the total cleared remanufactured da Vinci instrument portfolio to four products. The clearance expands the cost-savings alternative the company offers to hospitals running da Vinci Xi systems, and the structural pressure on the Intuitive instrument-replacement revenue model has measurable margin implications across the installed base over time. Intuitive’s installed base continues to absorb the remanufacturing competition, and the company has continued to outperform Wall Street’s forecasts on the recurring-revenue line through Q1 2026. The Restore Robotics expansion is a structural reminder that the recurring-instrument layer is not a uncontested margin pool. Founders engineering the recurring-revenue economics into a cleared system architecture have to design for the remanufacturing pressure that will eventually arrive against any commercially successful instrument portfolio, with the design-for-controlled-reuse and supply-chain-control levers built into the architecture from initial product design. The platforms that build the recurring-revenue layer with that structural foresight protect the margin against the remanufacturing competition that the Restore Robotics clearance preview makes inevitable. ## What the Numbers Tell Surgical Robotics and Interventional Founders The four developments together describe a market where the surgical robotics platform competition concentrates at the capital-sale layer, and the durable economics for interventional and surgical device founders sit in the consumable, instrument, and recurring-revenue layer that BD’s 7.3% Interventional growth just confirmed is healthy across the entire MedSurg-Interventional portfolio. Founders building component or instrument platforms that fit into the cleared surgical robotic system architectures, founders building peripheral interventional, urology, and surgical instrument products in adjacencies BD just delivered mid-single-digit growth across, and founders thinking about the strategic conversation in the second half of 2026 should weight their commercial planning toward the recurring-revenue layer rather than the system sale. The valuation conversations for platforms that arrive at the strategic discussion with a credible recurring-revenue economic model run on different multiples than the conversations for platforms whose economic story rests on capital placement. Dave’s take BD’s Interventional Q2 number is the data point I will be using with the surgical robotics and interventional founders I work with this quarter. The platform sale is where the venture press concentrates, and the durable medtech economics run through the instrument and consumable layer that BD’s 7.3% growth just confirmed is healthy across surgery, urology, and peripheral intervention. Build your commercial model around the recurring-revenue layer from initial product architecture, with the instrument and consumable economics engineered into the platform before the first system ships. From Dave’s video library Dave on the five numbers a business actually runs on. BD's consumables growth is a reminder that the number carrying durable medtech economics is rarely the one in the headline. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to The Build newsletter: [The Build →](/field-notes/the-build) --- # AI-Embodied Surgical Robots Force a Regulatory Framework Reset | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-05-08-ai-embodied-surgical-robots-regulatory-framework-reset [← All Field Notes](/field-notes) May 8, 2026 Industry Roundup A team led by Prof. Prokar Dasgupta of King’s College London published a Frontiers in Science article on May 7 arguing that AI-embodied surgical robots can revolutionize surgery if regulators rebuild the framework around adaptive systems that continue learning after approval. The thesis lands the same week that Medtronic surgeons completed first U.S. cases on the Stealth AXiS Autopilot system, which integrates AI-powered planning, real-time segmental tracking, navigation, and robotics into a single spine surgery platform, and Neptune Medical reported zero adverse events and 100% cecal intubation in the 50-patient first-in-human CARE 1 study of the Triton robotic endoscopy system at Digestive Disease Week 2026. Read across the three developments, the question is no longer whether AI is moving into the cleared surgical robotics architecture, the question is whether the regulatory framework keeps pace with what the systems already do at the operating-room level. For the founders building the next generation of advanced surgical robotics platforms, the regulatory architecture decision is now load-bearing on the timeline of the cleared platform. ## Frontiers in Science: AI-Embodied Surgical Robots and the Adaptive-System Gap Frontiers in Science published an article on May 7 by a team led by Prof. Prokar Dasgupta and Dr. Alejandro Granados of King’s College London arguing that AI-embodied surgical robots can revolutionize surgical practice through sensor-equipped operating rooms that enable spatial understanding, adaptive learning, real-time intra-operative guidance, and predictive outcome visualization, and that the regulatory framework currently built around static device approvals does not match the systems the field is now putting into production. The authors highlight that AI’s ability to learn presents a regulatory puzzle because devices continue to evolve after approval, and that the licensing pathways, device classifications, post-market monitoring requirements, and compliance standards need to be updated for adaptive systems that no longer fit the static-device framework. The team also flagged dataset bias reinforcing healthcare inequalities, research and industry concentration in resource-rich nations, liability clarification in clinical practice, and the integration of autonomous systems within surgical teams as the structural questions that the field has to answer. The recommended framework is direct. Regulatory reform on licensing pathways and post-market monitoring for adaptive systems, standardized clinical trial metrics that evaluate AI software and human-AI interactions together, collaborative training models that include lower-income healthcare systems, and a hard line that the human surgeon remains the chief decision-maker with AI providing decision support rather than autonomous control. The framing matters for surgical robotics founders building the next generation of cleared platforms because the regulatory pathway architecture is now a strategic decision that has to be made years before the system clears. The platforms designed for adaptive AI capability against a static-device regulatory framework will arrive at the clearance gate with a structural mismatch that the static framework cannot easily resolve, and the founders who plan the regulatory architecture against the Frontiers framing are the ones positioned for the second-half 2026 strategic conversation. ## Medtronic Stealth AXiS Autopilot: Integrated AI-Powered Surgery in Production Surgeons Christopher Good, Colin Haines, and Ehsan Jazini performed the first U.S. surgical cases on the FDA-cleared Medtronic Stealth AXiS Autopilot system at HCA Virginia’s Reston Hospital Center in late April, with the milestone procedure including a single-level direct lateral interbody fusion followed by a three-level posterior instrumented spinal fusion with laminectomy and bilateral SI joint fusion. Medtronic announced on April 29 that the first clinical cases on the broader Stealth AXiS surgical system were complete across multiple sites following February and March FDA clearances for spine, cranial, and ENT indications, and the company secured CE Mark on April 28 for European expansion. The Stealth AXiS Autopilot platform combines AI-powered surgical planning, segmental tracking that follows each vertebra continuously through the procedure, navigation, and robotic delivery into a single integrated platform. The structural read on Stealth AXiS Autopilot is that the platform is the cleanest current example of an AI-embodied surgical robotic system already cleared and operating against the existing FDA framework. The segmental tracking and AI-powered planning components are the elements the Frontiers team described as adaptive AI in the operating room, and the platform’s integration of those components into a single cleared-as-static-device architecture is what surgical robotics founders should study as the operational template for clearing AI-embodied systems against the framework as it actually exists today. The platforms that arrive at clearance with AI-embodied architecture compatible with the static-device framework get to the commercial gate first. The platforms that depend on adaptive post-clearance learning have to wait for the framework reform the Frontiers team called for, and that reform timeline is structurally uncertain. ## Neptune Medical Triton: First-in-Human Robotic Endoscopy and the AI-Ready Architecture Neptune Medical announced on May 5 that the 50-patient first-in-human CARE 1 study of the Triton robotic endoscopy system met both primary endpoints with no adverse events at 14-day follow-up and 100% cecal intubation across both phases of the study at H-T Centrum Medyczne in Tychy, Poland. The system also delivered a 67.5% polyp detection rate, a 54.2% adenoma detection rate, and 100% polypectomy success on polyps under 2 cm. Triton is designed for diagnostic colonoscopies, endoscopic mucosal resections, and endoscopic submucosal dissections, with results presented in a late-breaking oral session at Digestive Disease Week 2026 by Dr. Jason Samarasena of UC Irvine. The system is not yet authorized for sale in the United States or international markets, and Neptune raised $97 million in 2024 Series D funding led by Sonder Capital with Olympus participation. The company spun out the Jupiter Endovascular subsidiary to focus the operational discipline on the GI robotic platform. The CARE 1 first-in-human results are clean against the static-device framework. The near-term clinical impact is consistency, ergonomics, and procedural control in colonoscopy, and the longer-term opportunity flagged by the analysts following the readout is advanced therapeutic endoscopy with AI-enabled visualization layered onto the robotic scope-control platform. For surgical robotics founders watching the Triton trajectory, the operational lesson is that the Neptune architecture clears against the existing framework as a robotic scope-control platform, and the AI capability that the longer-term opportunity depends on layers onto the cleared platform later, in an order the static-device framework can absorb. The architecture decision to clear the robotic platform first and add the AI-embodied capability against the framework as it evolves is the same operational pattern Medtronic used with Stealth AXiS Autopilot, and it is the pattern the founders building the next generation of GI, cardiac, and interventional robotic platforms should be studying. ## The Cleared Static-Device Stack and the Adaptive AI Frontier The three developments together describe a market where the cleared surgical robotics platforms are absorbing more AI capability against a regulatory framework still organized around static devices, and the founders building the next generation of advanced surgical robotic platforms have to make the architecture decision between clearing the static-device version of the platform first and waiting on the adaptive-AI regulatory framework reform the Frontiers team called for. The Stealth AXiS Autopilot pattern shows that AI-powered planning, segmental tracking, navigation, and robotic delivery clear together as an integrated static-device platform when the architecture is designed for the framework as it actually exists. The Triton pattern shows that a robotic platform can clear first and absorb additional AI capability later as the framework evolves. The Frontiers team’s framing makes clear that the broader adaptive-AI surgical robotic frontier still requires regulatory reform, and that reform is structurally uncertain on timeline. The strategic read for founders is operational. Design the platform’s AI capability against the framework as it actually exists at the clearance gate, not against the framework you wish existed when the platform reaches clinical scale. The cleared platform that delivers integrated AI-powered planning, tracking, and navigation against the static-device framework gets to commercial scale first. The platform that depends on the framework reforming before clearance has to absorb the regulatory uncertainty into the operating timeline, and the strategic-acquirer evaluation in the second half of 2026 will price the regulatory architecture decision into the platform multiple. The recurring instrument and consumable revenue layer underneath every cleared platform, the BD Q2 Interventional 7.3% growth pattern, the Intuitive Surgical 11,395-system installed base, and the J&J Ottava and CMR Versius Plus expansion arcs all run against cleared static-device architectures that absorb AI capability incrementally rather than through framework reform. Dave’s take The Frontiers paper describes the regulatory frontier the surgical robotics field is heading toward, and the Stealth AXiS Autopilot and Triton clearance patterns show what the framework can already absorb when the architecture is designed for the framework as it exists. I will be using this contrast with the founders I am working with this quarter. Plan the AI capability into the cleared platform against the static-device framework, and design the architecture to absorb additional AI capability incrementally as the framework evolves. The platforms that bet the company on framework reform are the ones that arrive at the strategic conversation with regulatory uncertainty the buyer cannot price. From Dave’s video library Dave on AI systems that already make consequential decisions and why the override matters more than the intelligence. Adaptive surgical robots that keep learning after approval are the sharpest version of that problem regulators have faced yet. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to The Build newsletter: [The Build →](/field-notes/the-build) --- # Interventional Medtech Q1 2026 Earnings Reveal a Margin Tax Beneath the Topline | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-05-09-interventional-medtech-q1-margin-tariff-compression [← All Field Notes](/field-notes) May 9, 2026 Industry Roundup Teleflex reported Q1 2026 on May 7 with revenue up 32.3% on a GAAP basis to $548.3 million, the Interventional segment up 3% to $204.7 million, and adjusted gross margin down 470 basis points to 61.4% on tariffs, quality remediation charges, higher logistics costs, and product mix from the recently absorbed Vascular Intervention acquisition. The 2026 guidance carries a $33 million tariff headwind. Penumbra reported $374.8 million on May 6, up 15.6%, with the Boston Scientific merger close pending and no full-year guidance issued. Profound Medical and Sight Sciences both posted Q1 results that lift FY26 expectations on the interventional MRI and interventional eye care platforms. Read across the four reports, the picture is consistent. Interventional medtech topline growth is real, and the margin tax underneath the topline that founders building the next generation of advanced interventional platforms have to plan for is now a structural feature of the segment. ## Teleflex Q1: 32% Topline, 470 Basis Points of Margin Compression Teleflex reported first quarter revenue of $548.3 million on May 7, an increase of 32.3% on a GAAP basis and 5.1% on a pro forma adjusted constant currency basis year over year. The Interventional segment generated $204.7 million in revenue, a 3% rise year over year, supported by intraosseous, right heart, and complex catheters. The headline growth carries a substantial cost line underneath. Adjusted gross margin declined 470 basis points to 61.4% year over year, with management attributing the compression to tariffs, quality remediation charges, higher logistics costs, and product mix from the Vascular Intervention acquisition. Adjusted operating margin fell 510 basis points, pressured by the gross margin decline and operating expenses linked to the recent acquisition and higher R&D spend. Adjusted EPS came in at $1.39, a 3.5% decrease year over year. Tariffs alone represent a $33 million headwind in the 2026 guidance, with potential refunds for 2025 or 2026 payments treated as upside rather than baseline. For founders building advanced interventional platforms, the Teleflex Q1 read is the cleanest current data point on what the integration economics actually look like when the topline arrives via M&A and the cost lines underneath the integration absorb tariffs, quality remediation, and logistics simultaneously. The Vascular Intervention acquisition delivered the topline that produced the 32.3% GAAP growth, and the cost lines underneath the acquired business produced the margin compression that put adjusted EPS down year over year. The structural read is that interventional M&A on the Teleflex scale absorbs a multi-quarter integration tax even when the strategic logic of the deal is sound, and the 2026 guidance now has to absorb the tariff line on top of the integration line. Founders preparing for a strategic conversation with the Teleflex-class acquirers in the second half of 2026 should price the integration math the buyer is actually running, not the topline math that gets quoted in the press release. ## Penumbra Q1: 15.6% Growth Into the Boston Scientific Close Penumbra reported Q1 2026 revenue of $374.8 million on May 6, up 15.6% year over year, with thrombectomy revenue of $253.9 million up 12.1% and embolization and access revenue of $120.8 million up 23.8%. Gross margin held at 67.6%. Operating expenses were $215.2 million, including $9.4 million of acquisition-related expenses tied to the pending Boston Scientific transaction. The company is not providing full-year 2026 guidance and will not host an earnings call, given the pending Boston Scientific acquisition at $374 per share, valuing the transaction at approximately $14.5 billion with consideration mix targeted at roughly 73% cash and 27% Boston Scientific stock, subject to shareholder and regulatory approvals. The Penumbra Q1 print confirms what the Boston Scientific strategic logic depends on. The thrombectomy and embolization franchises are still growing through the merger announcement window, the gross margin profile is intact, and the only material new cost line on the income statement is the acquisition-related expense tied to closing the deal. The structural read for founders is that the strategic acquirer is paying $14.5 billion for a clean platform that delivers double-digit growth across two segments with a 67.6% gross margin, and the platform is being absorbed without margin damage to the underlying franchise. The platforms that arrive at the strategic conversation with the Penumbra-class operating profile get the Penumbra-class multiple. The platforms that arrive with margin damage from integration absorption get priced against the Teleflex-class profile instead. ## Profound Medical and Sight Sciences: Specialist Interventional Q1 Strength Profound Medical reported Q1 2026 results on May 7, with the company describing triple-digit revenue growth combined with strong gross margin and lower operating expenses, and projecting full-year 2026 revenue of approximately $25 million, representing 56% growth over the prior year. The TULSA-PRO interventional MRI platform anchors the growth profile, and the operating expense discipline produced gross margin expansion against the prior period. Sight Sciences reported Q1 revenue of $19.7 million on May 7, up 13% year over year, with gross margin at 86% and growth across both interventional segments. The company raised full-year 2026 revenue guidance to $83 million to $89 million on the strength of the reimbursed dry eye business and steady glaucoma execution. The two specialist interventional reports describe the operating posture that produces multiple-expansion outcomes in the current environment. Profound Medical is delivering category-defining growth on a focused interventional MRI platform with operating expense discipline that compounds through the year. Sight Sciences is delivering 13% growth at 86% gross margin with raised full-year guidance against a reimbursement structure that the operating plan has worked through. Neither company is absorbing the integration tax that compressed Teleflex’s adjusted gross margin by 470 basis points, and the operating profile each is producing matches the cleared-platform commercial template strategic acquirers are paying premiums for in the second-half 2026 conversation. The operating discipline that protects gross margin through the growth phase is the discipline that produces the multiple at the strategic conversation. ## What the Q1 Reports Mean for the Founder Operating Plan Read together, the four Q1 reports describe the structural choice every advanced interventional founder is now making in real time. Teleflex grew 32.3% on M&A absorption and lost 470 basis points of adjusted gross margin underneath the topline. Penumbra grew 15.6% organically into a $14.5 billion strategic acquisition with no margin damage. Profound Medical and Sight Sciences are growing on specialist platforms with gross margin discipline that lifts the operating profile through the year. The strategic acquirers in the second half of 2026 are reading these four operating profiles side by side and pricing the platforms that protect gross margin through growth at premiums to the platforms that absorb the integration and tariff tax through the operating line. For founders building the next generation of advanced interventional platforms, the operating plan question is how the gross margin profile and the cost line architecture will look at the moment the platform reaches commercial scale. The platforms that produce category-defining topline growth on cost structures that absorb tariffs, integration, and logistics through the gross margin line arrive at the strategic conversation at the Teleflex-class multiple. The platforms that produce double-digit growth on gross margins above 65% with operating expense discipline that compounds through the year arrive at the Penumbra-class or Sight Sciences-class multiple. The Q1 2026 cycle is the cleanest current evidence the segment has produced of which operating profiles the strategic-acquirer evaluation is actually pricing, and the founders running operating plans against those profiles now are the ones positioning the company for the second-half 2026 strategic conversation. Dave’s take Topline growth covers a lot of operating sins until the moment a strategic buyer reads the gross margin line. The Teleflex Q1 print shows what happens when integration, tariffs, and quality remediation hit the cost structure at the same time the acquired topline is producing the GAAP growth headline. The founders I am working with this quarter are running the same exercise on their own operating plans. Where does the gross margin profile actually sit at commercial scale, and which cost lines are absorbing the visible growth before it reaches operating income? From Dave’s video library Dave on turning messy records into one honest number. Teleflex just showed why it matters: 32 percent topline growth, and the number a strategic buyer actually reads is the gross margin line underneath it. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to The Build newsletter: [The Build →](/field-notes/the-build) --- # SS Innovations Perth-to-India Telesurgery Reframes Surgical Robotics Platform Economics | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-05-10-telesurgery-platform-economics-surgical-robotics [← All Field Notes](/field-notes) May 10, 2026 Industry Roundup Dr. Mohit Bhandari sat down at a portable robotic console in Perth, Australia on May 2 and performed a gastrojejunostomy on a patient in Indore, India, more than 4,500 miles away, using the SSi Mantra surgical robotic system and the SSi MantrAsana chair-based tele-surgeon console. The demonstration was staged at the Royal Australasian College of Surgeons Annual Scientific Congress and disclosed publicly on May 7. Three days earlier, on May 5, SS Innovations announced its installed base had crossed 10,500 cumulative procedures, including more than 500 cardiac operations, roughly 170 telesurgeries, and 125 pediatric cases, across more than 200 systems deployed in 11 countries since the system’s August 2022 commercial launch. On May 8, on a different layer of the interventional stack, BioCardia disclosed FDA alignment on two clearance pathways for its Helix Transendocardial Delivery Catheter following a Pre-Submission meeting. Read across the three disclosures, the structural lesson for surgical robotics and advanced interventional founders is consistent. Platform economics in this segment now run on the integrated stack of system, console, ultra-low-latency networking, and clinical evidence at commercial volume, and the regulatory posture for the cleared product is a function of how that stack was architected from the start. ## The 4,500-Mile Telesurgery From Perth to Indore Dr. Mohit Bhandari, President of IRCAD India, performed a robotic gastrojejunostomy on May 2 from Perth, Australia on a patient located more than 4,500 miles away in Indore, India, using the SSi Mantra surgical robotic system at the patient site and the SSi MantrAsana portable tele-surgeon console at the surgeon site. The procedure was staged at the Royal Australasian College of Surgeons Annual Scientific Congress and announced publicly on May 7. The MantrAsana is a chair-based, self-contained tele-surgeon console engineered for portability and a small physical footprint, designed to enable telesurgery from physician offices, satellite clinics, and other non-traditional surgical sites. SS Innovations characterized the console as engineered for real-time remote procedures through ultra-low-latency connectivity, high-definition visualization, and precise instrument control. For surgical robotics founders, the structural read on the Perth-to-Indore demonstration is that the unit of competition for the next phase of platform economics is not the operating-room robot in isolation. The unit of competition is the integrated stack of patient-side robot, surgeon-side console (including portable form factors), network engineering for sustained ultra-low-latency operation across continents, regulatory authorization in both the patient and the surgeon jurisdictions, and the clinical evidence base at procedure volumes that demonstrate the stack works at scale rather than only at staged conferences. The companies that finish in this segment are the ones that designed each layer of that stack to interoperate from initial product architecture, not the ones that built the robot first and treated the network, console, and cross-border regulatory layers as integrations to engineer later. ## 10,500 Procedures and the Commercial-Scale Milestone On May 5 SS Innovations announced its installed base had completed more than 10,500 cumulative procedures using the SSi Mantra system, including more than 500 cardiac operations, roughly 170 telesurgeries, and 125 pediatric surgeries. The company has trained approximately 2,100 physicians and deployed more than 200 systems across 11 countries since the SSi Mantra commercial launch in August 2022. CEO Dr. Sudhir Srivastava framed the milestone as a function of the system’s differentiated technical features, training accessibility, and cost profile, and emphasized the company’s mission to expand access to robotic surgical care globally. The structural lesson in the 10,500-procedure number for surgical robotics founders is operational, and it sits underneath the headline. Reaching commercial volume that produces the procedure data the strategic-acquirer evaluation actually prices requires the installed base, the surgeon training pipeline, and the geographic deployment to compound together over the years between initial commercial launch and the first strategic conversation. SS Innovations spent the four years from August 2022 building out the deployment footprint, the training pipeline, and the procedure variety the installed base now produces, and the telesurgery milestone last week is a clinical disclosure the operational footprint underneath was already in place to support. Founders building toward a comparable milestone on a comparable timeline are running the operational build now, not after the cleared platform reaches its first commercial customers. ## BioCardia Helix Pathway Alignment and the Interventional Regulatory Posture On May 8 BioCardia disclosed the outcome of a Pre-Submission meeting with FDA on its Helix Transendocardial Delivery Catheter and reported that the agency raised no concerns on Helix safety data, device performance, or compatibility with general classes of agents. The agency aligned with BioCardia on two pathways to clearance for Helix. The first is a simultaneous approval pathway running with the CardiAMP cell therapy system, which BioCardia identified as the agency’s preferred route. The second is a De Novo pathway following a follow-on Pre-Submission meeting. The Helix catheter is an enabling delivery platform for minimally invasive, targeted intramyocardial delivery of therapeutic and diagnostic agents directly to heart tissue, with a helical needle architecture engineered to stabilize within the beating heart and to reach regions other delivery modalities cannot access reliably. For advanced interventional founders, the BioCardia disclosure is the cleanest current example in the public record of what regulatory alignment looks like when the platform was architected as the delivery layer underneath a therapeutic stack rather than as a standalone catheter. The FDA aligned on two clearance pathways, raised no concerns on the safety or performance data, and identified the simultaneous approval pathway with the therapeutic system as the preferred route, which is the regulatory posture a delivery platform earns when the safety and performance evidence base was generated in the studies the therapeutic agents were already running. The platforms that arrive at the Pre-Submission meeting with the agency’s preferred pathway already on the table are the platforms whose evidence architecture and regulatory strategy were designed alongside the engineering work, not after. ## What the Telesurgery Milestone Means for the Founder Operating Plan Read together, the three disclosures describe the operating discipline that produces both the clinical milestone and the regulatory posture that strategic acquirers and clinical adopters in 2026 are pricing. SS Innovations produced a 4,500-mile cross-border telesurgery on a system architecture that designed the patient-side robot, the portable surgeon-side console, the ultra-low-latency network engineering, and the cross-border deployment footprint as one integrated stack, on an installed base that compounded to 10,500 procedures over four commercial years. BioCardia produced FDA alignment on the agency’s preferred clearance pathway on a Pre-Submission meeting where the safety and performance evidence already supported the simultaneous approval route. Both outcomes depended on architectural decisions made years earlier in the product, evidence, and regulatory strategy, not on integrations engineered after the platform was already in commercial use. For founders building advanced interventional and surgical robotics platforms in 2026, the operating plan question is which architectural decisions need to be made now to produce a comparable milestone in three to four years. The companies that finish in this environment design the integrated platform stack from initial product architecture, fund the operational build of the installed base and the training pipeline alongside the engineering work, and run the regulatory and evidence strategy to land the agency’s preferred pathway at the first Pre-Submission meeting rather than the second or third. The work is harder during the build phase because the architectural decisions compete for time and capital with the visible engineering progress that produces the next financing round. The compensation arrives at the moment a clinical milestone, a regulatory milestone, or a strategic conversation depends on the architecture, and the architecture has to already be in place for the milestone to land. Dave’s take The Perth-to-Indore demonstration is a clinical disclosure on a network and platform architecture that was designed years earlier to operate at intercontinental distance with ultra-low-latency control. The 10,500-procedure number is an operational disclosure on a deployment and training footprint that compounded over four commercial years. The BioCardia FDA alignment is a regulatory disclosure on an evidence architecture that was designed to land the agency’s preferred pathway. Each of the three is the visible result of architectural work the company did in the build years before the milestone landed. The founders I am working with this quarter are running the same exercise on their own platforms. Which architectural decisions sitting in front of you now produce the milestone three years from today, and which ones produce a slower path to the same outcome? From Dave’s video library Dave on why the unscalable, compounding work matters most right after early wins. A telesurgery demo across 4,500 miles reads as a moment, but it sits on four years of deployment and training work that never made headlines. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to The Build newsletter: [The Build →](/field-notes/the-build) --- # Neptune Triton GI Robot Validates the Specialist Surgical Robotics Path | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-05-11-neptune-triton-gi-robot-specialist-surgical-robotics-path [← All Field Notes](/field-notes) May 11, 2026 Industry Roundup Neptune Medical disclosed on May 6 that the 50-patient CARE 1 first-in-human study of its Triton robotic endoscopy system hit both primary endpoints with zero adverse events and 100% cecal intubation, with the data presented at DDW 2026 and the platform now positioned for the regulatory submission phase. On the same day, the FDA issued a Letter to Health Care Providers warning that a shortage of neurosurgical patties, sponges, and strips traceable to a Medline endotoxin remediation will continue through year-end. The day after, on May 7, Roche announced a definitive merger agreement to acquire PathAI in a deal worth up to $1.05 billion, with $750 million upfront and up to $300 million in milestones, building on a five-year partnership and consolidating Roche’s position in AI-enabled digital pathology. Read across the three disclosures, the structural lesson for surgical robotics and advanced interventional founders is consistent. Specialist platforms that finish are the ones whose first-in-human evidence is clean enough to anchor the regulatory submission, whose supply architecture survives a single-supplier disruption, and whose AI and software layers were designed for a strategic-acquirer evaluation against an integrated diagnostic stack. ## Neptune Triton CARE 1 and the Specialist Surgical Robotics Path Neptune Medical reported on May 6 that the CARE 1 first-in-human study of its Triton robotic endoscopy system met both primary endpoints, per MedTech Dive. The 50-patient single-arm study at a single site in Poland enrolled adults undergoing screening, surveillance, or diagnostic colonoscopy and followed each patient for 14 days post-procedure. The study reported zero adverse events related to the device and a 100% cecal intubation rate. Secondary endpoints included a 67.5% polyp detection rate and a 54.2% adenoma detection rate, both within the range supported by current clinical guidance for screening colonoscopy. The Triton platform is engineered to support diagnostic colonoscopy, endoscopic mucosal resection, and endoscopic submucosal dissection in the lower gastrointestinal tract, and the company has positioned the system to address physician fatigue and procedure throughput in the screening environment. Neptune Medical, based in Burlingame, raised $97 million in a Series D round in 2024 backed by Sonder Capital and Olympus. The structural read on the CARE 1 result for surgical robotics founders is that the specialist platform path is now producing clean first-in-human evidence on the same regulatory and clinical timelines as the integrated multi-specialty platforms. Neptune Triton is engineered specifically for the lower GI tract, designed against the procedural variety the gastroenterology specialty actually runs, and architected against the screening colonoscopy throughput pressure the U.S. health system is now under. The 100% cecal intubation rate and zero device-related adverse events are the cleanest first-in-human disclosure profile a specialist surgical robotics platform can produce, and the data lands the platform in the same regulatory submission readiness position the multi-specialty platforms reached through pivotal trials that took years longer to read out. Founders building specialist platforms in the surgical robotics segment now have a current public reference point for what the clean first-in-human disclosure profile looks like, and the platforms whose first-in-human evidence falls short of that profile will struggle to compete for the same regulatory pathway and strategic-acquirer attention. ## The FDA Neurosurgical Supply Shortage and the Single-Supplier Risk On May 6 the FDA issued a Letter to Health Care Providers warning of a shortage of neurosurgical patties, sponges, and strips that the agency expects to continue through the end of 2026, per MedTech Dive. The trigger was a March recall by Medline of neurosurgical sponges with higher-than-expected endotoxin levels, which removed a major share of the U.S. supply of the disposable products neurosurgeons use to absorb fluid and protect tissue during cranial and spinal procedures. The FDA recommended health care providers conserve use, open packages only when needed because unused devices cannot be returned to inventory, diversify supply sources, and report any patient impact through the agency’s shortage reporting channel. The disruption affects an estimated portion of cases across neurosurgery, microsurgery, and head-and-neck procedures and runs concurrently with the elective procedure backlog the U.S. health system is working through this quarter. The structural read for surgical robotics and advanced interventional founders is that supply architecture is now a board-level risk for any platform whose procedural use depends on disposables sourced from a small number of suppliers. The Medline endotoxin event was a single-supplier quality remediation that removed the disposable input across the neurosurgical procedure base for an entire planning year, and the platforms running neurosurgical procedures on integrated robotics or navigation systems lose procedure volume on every case the disposable shortage takes off the schedule. Founders building specialist surgical robotics platforms in neurosurgery, spine, ENT, or other procedural specialties with concentrated disposable supply now have to map the supply architecture against a credible single-supplier disruption scenario before the disruption arrives, and the platforms whose disposable supply, regulatory pathway, and customer commitments are concentrated in one supplier face a procedure-volume risk the integrated platform thesis has not been pricing. ## Roche-PathAI and the AI-Pathology Consolidation Layer Roche announced on May 7 that it entered into a definitive merger agreement to acquire PathAI, the Boston-based AI digital pathology company, in a transaction worth up to $1.05 billion, per MedTech Dive. The structure is $750 million upfront and up to $300 million in additional milestone payments, with the deal expected to close in the second half of 2026 subject to antitrust and regulatory approvals. PathAI will be integrated into Roche’s Diagnostics division. The transaction builds on a 2021 partnership between the two companies, expanded in 2024 to include co-development of AI-enabled companion diagnostic algorithms. Roche identified PathAI’s Image Management System and AI workflow capabilities as complementary to its digital pathology portfolio and identified the combined capability as a path to biomarker discovery, companion diagnostics, and translational research at scale. The deal positions Roche as the dominant integrated AI pathology stack, and industry analysts are watching whether the move triggers consolidation among remaining AI-pathology firms. For surgical robotics and advanced interventional founders, the Roche-PathAI deal is the cleanest current example of what the AI-enabled software layer is worth to the strategic acquirer when the AI capability has been integrated with the diagnostic and companion-therapy stack across a five-year partnership. The price Roche paid is the integrated-stack premium, not the standalone AI software premium, and the difference compounds across the surgical robotics, advanced interventional, and AI-medical-device segments as integrated platforms consolidate and standalone AI tools face downward pricing pressure. The signal for founders building AI-enabled surgical robotics or interventional platforms is that the strategic-acquirer evaluation prices the integrated stack of regulatory clearance, clinical evidence, workflow integration, and operating partnership history substantially above the price the standalone AI capability commands. Founders running AI-medical platforms on standalone roadmaps now need to think clearly about which strategic integration partnership they are designing the platform to anchor and how to start that partnership operating cadence in the build years before the strategic conversation. ## What the Three Disclosures Mean for the Founder Operating Plan Read together, the three disclosures describe the operating discipline that produces both the regulatory readiness and the strategic-acquirer multiple the surgical robotics, advanced interventional, and AI-medical-device segments are now pricing. Neptune Triton produced a clean first-in-human evidence base in a specialist procedural segment, on a 50-patient study designed against the primary endpoints the regulatory pathway and the strategic-acquirer evaluation actually read. The FDA neurosurgical shortage is a credible single-supplier disruption running through the rest of the planning year on disposable inputs concentrated in one supplier, and the surgical robotics platforms that scale through procedural specialties with concentrated disposable supply have to map the supply architecture against a comparable disruption scenario now. The Roche-PathAI deal is a $1.05 billion strategic-acquirer evaluation that priced the integrated AI-diagnostics stack at the multiple a five-year operating partnership produced, and the AI-medical platforms that arrive at the strategic conversation without a comparable integration history price at a different multiple. For surgical robotics and advanced interventional founders building specialist platforms in 2026, the operating plan question is which architectural and partnership decisions need to be made now to produce a comparable regulatory readiness, supply architecture, and strategic-integration profile in three years. The companies that finish in this environment design the first-in-human evidence base against the primary endpoints the regulatory pathway and the strategic-acquirer evaluation actually price, map the supply architecture against credible single-supplier disruption scenarios across the disposable inputs the platform depends on, and start the strategic-integration operating cadence with the prospective acquirer in the build years before the strategic conversation arrives. The work is harder during the build phase because each of the three lines competes for time and capital with the visible product engineering that produces the next financing round. The compensation arrives at the moment a regulatory milestone, an operating risk, or a strategic conversation depends on the architectural and partnership work, and the work has to already be in place for the milestone, the resilience, or the multiple to land. Dave’s take Neptune Triton just produced the cleanest first-in-human disclosure profile a specialist surgical robotics platform can produce, and it landed because the study was designed against the primary endpoints the regulatory pathway and the strategic-acquirer evaluation actually price. The FDA neurosurgical shortage shows what supply concentration costs when a single-supplier quality event removes a disposable input across an entire procedural specialty. The Roche-PathAI deal priced the integrated AI-diagnostics stack at $1.05 billion on a five-year operating partnership, not on a standalone AI capability. The three disclosures together describe the architectural and partnership discipline I am running with the founders I work with this quarter. Which evidence, supply, and strategic-integration decisions sitting in front of you now produce the readiness, the resilience, and the multiple three years from today? From Dave’s video library Dave on the difference between encouragement and proof. Neptune's first-in-human study worked because it was designed to produce the evidence acquirers and regulators actually price, not the evidence that flatters the technology. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to The Build newsletter: [The Build →](/field-notes/the-build) --- # IVL Platform Competition Splits Into Three Mechanisms for Calcium Modification | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-05-12-ivl-platform-competition-calcium-modification-2026 [← All Field Notes](/field-notes) May 12, 2026 Industry Roundup Stryker announced on May 7 that it had closed its acquisition of Amplitude Vascular Systems, adding the Pulse intravascular lithotripsy platform to a peripheral vascular portfolio that already includes the Inari Medical clot-removal business acquired last year. On May 11, a first-in-human report on the Boston Scientific Seismiq 4CE Coronary IVL Catheter was published in Circulation: Cardiovascular Interventions, with the RESTORE study showing feasibility and high procedural success in 41 patients with severely calcified coronary lesions. The same day, Johnson & Johnson disclosed 30-day follow-up data from a 30-patient prospective multicenter study of its Ottava robotic platform in gastric bypass, with all procedures completed robotically and no device-related adverse events, in a package now supporting the company’s FDA de novo submission. Read together, the three disclosures describe an interventional platform market that is rebuilding around three distinct technical mechanisms in calcium modification and a strategic-acquirer environment that is paying premiums for cleared platforms whose evidence base reads through the route the platform was designed to anchor. ## Stryker Closes Amplitude and Adds CO2 Pressure-Wave Lithotripsy Stryker announced on May 7 that it had completed the acquisition of Amplitude Vascular Systems, the Boston-based developer of the Pulse intravascular lithotripsy platform, per MedTech Dive. Financial terms were not disclosed, and the transaction closed less than a month after the definitive agreement was announced. Amplitude was running a U.S. pivotal study at the time of the deal and had raised $36 million in January 2025 to fund the device launch and trial program. The Pulse platform generates pressure waves through carbon dioxide delivered via a balloon catheter to fracture severely calcified arterial calcium and restore vessel patency, and the system has been designed for both peripheral and coronary applications. Stryker chief executive Kevin Lobo positioned the acquisition as a continuation of an active mergers and acquisitions cadence the company expects to maintain through the rest of 2026 and into 2027, with the Pulse platform anchoring the peripheral vascular portfolio alongside the Inari Medical clot-removal business Stryker acquired in 2025 for approximately $4.9 billion. The structural read for surgical robotics and advanced interventional founders is that calcium modification has now become a portfolio category large strategic acquirers are building around as a multi-platform stack rather than as a single-product line. Shockwave Medical, the original sonic IVL pioneer, was acquired by Johnson & Johnson in 2024 for $13.1 billion and now anchors the J&J Medtech peripheral and coronary vascular portfolio. Stryker has now built a competing peripheral vascular stack on Inari clot removal plus Amplitude Pulse IVL, and the company has signaled a continued acquisition cadence in the same calcium and clot category. The platforms whose calcium modification or vascular intervention technology fits inside the portfolio architecture of one of the major strategic acquirers gain a credible path to scale through the acquirer’s commercial channel. The platforms whose technology overlaps with the acquirer’s existing position face a steeper distribution build and a more cautious strategic-conversation profile when the platform reaches commercial readiness. ## Boston Scientific RESTORE First-in-Human and the Coronary IVL Path Boston Scientific reported first-in-human results for its Seismiq 4CE Coronary IVL Catheter on May 11, with the RESTORE study published in Circulation: Cardiovascular Interventions. The single-arm study enrolled 41 patients with a mean age of 71.7 years who presented with stable, unstable, or silent ischemia and were scheduled for percutaneous coronary intervention on severely calcified coronary lesions. The report showed that the Seismiq 4CE catheter delivered substantial improvement in luminal area, high rates of clinical procedural success, and excellent post-procedural stent expansion. The Seismiq platform uses laser energy to generate acoustic pressure waves that fracture calcium deposits, a mechanism distinct from the sonic-based Shockwave platform and the CO2-based Amplitude Pulse platform Stryker now owns. The Seismiq IVL system currently holds FDA approval for the treatment of peripheral artery disease and remains investigational in coronary indications, with the RESTORE first-in-human report positioning the coronary catheter for the next phase of the U.S. regulatory program. The structural read for surgical robotics and advanced interventional founders is that the coronary calcium modification market is now a three-way contest among technical mechanisms that were single-mechanism categories twelve months ago. The clinical performance data the RESTORE report produced is the kind of clean first-in-human evidence base the regulatory pathway and the strategic-acquirer evaluation both price at the premium end of the coronary intervention category, and the laser-acoustic mechanism gives Boston Scientific a differentiated technical position against the sonic and CO2 mechanisms the J&J and Stryker stacks anchor. Founders building advanced interventional platforms in the vascular space now have to position the platform within a coronary and peripheral category where three large strategic acquirers each anchor a distinct technical mechanism, where the strategic-acquirer evaluation prices the platform on the mechanism fit with the acquirer’s stack rather than on a generic IVL capability, and where the clinical-evidence base has to read against the specific mechanism category the platform will compete inside. ## J&J Ottava Gastric Bypass Study and the De Novo Submission Profile Johnson & Johnson reported on May 11 that a 30-patient prospective multicenter study of its Ottava soft-tissue surgical robotic platform in laparoscopic Roux-en-Y gastric bypass met its safety and efficacy endpoints through 30 days post-surgery, per MedTech Dive. Average weight loss across the 30-day window was approximately 30 pounds, and all procedures were completed robotically without conversion to a non-robotic surgical approach. No adverse events related to the Ottava device were reported. The 30-day follow-up data package is now part of the de novo authorization submission J&J filed with the FDA in January, with the submission targeting a general-surgery indication that covers Roux-en-Y gastric bypass, gastric sleeve, small bowel resection, and hiatal hernia repair, alongside an investigational device exemption trial in inguinal hernia repair that is running in parallel. J&J Medtech chief scientific officer Peter Schulam described the architecture as one designed to integrate into existing infrastructure and remove practical barriers to broader adoption. The company expects an Ottava U.S. launch by the end of 2026, pending FDA authorization. The structural read for surgical robotics founders is that J&J has now produced the second consecutive Ottava pivotal-quality readout in a major bariatric or upper-abdominal indication on a study architecture engineered against the de novo submission profile, and the platform is on a credible path to its first U.S. clearance by the end of the calendar year. For founders building multi-specialty soft-tissue surgical robotic platforms, the Ottava trajectory now describes the integrated-evidence profile the strategic-acquirer evaluation prices when the cleared platform reaches commercial submission readiness. For founders building specialist soft-tissue platforms, the trajectory now reframes the timeline pressure on the specialist platform whose first cleared indication is being read against a multi-specialty platform that has already produced two clean pivotal-quality readouts in bariatric procedures. The specialist platform that arrives at its first commercial customer in the second half of 2026 will be evaluated against an Ottava that hospital programs are at the same moment evaluating, and the procurement evaluation will read the specialist platform against the integrated J&J stack rather than against the standalone specialist platform alone. ## What the Three Disclosures Mean for the Founder Operating Plan Read across the IVL platform closes, the coronary first-in-human report, and the Ottava gastric bypass readout, the structural lesson for surgical robotics and advanced interventional founders is consistent across the three segments. The strategic-acquirer environment is paying premiums for cleared platforms whose evidence base reads against the specific route the platform was designed to anchor and whose technical mechanism fits inside the portfolio architecture the acquirer is building around the category. The clean first-in-human disclosure, the de novo submission package designed against the agency’s preferred clearance pathway, and the technical-mechanism positioning that matches the acquirer’s portfolio stack are now the three operating-plan lines that compound through the build years into the integrated-evidence and integrated-stack multiple the strategic conversation actually prices. For surgical robotics and advanced interventional founders running specialist platforms through 2026, the operating plan question is which architectural and partnership decisions need to be made now to produce the integrated-evidence and integrated-stack profile in three years. The Stryker-Amplitude close describes how the strategic-acquirer portfolio architecture extends into the next adjacent category once a foundational platform position is in place. The Boston Scientific RESTORE first-in-human report describes what the clinical-evidence base looks like when the study architecture is designed against the route the agency and the strategic acquirer both read. The J&J Ottava gastric bypass readout describes what the de novo submission profile looks like when the pivotal-quality readout architecture is engineered against the indication the agency is going to support. Founders building specialist surgical robotics or interventional platforms have to run the architectural, partnership, and evidence-design work as Day-1 strategy lines rather than as late-cycle deliverables, and the platforms whose work compounds through the build years arrive at the strategic conversation with the integrated profile the multiple actually prices. Dave’s take Calcium modification was a single-mechanism category twelve months ago. It is now a three-way contest among sonic, CO2 pressure-wave, and laser-acoustic mechanisms, with three different large strategic acquirers each anchoring a stack around a distinct technical position. Boston Scientific RESTORE produced the kind of clean coronary first-in-human report that lands a platform in the next phase of the regulatory program because the study was designed against the route. J&J Ottava produced a second consecutive pivotal-quality readout in a bariatric indication because the study architecture was engineered against the de novo submission profile. The vascular intervention and surgical robotics founders I work with this quarter are running the same evidence-design and stack-positioning discipline. Which technical-mechanism position is your platform actually building toward, and does the strategic acquirer whose stack you want to anchor already have a platform in that mechanism category? From Dave’s video library Dave's filter for picking one idea you can win with instead of five you can't. Calcium modification just split into three mechanisms, and each winner picked a single technical position and built the whole stack behind it. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to The Build newsletter: [The Build →](/field-notes/the-build) --- # Medtronic Splits Cardiovascular Into Two Operating Stack Units as Interventional Cardiology Therapies Takes Shape | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-05-13-medtronic-ict-cardiovascular-stack-restructure [← All Field Notes](/field-notes) May 13, 2026 Industry Roundup Medtronic disclosed a structural reorganization of its cardiovascular business this week, combining Structural Heart, Coronary, and Renal Denervation into a new Interventional Cardiology Therapies operating unit and merging Cardiac Surgery with Aortic into a new CardioVascular Surgery unit under newly promoted president Karim Bandali. MedTech Dive reported on May 12 that the company is closing its Santa Rosa, California, facility over the next two years, with operations shifting to Santa Ana, Minneapolis, and Galway, Ireland, and roughly 370 employees affected. Intuitive Surgical reported Q1 2026 results with 431 da Vinci systems installed globally and revenue of $2.77 billion, up 23% year over year. The Medtronic reorganization signals what kind of integrated stack the largest medtech strategic acquirer is committing to in the interventional cardiology category, and how the strategic-acquirer evaluation will read interventional platforms through the rest of 2026 and 2027. ## Medtronic Builds Interventional Cardiology Therapies as a Single Operating Unit Medtronic announced the reorganization through a spokesperson statement to MedTech Dive, with the company combining its Structural Heart, Coronary, and Renal Denervation businesses into a new Interventional Cardiology Therapies operating unit intended to strengthen focus and investment across the interventional cardiology continuum. Company spokesperson Justin Paquette positioned the integration as a way to advance the company’s strategy across these critical business areas. The three constituent businesses each anchor distinct interventional categories, with Structural Heart covering transcatheter aortic valve replacement and adjacencies, Coronary covering drug-eluting stents and percutaneous coronary intervention consumables, and Renal Denervation covering the Symplicity Spyral platform for resistant hypertension. The new operating unit consolidates strategic, commercial, and operating decision rights across the three businesses, with the practical implication that capital allocation, commercial channel architecture, and clinical-evidence investment now run through a coordinated interventional cardiology stack across the next operating cycle. The structural read for surgical robotics and advanced interventional founders is that the largest medtech strategic acquirer has now signaled at the operating-unit level that the interventional cardiology category is being run as an integrated stack across structural heart, coronary, and renal denervation. Founders building advanced interventional platforms in any of the three constituent categories now face a strategic-acquirer evaluation environment where Medtronic prices candidate portfolio additions against the integrated stack architecture the operating unit is built around, rather than against the standalone capability of the platform itself. The platforms whose technical mechanism, clinical-evidence base, and commercial channel architecture fit inside the integrated stack profile gain a credible operating-partnership pathway and a clear route to the integrated-stack multiple at the strategic transaction. The platforms whose profile overlaps with a position the integrated stack already anchors face a more selective evaluation and a more cautious operating-partnership conversation through the rest of 2026 and 2027. ## CardioVascular Surgery Unit and the Santa Rosa Footprint Decision Medtronic combined its Cardiac Surgery and Aortic businesses into a new CardioVascular Surgery operating unit under newly promoted president Karim Bandali, who previously led the cardiac surgery business. The integrated unit consolidates strategic responsibility across surgical aortic valve replacement, the heart-lung machine and cardiopulmonary bypass franchise, the ablation portfolio aligned to cardiac surgery, and the Endurant and Valiant endovascular grafts that sit alongside the aortic business. Paquette described the consolidation as a way to strengthen focus and investment across the cardiovascular care continuum and to partner more closely with vascular, cardiothoracic, and cardiac specialists across the integrated category. The CardioVascular Surgery unit and the Interventional Cardiology Therapies unit now operate as the two reorganized operating pillars beneath the cardiovascular segment Medtronic carries on its quarterly financial reporting. The Santa Rosa closure runs alongside the reorganization on the operating-footprint side. MedTech Dive reported that Medtronic will begin departing the Santa Rosa, California, facility in spring 2027 with full closure by spring 2028. The site has operated since Medtronic acquired stent manufacturer Arterial Vascular Engineering in 1998 and currently houses approximately 370 employees, with Press Democrat reporting confirming the workforce impact. Operations will move primarily to Santa Ana, California, Minneapolis, and Galway, Ireland, against the integrated operating profile the cardiovascular reorganization is engineered to produce. For founders building interventional platforms whose strategic-acquirer pathway runs through Medtronic, the operational consolidation runs in parallel with the operating-unit integration, and the strategic-acquirer evaluation in the next two years will read the candidate platform against an integrated stack and an integrated operating footprint that both got more concentrated this quarter. ## Intuitive Surgical Q1 2026 and the Established Stack Comparable Intuitive Surgical reported Q1 2026 results with 431 da Vinci systems placed globally and revenue of $2.77 billion, up 23% year over year. The company has continued the measured 2026 commercial expansion of the da Vinci 5 platform across the nine-indication FDA clearance package and is ramping training programs for cardiac teams at selected hospitals in advance of a limited U.S. cardiac rollout later this year. The Q1 print sits against a competitive backdrop in which the Medtronic Hugo platform holds an FDA clearance in urology and an active general-surgery pivotal program, CMR Surgical has launched the Versius Plus in the U.S. on the strength of the cholecystectomy 510(k), and Johnson & Johnson Ottava is in front of the FDA on a de novo classification request for general surgery. The structural read for surgical robotics founders is that the established Intuitive position prices the Q1 print at the integrated-stack multiple because the recurring instrument and consumable layer underneath each placement compounds across the installed base every quarter. Each placement adds to the recurring-revenue layer the integrated stack carries through the operating reporting, and each integrated indication adds to the procedure-mix breadth the strategic-acquirer evaluation reads as the durable economics of the platform. The Medtronic operating reorganization, the Intuitive Q1 print, and the Boston Scientific 2026 guidance cut earlier this quarter all describe a strategic environment in which the largest medtech operators are clarifying what kind of integrated stack each is committing to and pricing the rest of the category against that stack profile. The platforms that compound through the build years into an integrated-stack profile of their own arrive at the strategic conversation with a comparable the buyer reads at the integrated-stack multiple. ## What the Operating Reorganization Means for the Founder Stack Plan Read across the Medtronic operating-unit reorganization, the Santa Rosa footprint consolidation, and the Intuitive Q1 integrated stack read, the structural lesson for surgical robotics and advanced interventional founders is that the largest strategic acquirers are clarifying at the operating level what kind of integrated stack each is committing to over the next decade. The Interventional Cardiology Therapies integration tells founders building structural heart, coronary, or renal denervation platforms which adjacent categories Medtronic is committed to anchoring inside a single operating stack. The CardioVascular Surgery integration tells founders building cardiac surgical and aortic platforms which adjacent categories Medtronic is committed to running inside a parallel operating stack. The Intuitive print tells founders building soft-tissue surgical robotic platforms what the integrated established stack looks like once the installed base, the recurring instrument and consumable layer, and the procedure-mix breadth compound through the operating reporting. The founder operating plan question across surgical robotics and advanced interventional is which specific stack inside which specific strategic acquirer the cleared platform is being designed to fit, and the answer has to be embedded into the technical, clinical, and partnership architecture from initial product architecture. Founders who run the stack-positioning decision through the build years arrive at the strategic conversation with an integrated profile the acquirer reads at the integrated-stack multiple. Founders who defer the question to a downstream strategic conversation arrive with a standalone-capability profile the acquirer prices at the standalone-capability multiple regardless of the engineering quality of the platform. Dave’s take Medtronic just told the market what kind of integrated stack it is building for the interventional cardiology category over the next decade, and the operating-unit reorganization plus the Santa Rosa consolidation describe a strategic acquirer that is concentrating both its commercial intent and its operating footprint behind that stack. Founders building structural heart, coronary, or renal denervation platforms now have a clearer read on which specific gap inside which specific operating unit the platform has to fit, and the architectural, evidence-design, and partnership work that produces that fit has to start now. The cardiovascular-platform founders I work with this quarter are running the stack-fit question through the operating cadence the same way they run regulatory and reimbursement, because the strategic-acquirer evaluation in 2027 is going to read the integrated profile the architectural work produced rather than the standalone capability the engineering work alone would imply. From Dave’s video library Dave on selling as understanding the buyer's math before they explain it to you. Medtronic's reorganization is that math published as an org chart, and founders selling into the category should read it that way. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to The Build newsletter: [The Build →](/field-notes/the-build) --- # AI Imaging and Robotic Navigation Converge in the Interventional Stack | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-05-14-ai-imaging-robotic-navigation-interventional-stack [← All Field Notes](/field-notes) May 14, 2026 Industry Roundup Siemens Healthineers disclosed on May 12 that the FDA cleared six new systems in its Artis interventional imaging portfolio, all running a unified Optiq AI imaging chain that applies deep-learning-based noise reduction across fluoroscopy, acquisition, and digital subtraction angiography. The cleared portfolio spans the Artis vision platform in floor, biplane, ceiling, and pheno robotic floor-mounted configurations alongside floor configurations of Artis icono.explore and Artis genio. Stereotaxis reported Q1 2026 results the same day and described its definitive agreement to acquire Robocath, with management framing the combination as a unified robotic platform spanning magnetic and mechanical endovascular navigation across electrophysiology, interventional cardiology, and neurointervention. Alphatec used the Stifel Jaws of Life conference this week to detail its Valence spine navigation and surgical robot launch alongside fiscal 2026 guidance for $882 million of revenue, $134 million of adjusted EBITDA, and 15% top-line growth into a $10 billion spine market opportunity. Three disclosures across three days describe the same structural shift inside interventional and surgical robotic categories, with imaging chains, navigation platforms, and robotic mechanisms integrating into single operating stacks rather than competing as standalone capabilities. ## Siemens Optiq AI Pulls Six Interventional Imaging Systems Onto a Single AI Stack Siemens Healthineers received FDA 510(k) clearance for six new interventional imaging systems built around a unified Optiq AI imaging chain, with the disclosure published through the company press room and reported across the trade press on May 12. The cleared portfolio includes floor, biplane, ceiling, and pheno robotic floor-mounted versions of the Artis vision platform alongside floor configurations of Artis icono.explore and Artis genio. Optiq AI applies deep-learning-based noise reduction across fluoroscopy, acquisition, and digital subtraction angiography in real time, with the platform engineered to deliver consistent visualization at lower radiation doses across the procedural mix the interventional suite carries. The clearance also extends syngo DynaCT MORE motion correction to the 3D acquisition profile and standardizes the Touch UI tableside experience across the entire portfolio. Kris McVey, head of Angiography at Siemens Healthineers North America, told the trade press that the new portfolio delivers the image-quality profile clinicians require to guide complex interventional procedures. The structural read for advanced interventional founders is that the largest interventional imaging vendor has now standardized a single AI imaging chain across a six-platform portfolio that spans the entire procedural mix from coronary intervention through structural heart, peripheral vascular, neurointervention, and pheno robotic-mounted hybrid OR configurations. The interventional suite the founder is designing the platform to operate inside is now a Siemens-Optiq integrated stack across imaging hardware, AI denoising, motion correction, and tableside workflow rather than a collection of standalone imaging units the platform interfaces with separately. Founders building interventional devices, navigation systems, or imaging-adjacent capabilities now face an integrated stack environment in which the standalone interface profile reads against an integrated imaging stack the suite is being built around. The platforms whose technical profile, workflow integration, and clinical-evidence base read inside the integrated imaging stack the suite is built on land in the procedural workflow at the integrated-stack profile, and the platforms whose profile sits outside the integrated imaging stack run a more selective integration conversation with the interventional suite operator across the next two to three years. ## Stereotaxis Q1 and the Robocath Combination Push Robotic Endovascular Navigation Onto a Single Platform Stereotaxis reported Q1 2026 financial results on May 12 with revenue of $6.3 million split between $1.3 million of systems revenue and $5.0 million of recurring revenue, against a comparable Q1 2025 print of $7.5 million. The company logged a $5.9 million net loss for the quarter and described the period as a transition between legacy revenue dependency and the new product ecosystem the company is rolling out across 2026. CEO David Fischel described the quarter as the company essentially building a fresh start-up on the shoulders of the legacy platform, with two FDA clearances earlier in the year for the MAGiC cardiac ablation catheter in January and the Synchrony digital operating room in April. Demand for MAGiC has now run ahead of supply, with the company scaling manufacturing toward a 500-catheter monthly run rate by year end across U.S. and European deployments. The Robocath acquisition is the structural piece of the Q1 disclosure. Stereotaxis described the definitive agreement as the basis of a unified robotic platform combining the company’s magnetic navigation system with the mechanical endovascular navigation platform Robocath has built across percutaneous coronary and peripheral procedures, with a stated intent to deliver remote, automated, and fully robotic treatment across electrophysiology, interventional cardiology, and neurointervention. The structural read for surgical robotics and advanced interventional founders is that the specialist robotic-navigation category is now consolidating into single-vendor integrated platforms that span magnetic and mechanical mechanisms across multiple procedure categories rather than holding as standalone single-mechanism platforms inside discrete interventional categories. Founders building specialist endovascular or interventional robotic platforms now face an integrated-platform competitive environment, and the architectural decisions about which mechanism categories the platform spans, which procedure categories the platform anchors, and which integration profile the cleared platform reads at the strategic conversation become first-order operating decisions across the next two to three years. ## Alphatec Valence Adds Spine Robotics Onto an Integrated Lateral Procedure Platform Alphatec presented at the Stifel Jaws of Life conference this week and detailed the Q1 2026 launch of Valence, the company’s navigation and surgical robot system integrated into the existing lateral procedure platform across retractor placement, discectomy, endplate preparation, and interbody placement alongside navigated and robotic pedicle screw placement. CFO Todd Koning sized the addressable spine market at approximately $10 billion with company penetration just under 10%, and reiterated 2026 guidance of approximately $882 million of revenue, 15% top-line growth, $134 million of adjusted EBITDA, and $20 million of free cash flow. The company achieved positive adjusted EBITDA in 2024, runs cash flow positive in the current quarter, and refinanced approximately $200 million of term debt to reduce interest expense by approximately 300 basis points or $6 million annually. Valence is positioned against the broader spine robotics category by integrating across the entire lateral procedure rather than only the pedicle-screw placement step the established spine robotic platforms anchor. The structural read across the Alphatec disclosure is that the spine robotics category is now competing on integration depth across the full procedural workflow rather than on the standalone capability of the robotic mechanism. The companies that anchor a procedure category at the strategic-acquirer multiple in spine robotics are the ones whose platform spans the full surgical workflow inside which the robotic step is embedded, with the navigation, retractor placement, discectomy, endplate preparation, interbody placement, and screw placement steps integrated into a single operating profile the surgical team and the spine-implant business run together. The platforms whose profile spans only the robotic mechanism step run a different evaluation conversation than the platforms whose profile integrates the mechanism into the broader procedural workflow, and the spine-implant business that operates the procedural platform across the integrated workflow drives the per-procedure economics the strategic conversation prices. ## What the Convergence Means for the Founder Operating Plan Read across the Siemens Optiq AI clearance, the Stereotaxis-Robocath combination, and the Alphatec Valence launch, the structural lesson for surgical robotics and advanced interventional founders is that the integrated stack architecture pattern that has reorganized the strategic-acquirer landscape across cardiovascular and surgical robotics over the most recent year is now extending into the interventional imaging chain, the robotic-navigation platform layer, and the spine procedural workflow. The standalone capability that competes on technical profile alone now operates inside an integrated stack environment in which the imaging chain, the navigation platform, the robotic mechanism, and the procedural workflow are converging into single operating stacks. Founders designing interventional or surgical robotic platforms across the next two to three years run the architectural decision about which integrated stacks the platform reads inside, which gaps inside each stack the platform is built to anchor, and which integration profile the cleared platform delivers at the strategic conversation as a first-order operating decision rather than as a downstream commercial conversation. The companies that arrive at the strategic conversation with an integrated profile read at the integrated-stack multiple are the ones whose architectural and partnership work through the build years aligns the platform to a specific gap inside a specific integrated stack, with the technical mechanism, the clinical-evidence base, the integration interface profile, and the commercial channel architecture engineered against the operating stack the strategic acquirer or the procedural-suite operator reads. The companies whose architectural decision deferred the integration question to the strategic conversation arrive with a standalone-capability profile the buyer prices against the standalone-capability multiple regardless of the engineering quality of the cleared platform. Three disclosures across three days describe what the integrated-stack environment actually looks like at the operating level inside interventional imaging, robotic navigation, and spine procedural workflow. Dave’s take The Siemens portfolio clearance, the Stereotaxis-Robocath combination, and the Alphatec Valence launch all describe the same structural pattern from three different angles inside the interventional and surgical robotic categories. The standalone capability that competes on technical profile alone is being squeezed by integrated stacks that span imaging, navigation, robotic mechanism, and procedural workflow inside a single operating profile, and the founders I work with this quarter are running the integration architecture decision the same way they run regulatory and reimbursement. The companies that finish well in this environment design the integration profile from initial product architecture and run the operating-partnership cadence with the integrated stack operators through the build years rather than waiting for the strategic conversation to arrive. From Dave’s video library In this one Dave lands on the rule that people buy better working, not better looking. Standalone imaging and navigation tools are being squeezed by integrated stacks for exactly that reason: the buyer wants the thing that works end to end. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to The Build newsletter: [The Build →](/field-notes/the-build) --- # Hard-Tech Mega-Rounds Land on the Physical Layer of the AI Stack | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-05-15-hard-tech-mega-rounds-physical-layer-capital [← All Field Notes](/field-notes) May 15, 2026 Commercialization Anduril announced a $5 billion Series H on May 13 at a $61 billion valuation, led by returning investors Thrive Capital and Andreessen Horowitz, with total capital raised crossing $11 billion against $2.2 billion of 2025 revenue per company disclosures reported by TechCrunch. Rivian spinout Mind Robotics announced a $400 million Series B the same day at a $3.4 billion valuation, led by Kleiner Perkins, bringing the seven-month-old company past $1 billion of cumulative funding behind a foundation-model approach to industrial robotics deployed first at the Rivian Normal, Illinois plant per coverage in The Robot Report and SiliconANGLE. Iceotope Group closed a $26 million Series B on May 14, led by Two Seas Capital and Barclays Climate Ventures, scaling its precision liquid-cooling stack into AI data-center deployments where rack densities now push past one megawatt and air cooling no longer holds. Three rounds across forty-eight hours, three different hard-tech verticals, and one underlying pattern about how capital is pricing the physical layer of the AI and automation stack in 2026. ## Anduril $5B Series H Resets the Defense-Hardware Ceiling Anduril’s Series H closed at $5 billion on a $61 billion valuation on May 13, doubling the $30.5 billion valuation the company carried after its June 2025 round and pushing cumulative capital past $11 billion against trailing-year revenue of $2.2 billion that doubled year over year. The round was led by returning investors Thrive Capital and Andreessen Horowitz, with TechCrunch reporting the structural framing CEO Brian Schimpf gave to the financing as a recognition of how far defense venture has moved since the company was founded in 2017. The capital lands against a product portfolio that now spans Fury autonomous fighter jets, the Lattice command-and-control platform, and battle-management software, with active programs including the space-based Golden Dome missile defense architecture announced earlier this month, a Dutch Ministry of Defence contract, and a U.S. Army battle-management software award. For defense and dual-use hardware founders, the structural read is that the institutional ceiling on hardware-intensive defense companies has moved decisively. The historical objection that defense hardware could not absorb venture-scale capital at venture-scale valuations no longer holds at the round level. The newer question is what the bar for participation looks like at the $5 billion-round tier, and the Anduril operating profile gives the answer. The company arrived at the round with $2.2 billion of revenue, a multi-vendor program inside the Pentagon’s top architectural priorities, a vertically integrated manufacturing footprint, and a track record of converting research contracts into production programs across an eight-year operating history. Defense-hardware founders raising over the next twelve to eighteen months will be priced against that operating profile whether or not the round is anywhere near $5 billion, and the architectural and operating decisions that produce a comparable readiness profile are the ones that matter from the seed round forward. ## Mind Robotics $400M Series B Repositions Industrial Robotics as a Foundation-Model Bet Mind Robotics announced a $400 million Series B led by Kleiner Perkins on May 13, with returning investors Accel, Andreessen Horowitz, Eclipse, Prysm Capital, Bain Capital Ventures, and Greenoaks alongside new investors Meritech, Redpoint, SV Angel, Incharge Capital Partners, A-Star Capital, and Garuda Ventures. The Robot Report covered the round and confirmed the company has now raised over $1 billion across a $115 million seed, a $500 million Series A in March 2026, and the new Series B, with a $3.4 billion valuation reached less than six months after the company was spun out of Rivian in November 2025. CEO RJ Scaringe continues to run Rivian alongside the new company, and Rivian operates as both shareholder and the first deployment customer for the platform inside its Normal, Illinois manufacturing footprint. The product targets the judgment-dense tasks conventional industrial automation has historically failed at, including wiring-harness assembly, trim fitting, and connector mating, with a foundation-model architecture trained on production-grade manufacturing data the Rivian relationship supplies. For industrial robotics, physical-AI, and automation founders, the round describes the architectural framing the category is now being priced against. Capital is flowing toward platforms whose moat sits in the production-grade data flywheel and the foundation-model architecture, not in the mechanical novelty of the robot itself. SiliconANGLE captured the framing the company is leading with: that general-purpose industrial robots will be won by whoever can collect and learn from the most production-grade data rather than by whoever builds the flashiest humanoid. Founders building robotic platforms, fleet-management software, or physical-AI infrastructure now face a category in which the lead investor profile, the data architecture, and the anchor manufacturing partner are first-order operating questions years before commercial scale, and the architectural decisions that produce a comparable data and partnership profile are the ones that compound through the build years. ## Iceotope Series B Confirms the AI-Infrastructure Hardware Layer as Its Own Category Iceotope Group closed a $26 million Series B on May 14, led by Two Seas Capital and Barclays Climate Ventures, with participation from Edinv, ABC Impact, Northern Gritstone, and British Patient Capital. The company builds precision liquid-cooling systems for data centers and edge infrastructure, and the round funds product and engineering expansion, patent prosecution, and a partner ecosystem aimed at scaling deployments where next-generation accelerator platforms are driving rack densities toward one megawatt and beyond. The trade coverage framed the round as a notable infrastructure play in the AI build-out cycle, with thermal management shifting from routine data-center operations to strategic infrastructure as computational density increases. The round is far smaller than the Anduril and Mind Robotics financings, and that is part of the signal: climate-adjacent and infrastructure hardware companies addressing AI-driven physical bottlenecks are pulling institutional climate-tech and patient-capital pools alongside the headline mega-round investors. For climate-hardware and AI-infrastructure founders, the round is the clearest current signal that the physical layer of the AI stack is being priced as a category in its own right rather than as a sub-category of data-center capital expenditure. The institutional read for founders building liquid cooling, power conditioning, thermal substrates, photonics packaging, or other AI-physical-infrastructure platforms is that the patient-capital climate pools and the strategic AI-infrastructure investors are now meeting in the same syndicate, and the operating profile a credible Series B requires combines a defensible technical mechanism, a patent portfolio managed against the AI-infrastructure roadmap, and a partner ecosystem positioned to deliver into hyperscaler and enterprise deployments. The strategic conversation through the next two to three years is going to read candidate platforms against that profile, and the architectural and partnership work that produces the profile has to be designed in years before the round. ## What the Three Rounds Say Together Read across Anduril, Mind Robotics, and Iceotope, the structural pattern is that institutional capital is now pricing the physical layer of the AI and automation stack as the high-conviction category across the next decade rather than as a hardware sidecar to the software platforms it supports. Defense hardware, industrial robotics, and AI-infrastructure climate hardware are pulling capital at scales these verticals historically did not access, and the operating profiles that won the rounds are not the historical hardware-startup profile. Anduril arrived with multi-billion revenue and vertically integrated production. Mind Robotics arrived with a foundation-model architecture and a captive manufacturing data partnership. Iceotope arrived with a patent portfolio against an AI-infrastructure roadmap and an institutional climate-tech syndicate alongside strategic capital. The shared lesson for hard-tech founders is that the round profile that priced these three businesses at the levels reported is the round profile the strategic-conversation environment now reads candidate platforms against across defense, robotics, climate hardware, industrial physical AI, and adjacent regulated or safety-critical hardware categories. The companies that finish well in this environment design the operating profile against the institutional ceiling that is now visible. If you are building a defense-hardware platform, the architectural and operating-cadence work that produces a multi-vendor program profile inside a Pentagon top-line architecture priority is the work the strategic conversation will price. If you are building industrial robotics or physical AI, the data-partnership architecture and the foundation-model technical stance are the architectural decisions the strategic-conversation environment is reading at the round level. If you are building AI-infrastructure climate hardware, the patent and partnership profile against the hyperscaler and enterprise AI-deployment roadmap is the operating work the round environment is pricing. None of these profiles is engineered in the months before the round. They are engineered across the build years, and the founders running the build phase against the wrong round profile arrive at the institutional conversation with a different category of business than the round environment is now pricing. Dave’s take Three rounds across three hard-tech verticals in forty-eight hours is the cleanest signal I have seen all year that the institutional environment for hardware companies has structurally moved. The capital is real, the valuations are real, and the operating profiles those rounds reward are specific enough that you can reverse-engineer the build-phase work that produces them. The founders I am working with this month are running the question the same way: what does the operating profile look like at the round we plan to raise, and what architectural and partnership work has to be in motion now to produce that profile when the round arrives. From Dave’s video library Dave on why the only clock that matters is external proof, not months of building. The mega-rounds landing on hard-tech companies reward specific operating proof, and you can reverse-engineer the work that produced it. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to The Build newsletter: [The Build →](/field-notes/the-build) --- # Rocket Lab and RTX Confirm the Subcontractor Playbook for Defense Hardware | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-05-16-rocket-lab-rtx-defense-hardware-subcontractor-playbook [← All Field Notes](/field-notes) May 16, 2026 Defense Hardware Rocket Lab disclosed on May 15 that it has partnered with RTX as a subcontractor to bid for work under the $3.2 billion Space-Based Interceptor umbrella contract inside the Golden Dome missile defense program, leveraging its prime credentials on the predecessor Proliferated Warfighter Space Architecture tracking satellite work and supplying RTX with Electron and forthcoming Neutron launch capacity per coverage in The Motley Fool. AeroVironment’s Department of War Test Resource Management Center announced a separate $43 million, three-year award on May 12 for the company’s Phased Array Next-gen Telemetry Hypersonic Emitter Receiver to be installed across the Northrop Grumman SkyRange hypersonic test fleet of converted RQ-4 Range Hawks per coverage in The Defense Post. The Pentagon’s FY27 budget request, released in early April and dissected this month across Aviation Week, The Hill, and Defense One, proposes to fund the Defense Autonomous Warfare Group at roughly $54.6 billion against a current-year base of $226 million, with two new DARPA RFIs under the Materials for Physical Compute in Untethered Robotics and Decentralized Artificial Intelligence through Controlled Emergence programs telling the supplier base what the autonomy buy is going to require. Read the four signals together, and the defense-hardware buy-side is being rewired in public, with a participation profile founders building defense or dual-use hardware can now reverse-engineer from the disclosures. ## Rocket Lab and RTX Confirm the Subcontractor Path on Golden Dome Rocket Lab announced on May 15 that the company has partnered with RTX as a subcontractor to bid for work under the $3.2 billion Space-Based Interceptor umbrella contract that anchors the Golden Dome missile defense program, with the partnership timed to the company’s earnings cycle and disclosed in coverage from The Motley Fool. The structural detail is that Rocket Lab brings two assets RTX needs to satisfy the SBI bid profile and that RTX cannot organically supply. First, Rocket Lab already holds a prime contractor position on the predecessor Proliferated Warfighter Space Architecture tracking satellite program, giving the subcontractor relationship a credentialed track record on the closest comparable Space Force architecture. Second, Rocket Lab brings Electron launch capacity today and Neutron launch capacity through 2027, supplying RTX with a launch profile RTX cannot deliver inside the SBI program timeline without an external partnership. For defense and dual-use hardware founders, the structural read is that the prime-subcontractor architecture is now the primary participation channel for the mega-programs the Pentagon is building. The Golden Dome SBI bid set, the Sentinel ground-based architecture, the CCA autonomous-fighter program, the hypersonic-missile production lines, and the DAWG-aligned autonomy programs are not won on a clean prime bid by a pre-revenue defense-hardware company. They are won by primes who assemble a subcontractor stack of credentialed capability suppliers that close the capability gaps the prime cannot organically deliver inside the program timeline. The Rocket Lab disclosure shows the credentialed-capability profile the prime side is now looking for: a referenceable position on a predecessor program, a capability the prime cannot organically supply, and an operating cadence that can plug into the prime’s bid schedule. Founders building the next generation of defense or dual-use hardware platforms have to design the operating profile, the predecessor-program references, and the prime-side relationship cadence against that subcontractor-stack architecture, because the prime conversation will read candidate suppliers against the credentialed-capability profile rather than against the standalone-product profile that wins the founder-investor conversation. ## AeroVironment’s PANTHER Win Shows How Test Infrastructure Becomes a Procurement Lane The Department of War Test Resource Management Center announced on May 12 that AeroVironment received a $43 million, three-year award to install the company’s Phased Array Next-generation Telemetry Hypersonic Emitter Receiver across the Northrop Grumman SkyRange hypersonic test fleet of converted RQ-4 Global Hawk aircraft now operating as Range Hawks, with integration anchored at the GrandSKY business park in Grand Forks, North Dakota per coverage in The Defense Post. The structural detail is that PANTHER is not a weapons system or a deployed platform. It is a sensor-and-telemetry payload that enables the Pentagon’s hypersonic test ranges to capture data from multiple high-speed projectiles simultaneously, and the contract is part of a procurement lane that funds the test infrastructure the broader hypersonic-missile production buy is going to require. AeroVironment is also opening curriculum partnerships with Bismarck State College to develop PANTHER operations and sustainment specialists, indicating the company is positioning for the sustainment tail of the contract as a separate revenue layer beyond the initial integration award. For defense-hardware founders, the structural read is that the test-infrastructure procurement lane is now a defensible revenue channel inside the broader Pentagon hardware buy, and the operating profile that wins those awards is different from the operating profile that wins a deployed-system contract. The test-infrastructure lane rewards companies that supply a sensor, an instrumentation payload, a telemetry chain, a calibration capability, or a digital twin against a specific test program the Pentagon has already funded. The bid profile reads the candidate’s technical capability against the specific test-program requirement, the supplier’s ability to deliver inside the program timeline, and the operating-cadence track record on comparable test programs. Defense and dual-use hardware founders who are building sensor, instrumentation, telemetry, or calibration platforms that the hypersonic, autonomous, or directed-energy buy is going to require should be mapping their architecture against the test-program funding lines already disclosed in the FY26 and FY27 budgets, because those test-infrastructure lanes are spending real money against a defined supplier profile right now while the weapon-system production lanes are still consolidating. ## DAWG’s FY27 Budget Structure Reframes the Autonomy Buy at Categorical Scale The Pentagon’s FY27 budget request proposes to fund the Defense Autonomous Warfare Group at roughly $54.6 billion against a current-year base of about $226 million, with only $1 billion in the standard base defense budget and the remaining $53.6 billion proposed through the $350 billion defense reconciliation package per coverage in Aviation Week, The Hill, and Defense One. The structural detail is that DAWG is being stood up as a sub-unified command-style organization to consolidate the autonomous-systems buy across the services, and the budget construction signals two things at the supplier level. First, the categorical scale of the autonomy buy is now larger than the FY27 budget request for the Marine Corps, which means the operating-profile bar the Pentagon will read candidate suppliers against is going to compound rapidly across the next twenty-four months as the program offices spin up. Second, the funding split between base and reconciliation tells the supplier base the program-execution cadence will be skewed toward fast-cycle awards out of the reconciliation pool rather than the slower base-budget cadence, which favors suppliers who have already invested in the operating-cadence work that lets them respond to a fast-cycle RFP profile. For autonomy, robotics, AI-physical-infrastructure, and dual-use hardware founders, the DAWG budget construction is the cleanest current signal that the autonomous-systems buy has been moved into a categorical capital line. The downstream effect is that the prime conversation, the program-office conversation, and the institutional-investor conversation will read candidate suppliers against an operating profile aligned to a buy that has been categorically scaled rather than against the prior autonomy-as-experimentation framing the budget construction has now retired. Founders running autonomy or robotics platforms who are still operating against the prior framing will arrive at the prime-side and program-office conversations with an operating profile the FY27 environment is no longer reading. The architectural, partnership, and operating-cadence work that produces the new operating profile has to be in motion now, because the FY27 program-office buildout will start running the supplier-conversation cadence inside the next two to four quarters and the operating profile the program offices will read is being shaped today. ## DARPA’s New RFIs Tell You What the Pentagon Wants to Buy in 2027 DARPA released two new RFIs in early May alongside the DAWG budget framing, with Defense One reporting on May 4 that the Materials for Physical Compute in Untethered Robotics program targets technologies that enable autonomous systems to think and reason without relying on vulnerable external data-center connections, and the Decentralized Artificial Intelligence through Controlled Emergence program targets the technologies that enable robotic platforms to form peer-to-peer teams and execute complex missions without central command. The structural detail is that an RFI is the earliest formal Pentagon procurement signal an outside supplier can read. It identifies the specific technical problem the Pentagon wants the supplier base to solve, the operating profile the future program will read candidate suppliers against, and the rough timeline on which the program will move from RFI to formal solicitation. The two RFIs together tell the autonomy and robotics supplier base that the Pentagon’s FY27 and FY28 autonomy buy will reward platforms whose architectural stance is built around edge compute for untethered operation and peer-to-peer multi-agent coordination, rather than around centralized command-and-control architectures that depend on persistent data-center connectivity. For founders building autonomy, robotics, edge-compute, AI-physical-infrastructure, or sensor-fusion platforms, the structural read on the RFIs is that the architectural decisions you make this year inside your technical roadmap will determine whether the operating profile of the business reads against the FY27-FY28 program-office buy or against a prior generation of program-office requirements. Edge-compute for autonomous systems and peer-to-peer multi-agent coordination are not architectural niceties. They are the architectural framing the Pentagon has now told the supplier base it is going to buy against. Defense and dual-use hardware founders running an architectural review this quarter should be reading the RFIs as a primary input into the technical roadmap and the partnership-stack decision, because the founders whose platforms read against the RFI framing when the program offices spin up will participate in a buy that the founders running prior-generation architectures will not. ## What the Four Signals Read Together Read across Rocket Lab and RTX, AeroVironment, DAWG, and the DARPA RFIs, the structural pattern is that the Pentagon is rewiring the defense-hardware buy-side in public and is giving the supplier base unusually clear advance signal on the operating profile the rewired buy is going to read. The Rocket Lab disclosure shows the prime-subcontractor architecture that is now the participation channel for the mega-program SBI, CCA, Sentinel, and hypersonic-production bids. The AeroVironment award shows that the test-infrastructure procurement lane is spending real money against a defined supplier profile right now and rewards companies that can plug a sensor, telemetry, or instrumentation payload into a specific test program. The DAWG budget construction shows that the autonomy buy has been moved into a categorical capital line whose program-office cadence will start reading candidate suppliers inside the next two to four quarters. The DARPA RFIs show what architectural framing the autonomy buy will reward at the program-office level when those conversations spin up. The companies that finish well in this environment design the operating profile against the rewired buy-side that is now disclosed in public rather than against the buy-side that used to exist. For a defense-hardware founder, the architectural work that produces a credentialed-capability profile, a predecessor-program reference, and an operating-cadence track record the primes can plug into is the work the SBI-class bid environment will price. For an autonomy or robotics founder, the architectural framing aligned to the DARPA RFIs and the operating-cadence work aligned to the DAWG program-office buildout are the operating decisions the FY27 program-office conversation will read. For a sensor, instrumentation, or telemetry founder, the test-infrastructure procurement lane is a real revenue channel that rewards companies that have mapped their architecture against a specific test program already inside the budget. None of these participation profiles are engineered in the months before the bid opens. They are engineered across the years of build-phase work that precede the bid, and the founders running the build phase against the prior generation of buy-side framing will arrive at the rewired buy with a business the rewired buy is no longer reading. Dave’s take Four signals across four days that all point at the same thing is the cleanest read I can give you on where the defense-hardware buy-side is moving. The Pentagon is rewiring the buy in public, and the participation profile for defense and dual-use hardware founders is being written down inside the program documents, the budget structures, and the prime-side partnership announcements. The founders I am working with in defense and adjacent dual-use right now are running the same question across every architectural review: does the operating profile we are building today read against the buy that is being assembled, or against the buy that used to exist? From Dave’s video library Dave on getting attention for what you built when no one is looking for it. In defense hardware that problem runs through the primes: Rocket Lab just showed how a challenger borrows an incumbent's access to the buyer. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to The Build newsletter: [The Build →](/field-notes/the-build) --- # Physical AI Crosses From VC Bet to Incumbent Install Base | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-05-17-physical-ai-fanuc-google-incumbent-install-base [← All Field Notes](/field-notes) May 17, 2026 Industrial Robotics Fanuc disclosed on May 14 a partnership with Google Cloud and Google’s Intrinsic robotics unit to integrate Gemini Enterprise and the Intrinsic Flowstate platform into the company’s industrial robot systems, with shares trading up as much as 16 percent intraday to a record close per coverage in Bloomberg and Nikkei Asia. Three Figure AI Helix-02 humanoids ran a livestreamed autonomous package-sorting trial across roughly 50 hours of continuous operation with no teleoperation and no human intervention, with CEO Brett Adcock describing the run as “uncharted territory” per Bloomberg coverage on May 15. The HII High-Yield Production Robotics program launched in April at Sea-Air-Space 2026 with Path Robotics and GrayMatter Robotics, combining robotic welding, automated material movement, autonomous surface treatment, and autonomous quality checks into an assembly line aimed at lifting Navy shipbuilding throughput another 15 percent on top of last year’s 14 percent gain per HII’s newsroom release and Globe Newswire. Read the three signals together, and physical AI just shifted from a VC bet into an incumbent install-base play, with a participation profile hard-tech founders building robotics, autonomy, or industrial AI hardware can now reverse-engineer from the disclosures. ## Fanuc and Google Make the Installed Robot the Platform Fanuc and Google disclosed on May 14 a partnership to integrate Google Cloud’s Gemini Enterprise model layer and Google’s Intrinsic robotics platform into Fanuc’s industrial robot systems, with full compatibility planned for Fanuc’s robots inside the Intrinsic Flowstate development environment per coverage in The Next Web and Nikkei Asia. Fanuc shares rose as much as 16 percent intraday and closed at a record high per Bloomberg. The structural detail is that there are roughly 1.1 million Fanuc robots already installed in factories worldwide, and the upgrade opportunity for the Google AI layer is larger than any new robot Fanuc could sell. Fanuc has built a two-platform AI stack with NVIDIA Omniverse and Isaac simulation on the training-and-commissioning side, disclosed in March, and Google Gemini plus Intrinsic Flowstate on the foundation-model and orchestration side, disclosed this month. The result is that the installed base of Fanuc robots becomes the deployment surface for the physical AI stack, and the revenue line that grows from here is the AI subscription overlay rather than the next industrial arm Fanuc ships out the door. For founders building industrial robotics, autonomy, or any hardware that lives inside a manufacturing or logistics footprint, the structural read is that the incumbent OEMs have just told the market they are competing on the AI overlay on top of the installed base rather than on a clean new platform. ABB and NVIDIA disclosed the RobotStudio HyperReality partnership in March on the same architectural premise. Yaskawa and KUKA were named in the same NVIDIA GTC partner set in March. Fanuc, ABB, Yaskawa, and KUKA together operate well over two million installed robots worldwide and have now committed to converting that install base into a physical AI deployment surface. The participation profile for a founder building a new robotics or autonomy hardware platform changed the week the incumbent install bases became the AI distribution channel. If you are building a new robotic arm or a new manipulator from scratch, you are now competing not against the next-generation arm the incumbent will ship, but against the AI overlay the incumbent is rolling out across the two million arms already on factory floors. The architectural decision a founder makes between competing with the incumbent stack and integrating into the incumbent stack is the decision that will determine whether the platform reads against the deployment channel the install base now is, or against a deployment channel the install base no longer feeds. ## Figure AI’s 50-Hour Run Resets the Reliability Bar for Autonomy Three Figure AI humanoid robots running the company’s in-house Helix-02 neural network sorted small packages across roughly 50 continuous hours with no teleoperation and no human intervention, with the run livestreamed and the trial extending well beyond the original eight-hour target, per Bloomberg coverage on May 15 and additional reporting in TechRadar and the New York Times. CEO Brett Adcock framed the milestone as “uncharted territory” for humanoid autonomy and emphasized that the robots ran entirely on onboard inference rather than on cloud-side or operator-side control. The structural detail is that the autonomy bar for industrial-grade humanoid deployment used to be measured in minutes of continuous unattended operation, and a public livestreamed run lasting two days under variable package geometries and pace constraints lifts that bar into a different category. Figure customers running active pilots like BMW at Leipzig and X1 at Plano will read the run as commercial evidence that the platform can hold an industrial shift profile, and prospective customers comparing humanoid options against Agility Digit at Toyota Canada or Apptronik Apollo at Mercedes will read it as a forcing function on every competing platform’s reliability disclosure cadence. For autonomy, robotics, and physical AI founders building toward a production deployment, the structural read is that the customer-facing reliability disclosure cadence has been reset by a public competitor. The reliability conversation customers run with a humanoid or autonomous platform vendor now starts from a 50-hour livestream as the public benchmark, and a vendor whose private reliability data does not approach that benchmark will be read as behind. The architectural work that produces a 50-hour continuous-operation profile is not bolted on the year before commercial pilots open. It is the onboard inference stack, the perception model robustness, the autonomous-reset behavior, the fault-recovery cadence, and the data-capture loop that closes after each unattended run, and that work has to be in motion across the build phase that precedes the commercial conversation. The founders whose platforms can run a comparable disclosure cadence twelve months from now are the founders whose architectural decisions across the next four quarters are aimed at the reliability bar Figure just set in public, not at the bar that existed last month. ## HII’s HYPR Program Shows Physical AI on a National Industrial Base Line Item HII signed an MOU with Path Robotics in February to integrate Path’s Obsidian physical AI welding model into Navy shipbuilding operations, and at Sea-Air-Space 2026 in April HII expanded the engagement into the High-Yield Production Robotics program with Path Robotics and GrayMatter Robotics, per HII’s newsroom release and Globe Newswire on April 20. HYPR combines robotic welding, automated material movement, autonomous surface treatment, and autonomous quality checks into an assembly line aimed at lifting Navy shipbuilding throughput another 15 percent on top of the 14 percent gain HII reported in 2025. The structural detail is that the physical AI partner stack inside HYPR sits on a national industrial base line item. HII is a publicly traded prime building submarines and surface combatants for the U.S. Navy, and the Pentagon FY27 budget request explicitly funds the industrial base capacity programs HYPR is designed to feed. Path Robotics is a physical AI company supplying the Obsidian foundation model for welding inside a production environment that the Pentagon’s capacity buy is going to underwrite at categorical scale, and GrayMatter Robotics is supplying surface treatment and quality automation into the same production envelope. For founders building physical AI, autonomy, robotics, or industrial automation hardware, the structural read on HYPR is that the prime-led production-line architecture is now a participation channel for physical AI companies inside the defense and dual-use industrial base. Path Robotics did not enter shipbuilding by winning a clean prime contract from the Navy. It entered by building the Obsidian welding model against a production problem HII was already solving, and the supplier conversation closed on the technical evidence that Path’s model could run inside HII’s production envelope at the operating cadence HII needed. The participation profile that wins the HYPR-class supplier conversation is not built on a generic capability pitch. It is built on a foundation model trained against a specific production problem, an integration profile that plugs into the prime’s existing line architecture, and a reliability disclosure cadence the prime can underwrite to its own customer. Founders running physical AI platforms aimed at the industrial base buy should be mapping the model architecture, the integration profile, and the reliability disclosure against the specific prime production problems that the FY27 industrial capacity budget is going to fund, because those prime conversations are happening now while the funding lines are being scaled. ## What the Three Signals Read Together Read across Fanuc-Google, Figure AI, and HII HYPR, the structural pattern is that physical AI has crossed three thresholds in the same week. The incumbent OEMs are converting installed bases into AI distribution channels, with Fanuc layering Gemini and Intrinsic over 1.1 million arms and ABB, Yaskawa, and KUKA operating against the same architectural premise. The reliability bar for autonomous operation has been lifted into a category that customers will now read every vendor against, with Figure’s 50-hour livestream as the public benchmark. The defense and dual-use industrial base is operationalizing physical AI inside prime-led production lines, with HII HYPR funded against the Pentagon’s industrial capacity buy. The three signals together tell hard-tech founders building robotics, autonomy, or industrial AI hardware that the commercialization environment has reorganized around three concrete distribution surfaces, and the operating profile that wins inside this environment is different from the operating profile that wins inside the prior environment. The founders who finish well in this environment design the participation profile against the three distribution surfaces that are now disclosed in public. For a founder building a new robotics platform, the choice between competing with the incumbent stack and integrating into the incumbent stack is the architectural decision the next four quarters will price. For a founder building an autonomy or humanoid platform, the reliability disclosure cadence and the onboard inference stack that produces it are the architectural decisions the next twelve months of customer conversations will read. For a founder building a physical AI foundation model or perception system, the prime-led production-line conversation is a participation channel that rewards model architecture trained against a specific production problem rather than a generic capability pitch. None of these participation profiles get assembled the month before the customer conversation opens. They get assembled across years of build-phase work, and the founders running that build phase against the prior commercialization framing will arrive at the rewired buy with a business the rewired buy is no longer reading. Dave’s take When three different parts of the physical AI stack move in the same week, it is rarely a coincidence. The incumbents, the foundation-model builders, and the prime industrial customers are all telling the supplier base the same thing, which is that physical AI is now a deployment problem rather than a research problem. The founders I am working with in robotics, autonomy, and industrial AI right now are running the same question across every architectural review: does the operating profile we are building today read against the install base, the reliability bar, and the prime production-line participation channel that the market just disclosed, or against an environment the market has now moved past? From Dave’s video library Dave reads Anthropic's founder playbook and finds the hard part did not vanish, it moved. Physical AI just made the same move in public: the research problem became a deployment problem, and deployment is where companies get built. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to The Build newsletter: [The Build →](/field-notes/the-build) --- # FDA Commissioner Resignation Reframes Regulatory Pathway Risk for Hard-Tech Founders | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-05-18-fda-commissioner-resignation-regulatory-pathway-risk [← All Field Notes](/field-notes) May 18, 2026 Regulatory FDA Commissioner Marty Makary resigned on May 12 after a 13-month tenure that closed with most of the agency’s senior career officials already gone, per CBS News, Fierce Pharma, and CNBC coverage. Deputy commissioner for food Kyle Diamantas is acting in the role. The transition lands on top of the April 1 decision letter denying Harrison.ai’s petition for a 510(k) exemption across radiology computer-aided detection, diagnosis, and triage software, in which CDRH explicitly rejected the argument that prior clearances demonstrate manufacturer competency for future devices, per AuntMinnie and Radiology Business coverage. Read together with the Aidoc $150 million raise on May 5 for a clinical AI foundation model and the cascading Medline Namic angiographic syringe recall affecting four kit manufacturers and four serious injuries reported by mid-March, the regulatory environment around hard-tech founders has moved into a different operating posture. Hard-tech founders running FDA pathways have to read the regulatory environment as the moving operating variable it now is, with the leadership transition, the AI pathway hardening, and the post-market exposure inside the convenience-kit supply chain all priced into the operating plan. ## The Makary Resignation and the CDRH Continuity Question Marty Makary resigned as FDA Commissioner on May 12 after a 13-month tenure that included a public clash with HHS Secretary Robert F. Kennedy Jr. over the agency’s authorization of fruit-flavored e-cigarettes, per coverage in CBS News, the Washington Post, NPR, CNBC, and Fierce Pharma. The agency’s senior career staff had already turned over substantially across the first year of the second Trump administration. Deputy commissioner for food Kyle Diamantas is the acting commissioner, and the permanent replacement process is open. The structural read for hard-tech founders running an FDA pathway is that the commissioner-level transition is the visible event, and the lower-visibility staffing question is the one that determines how a specific submission gets reviewed. CDRH, the device center, runs its review processes through a stack of branch chiefs, division directors, and center leadership whose continuity has been the operating variable a device sponsor reads when planning a submission timeline. The leadership transition at the top compounds with the senior-career turnover that already happened across the prior twelve months, and the device sponsor question becomes whether the specific review team handling a specific submission has the continuity to move the file at the cadence the sponsor’s operating plan was built against. For founders building medical devices, diagnostics, or any FDA-regulated hardware where a clearance or approval anchors the commercial conversation, the structural read is that the timeline error bar around an FDA submission widened the day the commissioner resigned. Founders with active De Novo, 510(k), or PMA submissions should pull the assigned reviewer history and the recent decision pattern out of their regulatory team this week, identify the specific review team handling the file, and update the operating plan’s timeline assumption against the actual continuity inside the team rather than against the publicly stated review-clock window. Founders preparing pre-submissions should expect the Q-Submission cadence to absorb the leadership transition, which means the pre-submission feedback may carry less weight on novel pathway questions than it would in a stable leadership environment. The first-time founder running an FDA pathway against an outdated timeline assumption will discover the slippage when the operating plan needs the clearance to anchor a financing or a launch, and the regulatory clock has moved against the plan. ## The Harrison.ai Denial Tells the AI Device Category What FDA Now Reads On April 1 the FDA denied Harrison.ai’s petition to partially exempt four classes of radiology AI software from premarket review, including computer-assisted diagnostic software, medical image analyzers, computer-aided triage and notification software, and computer-assisted detection and diagnosis software. The agency rejected the argument that holding a prior 510(k) demonstrates manufacturer competency for future devices, told the petitioner that CADe, CADx, and CADe/x devices cannot be treated interchangeably because they require different evaluation methodologies, and pointed petitioners toward predetermined change control plans and the Q-Submission Program as the alternative pathway, per coverage in AuntMinnie and Radiology Business. The structural detail is that CDRH used the denial to make the operating principle visible. The agency is not interested in the prior-clearance argument as a competency proxy for future devices, and the AI-enabled SaMD pathway is going to keep running through the full 510(k) or De Novo evaluation with the additional rigor that the predetermined change control plan and the Q-Submission program supply. Aidoc’s $150 million raise on May 5 for a clinical AI foundation model lands inside that exact regulatory frame, with the financing structured against a pathway that will keep going through the full 510(k) evaluation for each new indication. For founders building AI-enabled medical devices, diagnostics, or any SaMD platform, the structural read is that the regulatory posture on AI clearances is hardening rather than loosening, and the pathway design has to budget for the predetermined change control plan, the Q-Submission cadence, and the additional clinical validation work the agency is now asking AI sponsors to run. The founders who were modeling a shortened pathway against an exemption that did not arrive should revisit the operating plan against the April 1 decision letter and pull the regulatory submission strategy back into a full pathway with PCCP and Q-Sub embedded as load-bearing operating tools. The founders modeling against the full pathway from the beginning can move faster than the founders rebuilding the model now, and the capital plan has to support the longer pathway cycle that the agency just confirmed. ## The Medline Namic Recall Cascade Shows How Post-Market Risk Now Travels Aligned Medical Solutions initiated a nationwide recall on April 2 of its AMS6908E and AMS6908F Angio Pack convenience kits containing Medline Namic Angiographic Rotating Adapter syringes, after the rotating adapter mechanism was identified as a source of unwinding during use that could cause loose or full disconnection between the syringes and the manifold, with biohazard exposure, blood loss, infection, or air embolism among the listed risks. Medical Action Industries and American Contract Systems issued matching urgent recalls of their own convenience kits containing the same syringe component, and Medline reported four serious injuries and no deaths associated with the issue as of March 13, per FDA recall notices and Becker’s Hospital Review. The structural detail is that the post-market exposure inside a single component traveled across four kit manufacturers in a matter of weeks, with each downstream packager carrying a notification obligation and an inventory remediation cost the founder operating plan did not price into the convenience-kit revenue line. For founders building interventional devices, diagnostics, surgical consumables, or any hardware that ships inside a convenience kit or as a kitted component, the structural read is that the post-market exposure now travels through the supplier-customer graph faster than the pre-market evidence does. The cascade across Aligned Medical, Medical Action, and American Contract Systems shows what it looks like when a component supplier triggers a recall that the downstream packagers have to absorb on their own quality and notification systems, and the operating cost of running that absorption shows up in the gross margin line for whichever quarter the recall lands in. Founders selling into the convenience-kit channel should review the supplier-side quality agreement and the downstream notification cadence in the next 30 days, with the specific question of whether the operating plan can absorb a recall-cascade event without rewiring the gross margin trajectory the financing was sized against. The founders whose contracts and quality systems were written for the prior-decade exposure profile are running a balance-sheet risk the post-market environment is now pricing more aggressively than the pre-market environment ever did. ## What the Three Signals Read Together for Hard-Tech Founders Read across the Makary resignation, the Harrison.ai denial, and the Medline recall cascade, the regulatory operating environment around hard-tech founders has tightened along three axes in the same six-week window. The leadership transition widens the timeline error bar on every active submission, the AI exemption denial confirms that the AI device pathway will keep running through the full evaluation framework with PCCP and Q-Sub as load-bearing tools, and the recall cascade shows that the post-market exposure now travels through the supplier-customer graph at a cadence the operating plan has to be sized against. The combined signal is not that the regulatory environment is hostile. It is that the regulatory environment is moving, and the founders who treat the regulatory plan as a static assumption locked in at the seed-stage operating model are running a different business from the founders who treat the regulatory plan as a quarterly operating variable. The founders running the second version pull the regulatory team into the operating cadence on a monthly review rhythm, hold the timeline, pathway, and post-market exposure assumptions against the most recent agency signal, and update the operating plan when a signal moves any of the three assumptions. The founders running the first version arrive at the financing or the launch with an operating plan the agency has already moved past, and the cost shows up at the moment the operating plan needs the regulatory assumption to be true. The participation profile inside a hard-tech business that runs the regulatory environment as an operating variable is different from the profile inside a business that runs it as a static assumption. The first business has a regulatory team embedded in the operating cadence, a quarterly review of the submission timeline against actual reviewer continuity, a pathway plan that prices PCCP and Q-Sub as load-bearing tools rather than as fallback options, and a post-market quality and notification architecture sized for the supplier-customer cascade exposure the current environment is producing. The second business has a regulatory consultant who delivers a strategy deck at the seed stage, a timeline assumption that has not been pressure-tested against the leadership transition, a pathway plan that did not budget for the PCCP work the AI device category now requires, and a post-market architecture sized for a prior decade. The first business can absorb a Makary-class event without rebuilding the operating plan. The second business cannot, and the moment the operating plan needs the regulatory assumption to be true is the moment the cost arrives. Dave’s take The hard-tech founders I work with treat the FDA as one of the operating environments their business runs inside, not as a process the business will encounter at a specific milestone. The Makary resignation is a single data point, the April 1 Harrison.ai denial is another, and the Medline recall cascade is a third, but the operating principle is the same across all three. The regulatory environment is moving on a quarterly cadence right now, and a founder who built the company against an operating plan that assumed it would be static is running a business different from the one the next twelve months are going to test. From Dave’s video library Dave on reading a dramatic government decision for what is real and what is noise. The FDA transition deserves the same read: a moving regulatory environment is an operating variable, not a milestone you schedule around. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to The Build newsletter: [The Build →](/field-notes/the-build) --- # Boston Scientific's MiRus Stake Marks MedTech's Milestone Era | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-05-19-boston-scientific-mirus-tavr-milestone-deal-structure [← All Field Notes](/field-notes) May 19, 2026 Commercialization Boston Scientific took a 34% equity stake in MiRus on May 18 for $1.5 billion in cash, with an exclusive option to acquire the rest of the company’s TAVR business for up to $3 billion in milestone-gated payments tied to the SIEGEL valve’s clinical and regulatory progress, per MedTech Dive and the company’s press release. Read alongside Stryker’s $835 million Amplitude Vascular close on May 7 (intravascular lithotripsy, $435 million cash plus $400 million in milestones), Embecta’s £150 million Owen Mumford close on May 15 (drug delivery, £100 million upfront plus £50 million in milestones), HistoSonics’s De Novo submission for kidney tumors on May 11, and the CMS-FDA RAPID Coverage Pathway now being operationalized, the MedTech commercialization environment has moved into a different operating posture. Large acquirers and the payer side are pricing the specific clinical and regulatory milestones a device hits, not the finished asset on the shelf. For hard-tech founders running a commercialization plan against an FDA pathway, the operating implication is that milestone delivery has become the unit of value the buyer and payer side now negotiate against. ## Boston Scientific Bought Optionality, Not a Finished TAVR Boston Scientific announced on May 18 that it had taken a 34% equity stake in MiRus for $1.5 billion in cash, with an exclusive option to acquire the rest of the TAVR business for an additional $3 billion in milestone-gated payments. The deal also includes options on MiRus’s mitral and tricuspid programs. MiRus’s SIEGEL valve is a nickel-free balloon-expandable transcatheter aortic valve replacement platform built on a rhenium alloy, currently in the STAR pivotal trial across low, intermediate, and high surgical-risk cohorts. This is Boston Scientific’s third TAVR attempt. The Lotus Edge program shut down in 2020 on delivery-system issues, the Acurate Neo2 program ended in 2025 after failing noninferiority against Edwards Sapien and Medtronic Evolut, and SIEGEL is the structured re-entry. The company is paying for optionality on a clinical and regulatory pathway, and the price of that optionality is sized against the risk that the next STAR readouts move against MiRus rather than for the asset that exists today. For founders building any platform where the commercial conversation with a strategic acquirer is on the table inside the next 24 months, the implication is that the deal structure has migrated from straight purchase price toward milestone-tied staged payments. What the acquirer pays for now resolves into a finished asset after a sequence of clinical and regulatory milestones that the term sheet specifies, and the operating plan that wins the deal is the one that makes those milestones legible, fundable, and verifiable to an external auditor. A surgical robotics founder, an interventional device founder, a diagnostics founder, or a defense-hardware founder negotiating with a tier-one acquirer inside this environment is going to encounter the same milestone-gated structure Boston Scientific used on MiRus, and a founder who built the operating model against a single-price-tag exit is going to be disappointed when the term sheet arrives. ## Stryker and Embecta Confirm the Pattern, Not the Outlier Stryker completed its acquisition of Amplitude Vascular Systems on May 7 for $435 million in cash plus up to $400 million in milestone payments, adding a CO₂-generated pressure-wave intravascular lithotripsy platform to its peripheral vascular portfolio, per the Stryker press release and Cardiovascular Business coverage. Embecta closed its acquisition of Owen Mumford on May 15 for £100 million upfront plus up to £50 million in milestone payments tied to net sales of the Aidaptus next-generation auto-injector platform over the three-year window post-close, per the Embecta SEC 8-K. The headline value of the Stryker deal is roughly 52% upfront, the headline value of the Embecta deal is roughly 67% upfront, and across both the structure is the same posture Boston Scientific used on MiRus. A fraction of the headline number transfers against the asset that exists today, and the remainder transfers as the target hits the specific clinical, regulatory, or commercial milestones the term sheet enumerates. Three deals inside ten business days tells the founder reading along that this is not a one-off structure designed for an unusual asset. It is the operating posture of the MedTech buyer side in May 2026. Founders selling a hard-tech business inside this environment should expect milestone-tied payment structure across roughly 30 to 50 percent of the headline value, with the milestones tied to clinical readouts, regulatory authorizations, manufacturing-yield commitments, and net-sales tranches the buyer can audit. The operating implication is that the founder team has to model the post-deal milestone schedule with the same operational rigor it uses to manage the FDA pathway today, because the milestone schedule is the cash-flow schedule for the founder, the early employees, and the equity holders who own the upside that has not closed yet. ## HistoSonics Shows What Platform-Milestone Sequencing Looks Like HistoSonics submitted a De Novo request to the FDA on May 11 seeking authorization to expand the Edison Histotripsy System into kidney tumor destruction, with clinical evidence from the 67-patient #HOPE4KIDNEY trial led by William Huang at NYU Langone Health, per Endovascular Today, Urology Times, and MedTech Dive. The Edison platform already holds a 2023 De Novo authorization for liver tumor destruction, and kidney is the second major indication the company is pursuing. The structural detail is that a single platform technology with a credible clinical-evidence trail becomes a multi-indication asset, where each subsequent indication is a fundable milestone the operating plan can sequence against rather than a single regulatory event that anchors the entire business. For founders building any platform technology where the underlying mechanism can address multiple clinical indications, the HistoSonics submission is a useful reference point for staging the regulatory pathway across indications. Each subsequent indication becomes its own commercial inflection point separate from the initial authorization, and the financing market prices each one against the readout that anchors it. The operating implication is that the platform plan should identify the secondary and tertiary indications inside the seed-stage operating model as a sequence of fundable milestones the company will deliver against once the first authorization arrives, rather than as a single-paragraph appendix at the end of a fundraising deck. ## RAPID Coverage Pathway Puts the Payer Side on Milestone Time The CMS-FDA RAPID Coverage Pathway, announced on April 23 and now operationalized, compresses the gap between FDA marketing authorization and a Medicare National Coverage Determination to 60 to 90 days, against a prior baseline that often ran past six months under the predecessor TCET program, per analysis from Latham & Watkins. The pathway is sized for roughly 40 currently qualifying breakthrough devices with up to 20 additional eligible, against TCET’s roughly five-per-year capacity. The proposed coverage determination is issued the same day as FDA marketing authorization, with the final determination 60 to 90 days later. The payer side is using the same milestone logic the acquirer side is using, with reimbursement timed against a specific FDA milestone and the founder operating plan now able to price the reimbursement event with materially less timeline uncertainty than the prior decade allowed. For founders building breakthrough-designated devices in any hard-tech category, the implication is that the regulatory and reimbursement events have collapsed into a single milestone window. The operating plan that treats FDA authorization and Medicare coverage as two separate inflection points is running against the prior operating reality. The plan that treats them as a single milestone, with the financing, inventory build, and commercial team sized against the joint event, is running against the operating reality the RAPID pathway just created. The founders who reorganize the operating plan against the joint milestone move materially faster through the commercialization curve than the founders running the legacy assumption. ## What the Four Signals Read Together for Hard-Tech Founders Read across the four signals, the MedTech environment has moved into a milestone-priced operating posture. Acquirers are pricing milestones inside the deal structure. Payers are pricing milestones inside the reimbursement timeline. Platform companies are sequencing milestones as the unit of value across multiple indications. The deepest credential inside this analysis is interventional cardiology and surgical robotics, but the milestone logic is going to land identically on a diagnostics platform, a defense-hardware platform, a climate-hardware platform, or an industrial-robotics platform once the acquirer side or the payer side starts pricing them the same way. The founders who built the company against a finished-asset assumption have to revisit the operating plan against the milestone-priced reality, and the founders who already structure the company against a quarterly milestone cadence are operating against the environment as it now is. The participation profile inside a hard-tech business that runs the milestone posture is different from the profile inside a business that runs the finished-asset posture. The first carries a board pack that displays the next four to six fundable milestones, an operating plan that prices each milestone against the cash-flow and financing event it anchors, and a sales-side conversation with the strategic acquirer that pre-frames the milestone structure before the term sheet arrives. The second carries a deck with a single exit number, an operating plan sized against a single revenue ramp, and a sales-side conversation that lands at the term-sheet table without the milestone language already negotiated. The first business runs faster. The second business runs slower, and the cost of the slower run shows up at every financing window between now and the exit. Dave’s take The shift I see in the term sheets and operating plans coming across my desk this month is that the milestone structure has become the operating language of MedTech commercialization, and the founders who have not built the company around it are negotiating in a vocabulary the buyer side already moved past. I’d rather see a hard-tech founder spend the next 90 days rebuilding the operating plan against milestone-priced cash flow than spend the same 90 days polishing a deck for a finished-asset valuation conversation that is no longer happening. From Dave’s video library Dave on why buyers pay for proof, not promises. MedTech deal structures now say it in the term sheet: Boston Scientific priced MiRus as milestones to be earned, not a finished asset to be bought. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Boston Scientific SEISMIQ IVL Pivotal Data Hits Endpoints | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-05-20-boston-scientific-seismiq-ivl-fracture-pivotal-evidence [← All Field Notes](/field-notes) May 20, 2026 Interventional Boston Scientific reported on May 19 that the FRACTURE IDE pivotal trial of the SEISMIQ 4CE coronary intravascular lithotripsy catheter hit both primary endpoints at the EuroPCR late-breaking session in Paris, with 93.3% freedom from major adverse cardiac events at 30 days against an 86.2% performance goal and 93.7% procedural success against an 85.8% goal across 420 patients at 46 sites, per the company press release and Medical Device Network. Endologix announced on May 19 that it acquired the Pounce thrombectomy platform from Surmodics in a carve-out that bundles the LP, standard, and XL configurations with a core Surmodics team transferring to Endologix, per the Orange County Business Journal and Yahoo Finance, terms not disclosed. Lauxera Capital Partners closed Lauxera Growth II at €520 million the same day, sized for 12 to 15 European healthtech companies pursuing US expansion with check sizes of €20 to €50 million each, per the firm press release and Fierce Biotech. Read together, the three signals describe what pivotal-grade clinical evidence, focused commercial scope, and growth capital can actually buy a hard-tech founder in interventional MedTech right now. ## SEISMIQ Bought a Credible Challenge to Shockwave’s Calcium Franchise The FRACTURE IDE trial enrolled 420 patients with severe calcified coronary artery disease across 46 sites in the United States and Europe, ran a prospective non-randomized single-arm design with 2-year follow-up, and met its prespecified safety and effectiveness performance goals with p-values below 0.0001 on each endpoint, per the May 19 Boston Scientific announcement and the Medical Device Network coverage. The safety endpoint registered 93.3% freedom from MACE at 30 days against an 86.2% pre-specified goal. The effectiveness endpoint registered 93.7% procedural success against an 85.8% goal, with 100% successful stent delivery and 94.2% average stent expansion at the most calcified segment. Chief medical officer Janar Sathananthan called the data pivotal evidence for the company’s regulatory submission, with the peripheral version of the SEISMIQ platform already cleared by FDA in 2025 and the coronary submission now staged on the strength of the FRACTURE readout. For an interventional device founder building any platform that has to compete head-to-head against a category incumbent the size of Shockwave (J&J) for the calcium-modification market, the SEISMIQ trial design is the reference template to read against your own IDE plan. A single-arm pivotal with pre-specified performance goals derived from the comparator’s historical evidence base, paired with a 30-day MACE endpoint and a procedural-success effectiveness endpoint, is the cleanest path from clinical evidence to FDA submission inside interventional cardiology in 2026. The trial doesn’t have to randomize against the incumbent, doesn’t have to demonstrate superiority, and doesn’t have to enroll the multi-thousand-patient cohort that a randomized non-inferiority study would require. It has to hit the pre-specified performance goal with a margin the agency reads as convincing, and it has to enroll across enough geographic and operator diversity that the result generalizes beyond a single high-volume center. ## Endologix and Surmodics Show the Specialist Carve-Out Geometry The Endologix purchase of the Pounce thrombectomy system from Surmodics, announced May 19, is the kind of transaction a founder building a peripheral or interventional device should study closely whether the founder is building toward an acquirer or building inside the acquirer’s adjacent footprint. Endologix CEO John Liddicoat described the Pounce platform as a strong strategic and clinical fit for the company’s vascular intervention portfolio, which already includes the Detour, Alto, and AFX2 platforms. The Pounce platform is FDA-cleared, ships across three configurations (LP, standard, XL), uses a dual-basket and nitinol collection funnel mechanism that doesn’t require capital equipment or aspiration, and brings a core Surmodics team to Endologix as part of the transition. Two readable implications for founders. First, the carve-out geometry of this deal tells you that a peripheral or interventional asset can move between a diversified medical-coatings company (Surmodics) that is no longer the natural commercial owner and a focused vascular-intervention operator (Endologix) where the asset fits the existing rep bag, the existing call point, and the existing post-market commitment. If you are building a specialty device that lands inside a strategic acquirer’s adjacent rather than core portfolio, the post-close exit path can run through a carve-out to a smaller, more focused operator rather than through a hold on the original acquirer’s balance sheet. Second, the team transfer detail matters more than the term sheet usually surfaces. A core operator team moving with the asset preserves the post-market quality system, the clinical-affairs continuity, and the commercial relationships the new owner needs to ramp without breakage. If you are negotiating a carve-out from the buyer or the seller side, the team-transfer clause is closer to the gross-margin trajectory than the headline price. ## Lauxera’s €520M Tells You What Growth Capital Now Pays For Lauxera Capital Partners’ close of its second fund at €520 million ($605 million), announced May 19, sets the European healthtech growth-equity check size at €20 to €50 million across 12 to 15 portfolio companies, with an explicit thesis around helping European companies expand into the US market. The firm, based in Paris and San Francisco, has already deployed twice from Fund II: a minority investment in German neurovascular device maker Acandis and a majority position in Swedish advanced-imaging CRO Antaros Medical, per the firm announcement and reporting in Ventureburn and Fierce Biotech. The fund covers commercial-stage healthtech across medical devices, pharma services, digital health, healthcare data and software, and life-science tools and diagnostics. For a European hard-tech founder running a US expansion plan, the Lauxera close confirms that growth equity sized for the US commercial ramp is available at meaningful scale, with a fund partner that has already structured the US-from-Europe pathway across multiple investments. The check size is large enough to fund the US regulatory submission, the initial US commercial buildout, and the first 12 to 18 months of the US sales motion without dilution against a US-only seed round. For a US founder competing against a European entrant in the same indication, the same close tells you the European competitor is going to arrive at the US market with growth capital sized for the ramp, with a strategic partner that has navigated the FDA pathway and the US payer side across multiple prior deals, and with an operating template the US entrant has to compete against on more than first-mover advantage alone. ## What Pivotal Evidence Buys You in 2026 Interventional MedTech Read across the three signals, pivotal evidence is the structural input that unlocks each of the downstream commercial events the operating plan is sized against. Boston Scientific’s SEISMIQ data unlocks the FDA submission and the calcium-franchise commercial conversation against Shockwave. Endologix’s Pounce purchase rests on an FDA-cleared platform with the post-market evidence base Surmodics built over years of commercial use. Lauxera’s growth capital prices European companies that already have CE-mark commercial-stage evidence and are now staging US pivotal trials or post-market US ramps. In each case the unit of value that opens the next commercial door is the clinical evidence base, sized and structured to read against the buyer-side or agency-side audit the next conversation will run. For founders building interventional, peripheral, surgical-robotics, neurovascular, or diagnostics platforms inside this environment, the operating implication is that the IDE and post-market evidence plan is closer to the financing plan than the deck usually treats it as. A pivotal trial designed to hit a pre-specified performance goal with a margin the agency reads as convincing is the precursor to the FDA submission, the precursor to the buyer-side conversation, and the precursor to the growth-equity round. A pivotal trial designed against the wrong performance goal, the wrong endpoint, the wrong follow-up window, or the wrong site mix delivers a readout the buyer side, the agency side, and the payer side all read differently than the operating plan modeled, and the cost of the difference shows up at every commercial and capital window between now and the exit. Dave’s take The FRACTURE readout is the cleanest 2026 example of how a well-designed single-arm pivotal can carry an interventional device founder through the FDA submission, the calcium-franchise commercial fight, and the buyer-side conversation in one sequence. I’d rather spend a full quarter with a founder rebuilding the IDE design against the agency’s read of the comparator’s evidence base than spend the same quarter polishing a deck for a round where the pivotal design is still wrong. The pivotal you ran is the pivotal the buyer side, the payer side, and the agency side all read, and there is no second draft. From Dave’s video library Dave on the one kind of proof nobody can argue with. The FRACTURE readout is that proof in interventional form: a pivotal designed well enough to carry the FDA case, the commercial fight, and the buyer conversation in one sequence. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Surgical Robotics Challenger Field Fills Out in 2026 | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-05-21-surgical-robotics-challenger-field-2026-launch-architecture [← All Field Notes](/field-notes) May 21, 2026 Surgical Robotics The soft-tissue surgical robotics market filled out across the first half of 2026. Intuitive Surgical reported Q1 2026 revenue of $2.77 billion (up 23% year-over-year), placed 431 da Vinci systems including 232 next-gen da Vinci 5 units, and grew the installed base to 11,395 systems with procedure volume up about 17%, per the company release and Modern Healthcare coverage. Medtronic completed its first US Hugo case at Cleveland Clinic in February with Duke and Atrium Health Wake Forest Baptist High Point already among the first US installs, per MedTech Dive. Johnson & Johnson reported in May that the Ottava bariatric pivotal trial met its primary safety and performance endpoints across 30 patients with all procedures completed robotically, building on the January FDA de novo submission, per Mass Device and the company press release. And CMR Surgical CEO Massimiliano Colella disclosed in a March MedTech Dive interview that he halted the gen-one Versius US launch in late 2024 to wait for Versius Plus, which received 510(k) clearance in December 2025 and entered US soft launch in 2026. Read across the four signals, the surgical robotics race stopped being “Intuitive versus everyone” in 2026 and became a multi-player field where each challenger picked a different launch architecture. ## Intuitive Is Accelerating, Not Holding The single most important data point for any surgical robotics founder reading the 2026 landscape is that the incumbent is accelerating. Q1 2026 revenue of $2.77 billion against $2.25 billion in Q1 2025 is 23% growth at a $11 billion run rate, and the 431 da Vinci system placements compared to 367 a year earlier is a 17% installation increase against an already-dominant installed base. The 232 da Vinci 5 units inside that placement number is the more important figure. The next-generation system shipped in volume across the first quarter, the installed base grew to 11,395 worldwide (12% year over year), and worldwide procedure growth held at 17% with Ion bronchoscopy procedures growing 39%. The headline-level read is that Intuitive raised its 2026 procedure-growth guidance against this backdrop, not lowered it. If you are building a surgical robotics platform aimed at any indication da Vinci already covers, the operating implication is that the incumbent’s reference customer at year-end 2026 has a da Vinci 5, an upgrade path the incumbent is already executing, and a procedure-volume ramp that creates surgeon-side switching cost the new entrant has to overcome on more than parity alone. The pitch that the new platform is “da Vinci-equivalent at a lower price” was a viable thesis when da Vinci installations were tracking 350 systems a quarter and the incumbent platform was Xi or earlier. It is a thinner thesis when the incumbent is placing 232 next-gen units a quarter and growing procedures on the prior-generation install base at the same time. The differentiated value proposition has to live somewhere the incumbent isn’t already executing on. ## Medtronic Picked Indication Sequencing Over Big-Bang Launch Medtronic’s Hugo platform completed its first US commercial surgical case at Cleveland Clinic in February 2026, with Jihad Kaouk performing a prostatectomy and the patient discharged the next day, per the company announcement and MedTech Dive. Duke University Hospital and Atrium Health Wake Forest Baptist High Point are among the first US installs alongside Cleveland Clinic. The FDA clearance covered urology indications first; gynecology and general surgery (including hernia repair) are the next expansion targets, with the urology and hernia trials already reporting on primary safety and effectiveness endpoints. The platform ships across 30+ countries and has been used in tens of thousands of urology, gynecology, and general surgery procedures outside the US. For a soft-tissue surgical robotics founder, the Medtronic launch architecture is the most readable template for a credible challenger entering against da Vinci with the deep-pocketed-incumbent commercial-engine playbook. Medtronic chose to lead with urology (the narrowest indication, the one with the cleanest comparator evidence base, and the one Hugo had the longest commercial-use history in outside the US), to anchor the launch on academic-medical-center reference customers (Cleveland Clinic, Duke, Wake Forest), and to stage indication expansion behind already-completed pivotal trials in gynecology and general surgery rather than against a single multi-indication submission. The sequencing matters because the reference-customer base built across the urology launch is the channel the gynecology and general-surgery expansion will run through. Founders running a smaller-balance-sheet version of the same architecture should size the urology launch against the reference-customer evidence base that will carry the next two indications, not against the urology revenue line alone. ## J&J Ran the Pivotal Trial Where the Architecture Question Lives Johnson & Johnson’s Ottava bariatric pivotal trial, reported in May 2026 at the ASMBS Annual Meeting, met its primary safety and performance endpoints across 30 Roux-en-Y gastric bypass patients with all procedures completed robotically on the platform and no conversion to a non-robotic approach. Average weight loss at 30 days was 30 pounds. The detail that matters more than the headline endpoint is geographic: in five of the six study sites, procedures were performed in operating rooms that had not previously been used for robotic surgery, including rooms historically considered too space-constrained for the boom-mounted form factor da Vinci runs on. J&J submitted Ottava to FDA for de novo classification in January 2026 covering multiple upper-abdomen general-surgery procedures including gastric bypass, gastric sleeve, small-bowel resection, and hiatal hernia repair. The Ottava launch architecture chose a different lane from Medtronic’s. The platform’s differentiated value sits in the table-integrated robotic-arm form factor that opens the addressable OR base from the high-end purpose-built robotic suite to the broader hospital and ASC inventory of space-constrained ORs. The pivotal trial was designed to demonstrate not just safety and effectiveness in bariatric surgery but specifically that the platform performs in the OR environments the incumbent platform cannot operate in. For a surgical robotics founder whose differentiated value proposition is form-factor or site-of-care driven (ASC-ready footprint, mobile deployment, smaller OR compatible), the Ottava trial design is the reference template for staging the pivotal readout to validate the architecture question, not just the clinical claim. ## CMR Surgical Bought Discipline by Waiting CMR Surgical CEO Massimiliano Colella, appointed permanent CEO in January 2025, told MedTech Dive in March that he halted the planned US launch of the first-generation Versius robot when he took over, choosing to wait for Versius Plus rather than enter the US market with gen-one hardware. Versius Plus received 510(k) clearance for cholecystectomy in December 2025, the company submitted an additional 510(k) for gynecology in April 2026, and the US soft launch is underway in 2026 with select hospital partners. Outside the US, Versius is the second-most-adopted soft-tissue robotic platform with 40,000-plus procedures across 30-plus countries. Colella described the soft-tissue robotics market as one where the company cannot afford a single misstep on the US entry. For a hard-tech founder running a US market entry plan against an entrenched incumbent, the CMR decision to delay launch by 12 to 18 months in exchange for shipping gen-two rather than gen-one hardware is the case study that should sit on top of the operating-plan stack. The cost of delaying the US revenue line by a year was real and visible to the board, the investors, and the global commercial team. The cost of launching gen-one against an incumbent who would have read the gen-one feature set against da Vinci’s gen-five reference standard would have been larger and structural. Operating discipline at the launch-timing decision was the call that made the US launch survivable. Founders who reach the same crossroads on a smaller balance sheet should treat the operating cost of delay as the price of preserving the right to compete, not as a failure of speed. Dave’s take The most useful read across the four 2026 surgical robotics signals is that there is no single right launch architecture against da Vinci. Medtronic ran indication sequencing. J&J ran a form-factor-driven pivotal trial against space-constrained ORs. CMR Surgical waited a year and shipped gen-two. Each one is a credible architecture, and the founder lesson is that the company has to pick the architecture first, then design the pivotal trial, the reference-customer base, and the commercial sequence against the architecture. The companies that get into trouble are the ones that pick the architecture last, after the pivotal is already in the field. I have seen that mistake compound across two and three financing rounds, and it is the one structural decision I would spend the first quarter of a Fractional CPO engagement getting settled. From Dave’s video library Dave on why one coherent bet beats five plausible ones. Four challengers are attacking da Vinci with four different launch architectures, and the lesson is not which one is right, it is that each company committed to exactly one. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Guardant360 Liquid CDx and the Companion Diagnostic Platform Moat | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-05-22-guardant-liquid-cdx-companion-diagnostic-platform-moat [← All Field Notes](/field-notes) May 22, 2026 Diagnostics The FDA approved Guardant Health’s new Guardant360 Liquid CDx on May 20, the largest FDA-approved liquid biopsy panel to date with what Guardant described as a roughly 100x wider genomic footprint than the prior Guardant360 CDx, per the company press release. Seven previously cleared companion-diagnostic indications transfer to the new test, results are returned in as little as seven days, and the platform integrates genomic and epigenomic profiling from a single blood draw, per the company release and labmedica.com coverage. The approval landed two and a half weeks after the FDA approved Guardant360 CDx as a companion diagnostic for Arvinas and Pfizer’s vepdegestrant (VEPPANU) in ER+/HER2- advanced breast cancer with ESR1 mutations on May 4, Guardant’s third ESR1 CDx clearance, per the BioSpace and Guardant Health investor-relations release. Three weeks earlier on April 30, Guardant announced a multi-year strategic CDx collaboration with Nuvalent to support targeted-cancer-therapy commercialization on the Guardant Infinity platform, per the BusinessWire announcement. Read together, the three signals are about more than a single product launch. The diagnostic platform moat in 2026 is built out of pharma-partnership depth, not panel size. ## What the New Guardant360 Liquid CDx Actually Carries The headline number is the 100x genomic-footprint expansion against the prior Guardant360 CDx, but the load-bearing data point for any diagnostics founder is what the new approval inherits. Seven companion-diagnostic indications that were already cleared on the prior Guardant360 CDx transfer to the new test, per the Guardant press release and Yahoo Finance coverage. That means the new test enters the market on the day of approval with a working CDx footprint across multiple non-small-cell lung cancer and colorectal cancer indications, plus the only FDA-approved CDx for ESR1-mutated advanced breast cancer, rather than entering with the empty label that a brand-new approval would normally carry. The Smart Platform integrates genomic and epigenomic profiling from a single blood draw and delivers what Guardant describes as a several-fold sensitivity improvement for circulating tumor DNA detection. Results are returned in as little as seven days regardless of tissue availability, line of therapy, or care setting, per the Morningstar and labmedica.com coverage. If you are building an IVD startup positioning against an entrenched incumbent CDx platform, the load-bearing read is that the regulatory work the incumbent did on prior generations of the product compounds into the next generation. The cleared CDx indications carry forward, the pharma-partner relationships those indications represent stay attached, and the prior trial data feeds the labeling work for the next launch. The lesson for a diagnostics founder running a much smaller capital base is to design the first CDx clearance with the second-generation transfer already in mind, so that the regulatory effort on version one is structurally reusable rather than an isolated launch milestone the team has to repeat from zero on the next product. ## The VEPPANU CDx Is a Second Data Point on the Same Moat On May 4, the FDA approved Guardant360 CDx as a companion diagnostic for Arvinas and Pfizer’s vepdegestrant (VEPPANU) in adults with ER+/HER2- advanced or metastatic breast cancer with ESR1 mutations and disease progression after at least one line of endocrine therapy, per the BioSpace and Guardant investor-relations announcement. Vepdegestrant is a PROTAC heterobifunctional protein degrader co-developed by Arvinas and Pfizer, and the approval makes Guardant’s test the only FDA-approved CDx for targeted therapy in ESR1-mutated advanced breast cancer. The same release noted that this was Guardant’s third ESR1 CDx clearance, the previous two attached to other therapies in the same disease segment. For a diagnostics founder, the structural point is that Guardant has accumulated three FDA-cleared CDx labels in a single mutation segment across multiple therapies and multiple pharma partners. Each new approval makes the platform harder to displace by a single-shot competitor, because the competing test has to clear not one but three CDx labels and replicate the multi-partner pharma relationships behind them. The diagnostic moat compounds in the same way the surgical-robotics installed-base moat compounds. The first cleared CDx is a label. The third cleared CDx in the same segment is a category position the next entrant has to displace as a whole. If you are building a competing diagnostic, the segment you choose has to be one where you can credibly accumulate at least three CDx clearances across multiple pharma partners on the way to the fourth, not one where you are betting on a single CDx clearance to carry the company. ## The Nuvalent CDx Collaboration Tells You How the Moat Gets Built On April 30, Guardant Health announced a multi-year strategic CDx collaboration with Nuvalent to support development and potential commercialization of Nuvalent’s targeted-cancer-therapy pipeline on the Guardant Infinity platform, per the BusinessWire announcement. The agreement covers CDx development across the Nuvalent pipeline rather than a single therapy. For a hard-tech founder reading the structure of the deal, the takeaway is that the Guardant moat is not built one CDx clearance at a time. It is built through multi-year pharma agreements that commit pipeline-level CDx work to the platform, generate the regulatory submissions that feed the next round of label expansion, and lock the platform in as the preferred CDx reference at the pharma partner across the next several FDA approvals. The translation for a diagnostics founder is direct. The CDx business model is not a transactional sale of an individual test to an individual pharma program. It is a pipeline-level partnership that turns each pharma collaboration into multiple cleared CDx labels across the next three to five years. Founders building a competing platform should design the first pharma partnership as a multi-year, pipeline-level commitment from day one, not a single-program engagement that has to be renegotiated for each therapy. The cost of negotiating one pipeline-level partnership is roughly the same as negotiating one program-level partnership, and the difference between the two structures shows up as label volume and reference-customer depth across the entire build phase. ## Q1 2026 Medtech Turbulence Is the Background The diagnostic platform moat reads especially clearly against the Q1 2026 medtech earnings season MDDI characterized as a quarter that analysts wanted to leave in the rearview mirror. Boston Scientific and Abbott both cut full-year 2026 guidance despite Q1 beats, per MedTech Dive and MassDevice coverage. Even companies delivering beats and raises faced steep sell-offs as the bar for investor satisfaction climbed, with pulsed-field ablation and diabetes technology hit particularly hard. Johnson & Johnson’s cardiovascular medical-device unit was the only major medtech segment to grow by double digits in Q1. The diagnostic platform moat matters more in that environment, not less. When the buy-side is repricing growth and pricing structural defensibility into the multiple, the platform that ships into the next quarter with a 100x expanded panel, seven inherited CDx indications, three ESR1 CDx clearances, and a multi-year pipeline-level pharma partnership reads as a different category of asset than the IVD point-product competing on assay sensitivity alone. For a diagnostics or IVD founder pricing the next round in 2026, the structural defensibility argument has to be assembled the same way Guardant has assembled it: across multiple CDx clearances in the same segment, across multiple pharma-partner relationships, and across a multi-generation platform that lets each clearance carry forward to the next product launch. Dave’s take I read the three Guardant approvals across April and May as a single sustained demonstration of how a diagnostic platform moat is actually built. The 100x panel expansion is the headline. The seven inherited CDx labels, the third ESR1 CDx clearance, and the Nuvalent pipeline-level partnership are the structural assets. If I were sitting with the founder of a diagnostics company at the seed stage today, the conversation I would want to have first is which mutation or biomarker segment they intend to accumulate three cleared CDx labels in across the next 36 months, which pharma partners they intend to commit at the pipeline level rather than the program level, and how the regulatory work on the first clearance is designed so that the second and third clearances inherit rather than repeat it. Those three architectural decisions, made at the seed stage, are what produce the moat the buy-side reads at the deal stage. From Dave’s video library Dave on the unglamorous work that compounds after the first win. Guardant's moat was not the panel size in the press release, it was years of pharma partnership work stacking up underneath it. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Medtronic’s $650M SPR Deal and the Chronic Pain Platform Consolidation | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-05-23-medtronic-spr-chronic-pain-neuromodulation-platform [← All Field Notes](/field-notes) May 23, 2026 Medical Devices Medtronic announced on May 20 that it intends to acquire Cleveland-based SPR Therapeutics for approximately $650 million in cash upfront for all outstanding equity in the company, per the Medtronic newsroom release and MedTech Dive coverage. SPR’s Sprint peripheral nerve stimulation system was cleared by the FDA in 2018, has reached roughly 50,000 implants by the end of 2025, and delivers up to 60 days of temporary, percutaneous therapy without a permanent implant, per MassDevice and MedTech Dive. The deal closes the third leg of Medtronic’s 2026 buying run, following the $585 million CathWorks deal that closed in April and the $550 million Scientia Vascular agreement announced in March, for a combined roughly $1.78 billion across coronary detection, neurovascular, and chronic pain. Boston Scientific finished its own $533 million Nalu Medical acquisition earlier this year on the permanent-PNS side, sized against more than $60 million in 2025 sales and 25 percent year-over-year growth in 2026, per the Boston Scientific newsroom and bioxconomy.com. Read together, the two PNS deals describe a category that is consolidating around the temporary-versus-permanent therapy split inside an $11 billion 2026 neuromodulation market, per Precedence Research, and the structural read for a hard-tech founder building in pain management is that the platform decision now sits one level above device performance. ## What the SPR Deal Actually Adds to Medtronic Medtronic interim neuromodulation president Domenico De Paolis described the Sprint system in the Medtronic release as a way to broaden access to treatment by allowing physicians to intervene earlier in the chronic pain pathway, per MedTech Dive. The structural read is more specific than the access language. Sprint is a 60-day, percutaneous, no-implant therapy. Medtronic’s existing neuromodulation portfolio is anchored by permanent spinal cord stimulation and deep brain stimulation systems. The acquisition gives Medtronic a step-down therapy that sits in front of the implanted product, a referral funnel into the permanent device when the temporary therapy stops working, and a presence in pain-management practices that do not currently implant. The 50,000-implant install base SPR has already built is the entry ticket. The economic argument inside Medtronic is that owning the temporary-PNS shelf changes the cost of customer acquisition for the permanent neuromodulation portfolio, not that the standalone temporary-PNS business will return $650 million in revenue inside the next two years. ## Boston Scientific Picked the Other Side of the Same Category Boston Scientific announced the Nalu Medical acquisition in October 2025 and closed it earlier in 2026, with Nalu’s permanent peripheral nerve stimulation system slotting in next to the WaveWriter Alpha spinal cord stimulator and the basivertebral nerve ablation portfolio, per the Boston Scientific newsroom and Neuro News International. Nalu’s 24-month randomized data published in January 2026 showed 85 percent of patients in the active and crossover arms hitting at least a 50 percent reduction in pain and an average pain reduction of 67 percent, per Medical Economics coverage. The Nalu acquisition is sized against more than $60 million in 2025 sales growing more than 25 percent in 2026, per the Boston Scientific announcement. The two PNS acquisitions sit on the same shelf in opposite directions. Medtronic owns the entry-level temporary therapy that sets up a permanent implant. Boston Scientific owns the permanent micro-IPG implant that competes against Medtronic’s own permanent neuromodulation portfolio. The category is no longer organized around point devices. It is organized around portfolio-shaped pain-management pathways the buyer can assemble across the patient’s 12- to 36-month treatment timeline. ## Nevro Picked AI as the Third Lane Nevro’s HFX AdaptivAI platform, FDA-cleared in 2024 and now scaling across the HFX iQ spinal cord stimulation system, runs closed-loop pain therapy off real-time patient input and a training dataset Nevro describes as more than 100 million data points across 100,000-plus patients, per the Nevro release and Medical Design & Development. The AI lane reads as the mid-sized challenger’s play to differentiate against the two big incumbents at the algorithmic layer rather than the form-factor layer. The structural question for the next entrant is which of the three lanes the venture and the regulatory clock can defend. The portfolio-acquisition lane is closed unless a founder is willing to sell the company in the second or third year of revenue. The closed-loop AI lane requires the patient-data flywheel Nevro spent years assembling. The remaining lane for a new entrant is a clinical or anatomical pocket where a single-segment device clears against the existing portfolios rather than competing with them. ## The Q1 2026 Earnings Backdrop Is Still the Frame The Q1 2026 medtech earnings cycle is the environment all three deals were sized against. Boston Scientific and Abbott both cut full-year 2026 guidance despite beating Q1, and analysts characterized the quarter as one investors wanted to leave in the rearview mirror, per prior MedTech Dive and MassDevice coverage. In that environment, the buy-side is repricing structural defensibility into the multiple and treating a single-product PNS thesis as a transactional asset rather than a platform. Both Medtronic and Boston Scientific are paying mid-nine-figure prices to consolidate a category that the buy-side is now pricing on the depth of the pain-management shelf rather than on the unit economics of a single implant. For a hard-tech founder running a much smaller capital base inside chronic pain or any related neuromodulation segment, the read is that the comparable inside the category is no longer a single 510(k) clearance or a single De Novo. The comparable is a shelf the strategic acquirer can plug the device into. ## The Read Across to Other Hard-Tech Verticals The same shelf-shaped category logic is showing up across most of the hard-tech verticals BRG covers. Interventional cardiology is consolidating around full-stack vessel-prep, valve, and imaging combinations rather than single-device launches. Surgical robotics is consolidating around installed-base ecosystems rather than standalone platforms. Diagnostics is consolidating around CDx label inheritance and pipeline-level pharma partnerships rather than single-assay clearances, as the Guardant sequence the prior week showed. The chronic pain category in May 2026 is the same pattern landing in a different specialty. For founders in defense hardware, climate hardware, industrial robotics, and physical AI watching the medtech category move, the structural takeaway is that the consolidation logic is the same one the buy-side has been pricing across hardware platforms for the last two years. The acquirer is not paying for a device. The acquirer is paying for a slot in the pathway the customer walks through. Dave’s take If I were sitting with the founder of a chronic pain neuromodulation startup right now, the first question I would want to settle is which slot in the pathway the company is competing for. The temporary-PNS slot is now owned by Medtronic. The permanent micro-IPG slot is owned by Boston Scientific. The closed-loop AI overlay is owned by Nevro. The only remaining venture-backable position is a specific anatomical or clinical pocket the three incumbents are not credibly servicing, and the operating plan has to be designed to clear into that pocket with a label the incumbents cannot replicate from inside their existing portfolios. The same exercise applies to any hard-tech founder in a category the strategics have started consolidating around. The work is to identify the slot, document why the incumbents cannot fill it, and architect the engineering, regulatory, and commercial plan against the slot from day one. From Dave’s video library Dave on identifying the assumption that has to be true before anything else gets built. For a chronic pain startup today that assumption is which slot in the treatment pathway is actually still open, because three of them just got bought. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Software Cadence Is Becoming the Surgical Robotics Moat | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-05-24-surgical-robotics-software-cadence-platform-moat [← All Field Notes](/field-notes) May 24, 2026 Surgical Robotics Intuitive Surgical disclosed on May 22 that it has more than 100 updates and user-experience improvements queued for the da Vinci 5 platform, with measured launches in the US starting in June and an international rollout to follow, per MedTech Dive coverage of the company release. Medtronic announced two days earlier that it is opening a European software development hub in Galway focused on cardiac digital health, adding 85 jobs initially and slotting in as the company’s first European software team alongside existing US and India teams that already employ around 1,000 people, per The Irish Times and MassDevice. Lithuania-based Sentante won a CE mark for its haptic endovascular robotic platform in the same week, with the system positioned around remote-network architecture for thrombectomy and other vascular procedures, per Medical Device Network and the company release. Three stories, three different lanes inside the surgical-robotics platform race, and one underlying structural read for any hard-tech founder building inside the category. The hardware contest the new entrants spent the last five years training to win is shifting up a layer. The next twelve months of share will be decided by who can ship updates against an installed base, not who shipped the cleanest robot. ## Intuitive’s 100-Plus Updates Reframe What Owns the Installed Base Intuitive’s May 22 disclosure covers a deliberate set of capability additions to a platform that already has more than 1,400 da Vinci 5 systems installed and more than 12,800 surgeons running over 380,000 procedures globally, per MedTech Dive. The visible list includes a remote operating-room camera, an enhanced telepresence cursor and audio loop, mobile phone login with multifactor authentication, six new SimNow 2 simulation exercises, increased reusable counts on five of six Force Feedback instruments from six uses to fifteen, and a digital intraoperative ruler and surgeon-initiated tool eject queued for later this year subject to additional clearance. Iman Jeddi, the company’s general manager of da Vinci platforms, told MedTech Dive that the da Vinci 5 carries roughly “10,000 times the computing power” of the older Xi system, framing the update list as evidence that the new hardware was sized for years of software extension. For a founder running a much smaller surgical-robotics platform against Intuitive, the relevant variable is not the headline feature in the update list. The relevant variable is the cadence. Intuitive has now demonstrated it can ship a measurable software refresh inside a single quarter window, and the second-order effect is that the time a competitive launch has to look modern before it starts to look outdated has shortened. The installed-base moat is no longer the count of systems. It is the cadence of meaningful updates the installed base receives. ## Medtronic Plants a European Software Hub Behind Hugo Medtronic’s Galway announcement, reported by The Irish Times and MassDevice on May 20, sets up a Patient Care Systems hub focused on cloud-based platforms and clinical software for implanted cardiac devices, including the software used to program pacemakers and implantable defibrillators at the point of insertion. The headline number is 85 new jobs initially, with leadership, software engineering, and systems reliability roles, and the company has signaled more growth over the next three years. The structural read is more interesting than the headline. This is Medtronic’s first European software development hub. The existing US and India teams already total roughly 1,000 people. Galway gives Medtronic a third time zone that can hand work to India in the morning and to North America at the end of the European day, which is a deliberate sequencing choice for a software organization that needs to ship continuous updates against an installed base of regulated devices. The Hugo robotic-assisted surgery system cleared the FDA in late 2025 and ran its first US cases in February 2026. Medtronic is now putting the infrastructure behind it that the next phase of the competition is going to require. If Intuitive has shown that the update cadence is the moat, Medtronic is building the time-zone-spanning engineering organization needed to keep up with it. ## Sentante’s CE Mark Picks the Specialist Network Lane Sentante received its CE mark in May 2026 for an endovascular robotic platform that connects a haptic remote workstation to a bedside robot through a secure network, with standard cath-lab guidewires and catheters in the device path, per Medical Device Network, Yahoo Finance, and the Vascular News write-up. The CE submission was supported by the ESSENTIAL study, which completed first-in-human cases in July 2024 at Paula Stradiņš Clinical University Hospital in Riga and covered remote peripheral interventions including balloon angioplasty and stent delivery. Sentante also holds FDA Breakthrough Designation from September 2025 and was accepted into the FDA TAP program in February 2026, three months before the CE mark landed, per the Medical Device Network coverage. The structural framing in Sentante’s pipeline is the more important detail. The company is positioning the platform around remote thrombectomy for acute ischemic stroke at hospitals that do not have local thrombectomy capability or a reliable transfer pathway to a center that does. That positioning is not a robot-against-robot story. It is a network-architecture story where the value the device captures is the ability to deliver a specialist across a network rather than across a building. Sentante is competing inside the same surgical-robotics category Intuitive and Medtronic dominate at the platform level, but it is competing for a slot the platform robots cannot reach without a comparable remote-network stack. ## The Shelf Logic Is Now Above the Hardware Layer The three stories read together describe a category whose competitive surface is moving up a layer. The hardware-only contest a new surgical-robotics platform won by clearing the FDA, qualifying the supply chain, and showing reliability data is no longer a winning posture by itself. The incumbents are now ranking competitors on how fast the platform can ship meaningful updates against an installed base, how deep the software organization is that produces those updates, and how credible the remote-network architecture is for the use cases the bedside robot alone cannot reach. The same logic showed up the prior week in the chronic-pain neuromodulation deals where Medtronic, Boston Scientific, and Nevro divided three slots that each required a different architectural choice. The surgical-robotics version of that pattern is the same shape with different specific slots: installed-base update cadence, time-zone-spanning software engineering, and remote-network specialist coverage. For a founder building inside the category in 2026, the launch documents the company is drafting need to defend the position inside all three layers, not only the hardware. ## The Read Across to Other Hard-Tech Verticals The same shelf logic is landing in adjacent hard-tech verticals on a comparable timeline. Diagnostics is consolidating around pipeline-level pharma partnerships and label-inheritance strategies, as the Guardant sequence two days ago showed. Defense hardware is consolidating around platform-attached autonomy and the long-term sustainment contract behind it, as the Rocket Lab and Anduril precedents earlier this month indicated. Industrial robotics and physical AI are consolidating around installed-base data flywheels, which the Fanuc and Google Cloud partnership demonstrated last week. Climate hardware is consolidating around grid-integration software and service contracts on top of the asset rather than the asset itself. For founders building in any of these adjacent verticals, the read across from this week’s surgical-robotics news is the same one the category itself just delivered. Hardware quality is now table stakes. The defensible slot is one layer above that, in the software, data, and network architecture that determines what the hardware can do across an installed base of customers the strategics already serve. Dave’s take If I were sitting with the founder of a surgical-robotics company today, the first conversation I would want to have is about the software roadmap, not the robot. I want to know what the company can ship into the installed base in the first twelve months after clearance, who is on the engineering team that ships it, and what the network architecture looks like for the procedures the bedside robot alone cannot reach. The same conversation applies whether the company is in surgical robotics, defense hardware, industrial robotics, or climate hardware, because every one of those categories is consolidating around the same layer. The robot, the platform, or the asset is the table-stakes deliverable. The defensible business is built one layer above that. From Dave’s video library Dave on why customers stay for better working, not better looking. A surgical robot's software cadence after clearance is the medtech version: what you ship into the installed base is what keeps the installed base. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Pentagon Deal Team Six and the New Path to Market for Defense Hardware Founders | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-05-25-pentagon-deal-team-six-defense-hardware-founder-path [← All Field Notes](/field-notes) May 25, 2026 Defense Hardware The Pentagon’s Deal Team Six negotiating unit, housed inside the Economic Defense Unit and led by former Cerberus defense head George Kollitides, stood up in early April and is now visibly reshaping how contractor offers get scored, per Military Times and Axios reporting from earlier this month. The Pentagon awarded California-based Perennial Autonomy a three-year, $500 million ceiling IDIQ contract for counter-drone systems on May 19, with Merops interceptors, Bumblebee quadcopters, and Hornet midrange strike drones already in operational use across US Central Command, per DefenseScoop and Military Times. The FY27 budget request lifts the Defense Autonomous Warfare Group, which absorbed Replicator at the end of 2025, from $225.9 million to $54.6 billion, with about $53 billion of that sitting in a flexible five-year reconciliation pot rather than the standard base budget, per Defense One. Anduril closed a $5 billion round at a $61 billion valuation on May 13 led by Thrive Capital and Andreessen Horowitz, and CEO Brian Schimpf disclosed $2.2 billion of 2025 revenue in the announcement post, per TechCrunch and CNBC. Four announcements, one structural read for any founder building defense hardware in 2026. The path to market that worked through 2024 is being rewritten in the same calendar quarter on three dimensions at once. The contract shape is different. The capital plan is different. The defensible slot the buy-side is pricing is different. ## Deal Team Six Moves Factory Capex onto the Contractor’s Balance Sheet Defense Secretary Pete Hegseth introduced Deal Team Six in a November 2025 memorandum, and the team formally stood up in early April under George Kollitides, per Military Times. The mandate is to renegotiate contractor terms so that contractors fund their own factory expansions, new assembly lines, and plant construction, with the Pentagon offering steady long-term orders at flat prices in return. Hegseth’s framing, quoted by Military Times, is that the department has spent decades letting contractors charge taxpayers for both the factory and the final product. The Economic Defense Unit was funded at roughly $266 million in FY26 for research, development, test, and evaluation, with the FY27 request lifting that line to more than $593 million, per the same coverage. The structural read for a hardware founder is that the program-of-record assumption many capital plans were written against has changed shape. The contract is still real, but the assumption that the customer funds the factory is gone. For a founder running a defense hardware company in 2026, the capital plan that worked in 2023 needs to be rewritten with the contractor-funded expansion sitting on the company’s side of the line and the long-term order sitting on the Pentagon’s side. The companies that read the shift early get more of the long-term order. The companies that read it late get less of it. ## Perennial Autonomy’s $500M IDIQ Shows the New Contract Shape Joint Interagency Task Force 401 awarded Perennial Autonomy a three-year IDIQ contract on May 19 with a $500 million ceiling for AI-enabled counter-unmanned aerial systems, per DefenseScoop. The product line includes the Merops interceptor, which has been credited with downing over 4,000 Russian drones in Ukraine since mid-2024, alongside the Bumblebee quadcopter and the Hornet midrange strike drone, all of which are described as currently employed by forces operating in US Central Command. Brigadier General Matt Ross described the partnership as providing state-of-the-art counter-UAS capability to remain lethal on the modern battlefield, per the DefenseScoop coverage. The contract structure is worth reading carefully. An indefinite-delivery, indefinite-quantity ceiling at $500 million is not a $500 million purchase order. The Pentagon is reserving the right to order against the ceiling at the pace operational demand justifies, with deployment already in progress against systems Perennial is already producing. For a founder building defense hardware in an adjacent category, the precedent worth tracking is that the buy-side is signaling appetite for capability that is already proven in operational use, contracted under a flexible ceiling rather than a fixed quantity, and tied to a counterpart task force with a clear procurement mandate. That is a very different shape from the traditional fixed-quantity program-of-record and it rewards a very different go-to-market sequencing. ## DAWG Relocates the FY27 Demand Curve Above the Airframe The Defense Autonomous Warfare Group absorbed the Replicator initiative at the end of 2025, and the FY27 budget request lifts its funding from a $225.9 million FY26 line to $54.6 billion, per Defense One. Only about $1 billion of that sits in the standard base budget. The remaining $53 billion sits in a flexible reconciliation pot that DAWG has up to five years to obligate as the autonomous-systems portfolio matures, with spending allocated across procurement, operations, maintenance, training, and sustainment. The detail worth pulling out is the shift in what DAWG is buying. Defense One frames the institutional read as a pivot away from procuring finished drone airframes and toward orchestration tools that can be deployed across multiple airframes, with Shield AI’s integration into the LUCAS drone program and onto the Fury fighter platform as the working examples. For a founder building defense hardware in 2026, the implication is that the demand curve the buy-side is most willing to pay against is moving one layer up. The airframe still has to clear, but the persistent procurement appetite is for the autonomy stack that runs across airframes, the data plumbing behind it, and the sustainment contract that pays for keeping it current. A capital plan that funds the airframe to clearance and then waits for a follow-on software contract has the sequencing inverted from where the FY27 demand is pointing. ## Anduril at $61 Billion Sets the Comparable for the Software-Layer Moat Anduril closed a $5 billion round at a $61 billion valuation on May 13, led by Thrive Capital and Andreessen Horowitz, with CEO Brian Schimpf disclosing $2.2 billion of 2025 revenue in the announcement, per TechCrunch and CNBC. The valuation roughly doubles the company’s prior $30 billion mark and lands two months after the US Army awarded Anduril a ten-year enterprise contract with a $20 billion ceiling for battle-manager software running on the company’s Lattice platform. Earlier this month, Anduril was named as part of the consortium developing the space-based Golden Dome missile-defense architecture and also announced contract wins with the Dutch Ministry of Defence and a US Army battle-manager program. The comparable matters for any defense hardware founder running a fundraise or planning one. The valuation the buy-side priced is anchored to the software-and-orchestration layer, not to the airframe count. Schimpf’s revenue disclosure also doubled year over year, which sets the implicit growth bar the comparable is going to be applied against when a smaller round gets priced this summer. The companies that read the comparable cleanly are the ones that can show how the software layer they own compounds against the airframe, sensor, or platform they ship. The companies that try to be priced as airframe vendors against the Anduril comparable are going to find the math very unfriendly. ## The Read-Across to Adjacent Hard-Tech Verticals The same procurement-shape shift is landing in adjacent hard-tech categories on a comparable timeline. In MedTech, the CMS-FDA RAPID Coverage Pathway has compressed Medicare coverage to 60 to 90 days after FDA marketing authorization for breakthrough devices and is starting to pull commercial timing forward in the same way the IDIQ ceiling structure pulls defense procurement forward. In diagnostics, the Guardant sequence from earlier this week showed the moat moving to pharma-partnership depth and pipeline-level CDx collaboration rather than panel size. In industrial robotics and physical AI, the Fanuc and Google Cloud partnership showed the same demand-curve shift toward the software and data layer above the installed asset. For a founder building any of these categories, the takeaway is not that defense procurement has changed. The takeaway is that the institutional buyers across regulated and capital-intensive hardware are converging on the same posture. They will pay above the asset for the durable cadence and orchestration layer the asset runs inside, and they will pay less than the founder expects for the asset itself. A founder whose capital plan, contract structure, and engineering organization are still organized around the asset is misreading the buyer they are actually selling to. Dave’s take If I were sitting with a defense hardware founder this week, the conversation I would want to have is about the capital plan against the Deal Team Six environment, not about the next prototype milestone. The factory expansion sits on your balance sheet now, and the long-term flat-priced order sits on theirs, so the financing question is which credible institutional capital is willing to fund a contractor-financed capacity build with a long-tail counterparty. The same conversation applies across MedTech, diagnostics, climate hardware, and industrial robotics, because every one of those buyers is rewriting their procurement shape in the same direction. The asset is necessary. The defensible business is built one layer above it and financed against a contract structure most founders are not yet reading correctly. From Dave’s video library Dave on the small set of numbers that actually run a business, including the runway math. Deal Team Six just changed the financing question for defense founders: the factory sits on your balance sheet, and the numbers have to carry it. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Physical AI Foundation Models Pull $400M to $700M Checks Above the Robot | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-05-26-physical-ai-foundation-models-platform-bet [← All Field Notes](/field-notes) May 26, 2026 Industrial Robotics Brett Adcock’s AI hardware company Hark closed a $700 million Series A on May 21 at a $6 billion post-money mark, led by Parkway Venture Capital with Nvidia, AMD Ventures, Intel Capital, Qualcomm Ventures, ARK Invest, Greycroft, and Salesforce Ventures all participating, per TechCrunch. RJ Scaringe’s Mind Robotics, the Palo Alto general-purpose industrial robotics spinout from Rivian that incorporated in November 2025, closed a $400 million Series B led by Kleiner Perkins in mid-May, taking total raised across seed, Series A, and Series B to roughly $1 billion in six months with Rivian as both customer and shareholder, per The Robot Report. Stuttgart-based Sereact raised a $110 million Series B on April 27 led by Headline, a four-times step-up from its January 2025 Series A, with the round funding deployment of its Cortex 2 vision-language-action model across logistics, manufacturing, and humanoid platforms, per Bloomberg and Sifted. Physical Intelligence is now sitting at $1.1 billion of total capital raised after a $600 million Series B at $5.6 billion in late November led by CapitalG and Lux, with PYMNTS reporting in the past week that the company is in market for a follow-on round targeting $1 billion in fresh capital. Four data points in five weeks, one structural read. The capital that is showing up in industrial robotics and physical AI is pricing the foundation-model and orchestration layer above the robot, not the robot itself, and the implication for any hard-tech founder building hardware is that the durable slot the buy-side is paying for is one layer up from where the engineering plan probably points. ## Hark’s $700M Series A Shows the Strategics Are Pre-Loading the Stack Brett Adcock, the founder behind Figure AI and Archer Aviation, raised $700 million for Hark in a Series A that valued the company at $6 billion post-money roughly six months after the company was founded, per TechCrunch. Parkway Venture Capital led, with Nvidia, AMD Ventures, Intel Capital, Qualcomm Ventures, ARK Invest, Greycroft, Salesforce Ventures, Brookfield, Prime Movers Lab, and Tamarack Global all participating. Hark currently runs a data center on Nvidia B200 GPUs and plans to release multimodal models this summer that will power a personal AI platform with hardware devices to follow, per the same coverage. The detail worth pulling out is the investor list. Three of the four major US AI-chip strategics, plus a major mobile-compute strategic, are all on the same cap table inside the same round. That is the strategic supply chain pre-loading the company that is going to drive demand for their silicon. For a hard-tech founder thinking about strategic capital, the read is that the chip strategics are no longer waiting until you have a product to write a check. They are funding the company that is going to build the demand for what their roadmap was already going to ship. The window to bring a chip strategic in early at a defensible mark is open right now, and the founders who are reading the signal cleanly are getting commitments at rounds the next cohort will not be able to replicate. ## Mind Robotics Raised $1B in Six Months by Pricing the Platform, Not the Hardware Mind Robotics was founded in November 2025 by RJ Scaringe, the Rivian co-founder, and has now raised a $115 million seed in late 2025, a $500 million Series A in March 2026, and a $400 million Series B in May 2026 led by Kleiner Perkins, per The Robot Report. Total raised sits at roughly $1 billion in six months. Rivian is both the company’s first customer and a shareholder, providing what the company describes as a live, high-volume manufacturing environment for model training and deployment. Scaringe’s public framing, quoted in the Robot Report coverage, is that the company is not building single-task machines and is instead developing a platform that generalizes across core manufacturing tasks and scales across domains. The structural read is that the round is being priced as a foundation-model and deployment-infrastructure company that happens to use industrial robots as the form factor, not as a robotics company that has a customer. The Rivian factory floor is the training dataset that compounds across every future deployment. The economics of the second customer are very different from the economics of the first, because the model that lands on the second customer’s floor was already trained on Rivian. For an industrial robotics or physical AI founder weighing whether to verticalize against one customer or platform-ize against many, Mind Robotics shows the buy-side is willing to fund either path. The trick is being explicit about which one the company is building and writing the engineering plan, the contract structure, and the capital plan accordingly. ## Sereact’s 4x Step-Up Validates the Robot-Brain-as-a-Service Slot Sereact closed a $110 million Series B on April 27 led by Headline, with Bullhound Capital, Felix Capital, and Daphni joining as new investors and Creandum following from the January 2025 Series A, per Bloomberg. The round is a roughly four-times step-up from the Series A in fifteen months. Use of funds is to scale Cortex 2, the company’s vision-language-action model, across logistics and manufacturing customers and into humanoid robot platforms, with a US office opening in Boston as the commercial wedge into the American market, per Sifted. Sereact reports more than 200 of its systems deployed across Europe and more than 1 billion production picks executed, with roughly one in every 53,000 requests requiring remote human intervention. The interesting precedent is the deployment model. Sereact does not build a robot. Sereact builds the intelligence layer that any hardware OEM’s robot can use to handle the variability that previously required a fixed-purpose end-of-arm tool plus a human operator. The implication for a hardware founder is that there is now a credible third-party robot-brain layer that can ship inside the form factor the company already builds. Whether that is good news or bad news depends on which side of the OEM-or-platform line the founder is on. The founder who is building a hardware platform that needs to scale across customers can ship Sereact on day one and recover the engineering years. The founder who thought the intelligence layer was the moat now has to make a structural decision about whether they are competing with Sereact or partnering with them. ## Physical Intelligence’s $1B Follow-On Tells the Cross-Cutting Story Physical Intelligence raised a $600 million Series B at a $5.6 billion valuation in late November led by CapitalG and Lux, with Bond, Redpoint, and Sequoia participating and Bezos, OpenAI, T. Rowe Price, and Thrive Capital following from prior rounds, per The Robot Report. Total raised to date is roughly $1.1 billion. PYMNTS reported this past week that the company is now in market for a follow-on round targeting approximately $1 billion in fresh capital, which would push total raised toward $2 billion before a single end customer has been disclosed. The company’s product is robot foundation models, including the open-sourced π0 vision-language-action model series, with version 0.6 released earlier this year. Real-world demonstrations the company has shown include making espresso, folding laundry, assembling cardboard boxes, and inserting filters into machines. CapitalG’s framing, quoted at the time of the Series B, is that the approach is transformative because it builds a single generalist intelligence that manifests in any physical form to solve any real-world problem. The read for a hardware founder is that the buy-side is willing to pay foundation-model multiples for the intelligence layer well in advance of revenue, which is a structurally different price than the hardware layer can support. The founders who are getting the cleanest rounds in this category are the ones who can credibly argue that what they ship is the layer that compounds, not the layer that ships in a box. ## The Read-Across to MedTech and Other Regulated Hardware The same posture is now visible in adjacent regulated and capital-intensive hardware on a comparable timeline. In MedTech and surgical robotics, the software-cadence shift documented earlier this week, with Intuitive Surgical’s 100-plus queued da Vinci 5 updates and Medtronic’s new Galway cardiac digital health hub, is the same demand-curve direction as DAWG’s shift toward orchestration above the airframe and Mind Robotics’ pricing as a platform rather than a machine. In diagnostics, the Guardant Liquid CDx and pharma-partnership pattern from May 22 is the same structural shape, where the moat compounds in the pipeline-level collaboration layer rather than in any one panel. In defense hardware, the Anduril $61 billion comparable from May 13 anchored the same logic at the software-and-orchestration layer above the airframe. For a founder building hardware in any of these categories, the takeaway is not that the hardware is unimportant. The hardware still has to ship, has to clear regulatory, and has to perform. The takeaway is that the institutional capital and the strategic acquirers are converging on a posture where the multiple they will pay sits at the layer above the asset. A founder whose engineering organization, contract structure, and capital plan are still organized around the asset itself, with the platform layer treated as a phase-two roadmap promise, is going to find the next round priced against an unfriendly comparable. Dave’s take The way I would frame this for a founder I sat down with this week is that the platform-versus-hardware decision is no longer an academic question about long-term strategy, it is the question that determines how the next round gets priced. The strategic chip cap table on Hark, the Rivian customer-and-shareholder structure on Mind Robotics, and the third-party robot-brain pattern on Sereact each show a different shape of the same trade. My job working with a Fractional CPO client building physical AI or industrial hardware is to make sure the engineering organization, the customer architecture, and the capital plan all read the same answer to that question, so the next round lands on a comparable the company can actually live inside. From Dave’s video library Dave on the difference between real value and fabricated value, whatever the valuation says. The platform-versus-hardware question in physical AI is now a pricing question for your next round, and the honest answer is worth more than the bigger number. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # MDUFA VI Just Reset the De Novo and Pre-Sub Mechanics for Device Founders | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-05-27-mdufa-vi-de-novo-presub-credits-founder-mechanics [← All Field Notes](/field-notes) May 27, 2026 Regulatory The FDA and the medical device industry reached an agreement in principle on MDUFA VI at MedCon on March 18, the user-fee framework that will govern device submissions from fiscal 2027 through 2032, with the formal commitment letter now being drafted for Congressional submission by January 15, 2027, per MedTech Dive and RAPS coverage. The operational details that surfaced through the MedCon panel and the subsequent regulatory press matter more to a founder running a De Novo or PMA submission than the fee numbers do. A new PreSTAR template and 30-day walkthrough are coming to the De Novo program, alongside a “not grantable” decision option that gives an applicant a second pass before a decline letter, mirroring the “not approvable” structure that already exists for PMA. Pre-submission fees are being raised, with the fee functioning as a credit toward a future marketing submission, and Breakthrough and STeP-designated devices are exempted. TAP 2.0 is launching as the next version of the Total Product Life Cycle Advisory Program. Seven priority expansion areas are codified in writing for the first time, including real-world evidence, consensus standards, global alignment, digital infrastructure, digital health, and patient science. Most firms will see lower fees in fiscal 2028 as the statutory structure shifts, with submission fees and domestic registration fees coming down and an “America-First” split applied to establishment fees between foreign and domestic manufacturers. None of this is the headline a founder reads on launch day. All of it is the structural ground the company is going to run on for the next five years. ## The De Novo “Not Grantable” Letter Is the Single Biggest Founder-Level Change The De Novo program has historically operated as a binary clearance gate. The agency either granted the request and let the company move into commercial launch, or it declined the request and the company was forced into a Type C meeting cycle, a resubmission, or a pivot to a different pathway entirely. MDUFA VI introduces a third option. A “not grantable” letter will allow the applicant another pass at the submission before the agency issues a formal decline, structurally identical to the “not approvable” letter that already exists for PMA applicants and that has functioned, in practice, as a survival mechanism for companies that would otherwise run out of capital before they could reposition. The PreSTAR template and the new 30-day walkthrough are designed to front-load the conversation, so the company arrives at the substantive review with the agency’s view on the package already understood. For a first-time hard-tech founder whose entire capital plan is sized to the De Novo timeline, the existence of the “not grantable” letter as a real option changes how the pre-submission strategy gets written. The pre-sub program is no longer the place where the company tries to prove it does not need a second pass. It is the place where the company designs the architecture of the second pass in advance, so that if it is needed, the additional six to nine months it buys do not come with a forced equity raise at a punitive mark. ## Pre-Sub Fees Become Credits, With Breakthrough and STeP Exempted The agency is increasing user fees on pre-submissions to promote what RAPS’s coverage described as “strategic and thoughtful utilization,” with the fees functioning as credits toward future marketing submissions. A focused 45-day follow-up timeline applies to certain pre-subs. Breakthrough Device and STeP-designated devices are exempt from the increased pre-sub fees. The structural effect is to penalize the historical pattern of running an inflated number of pre-sub interactions as a free informational fishing expedition, while preserving the program for companies that come in with focused, well-formed questions. For a founder, the operational read is that the pre-sub strategy now has to be designed against a cost-and-credit ledger, not just against a calendar. The Breakthrough and STeP exemption tells you which designation work is worth front-loading. A team that secures Breakthrough designation early not only buys the well-documented review-time and CMS-coverage benefits but also takes the pre-sub fee out of the equation entirely, which is a material cash-flow advantage during the lowest-revenue, highest-burn phase of the company. ## TAP 2.0 and the Seven Priority Expansion Areas The next version of the Total Product Life Cycle Advisory Program is being built into MDUFA VI as TAP 2.0, with FDA committing in writing to seven priority expansion areas the agency and industry have agreed will receive dedicated work over the five-year cycle. Real-world evidence is on the list, which signals that post-market data generation is being formalized as an input to ongoing regulatory decisions rather than treated as a separate research activity. Consensus standards and global alignment are on the list, which matters for any device founder whose commercial plan includes both US clearance and CE mark inside the same product cycle, because the cost of duplicate regulatory programs has been a structural drag on hard-tech companies trying to scale internationally. Digital infrastructure and digital health are on the list, which is the agency acknowledging that the device under review is increasingly inseparable from the software platform it ships inside. Patient science is on the list. The read for a founder is that the agency is publicly committing to working on the questions that have, for years, fallen between the cracks of formal submissions. A company that engages TAP 2.0 with a concrete real-world-evidence plan or a digital-health architecture question has a meaningfully better shot at productive agency dialogue than one that waits for a problem to crystallize inside a 510(k) or De Novo review. ## Lower Fees in Fiscal 2028 and the Foreign-Domestic Split Under the agreement in principle, total user fee funding stays largely flat between MDUFA V and MDUFA VI, but the statutory structure changes so that most firms will see lower submission and domestic registration fees in fiscal 2028, per MedTech Dive. The redistributive move is a split establishment fee, with overseas manufacturers paying a higher rate than domestic ones under what FDA is describing as an “America-First” approach. For a US-based hard-tech founder whose contract manufacturing is domestic, the FY2028 cost line gets easier. For a founder whose finished-device manufacturing is offshore, the same line gets harder, and the operating plan has to account for that. The number is not the issue. The issue is that the cost structure differential will start to show up in pitch deck comparables and in pricing-power assumptions, and any company whose next round will be marked off a comparable that is itself absorbing the new fee structure needs to have that math worked out before the term sheet conversation begins. ## The Read-Across to Diagnostics, Combination Products, and Adjacent Regulated Hardware MDUFA covers medical devices specifically, but the structural shifts inside it shape adjacent categories that share infrastructure with the device center. For a diagnostics founder running an IVD through 510(k) or De Novo, the “not grantable” option and the PreSTAR template are operational realities that need to be incorporated into the regulatory program plan. For a combination-product founder whose primary mode of action sits inside CDRH, the same is true, with the additional consideration that the cross-center review pathway has not been formally rewritten and will continue to require its own coordination plan. For a defense or climate hardware founder whose product has any medical or biological adjacency, like an autonomous medic system or an emergency-response wearable, the TAP 2.0 expansion areas around real-world evidence and digital infrastructure are the door the agency has just opened for early dialogue on a non-traditional product. The pattern is consistent with what RAPS framed as a “historic timeline” on the regulatory side, where the agency and industry are codifying in writing the structural changes that have been negotiated in the field for years. The founders who will read this cleanly are the ones who treat the commitment letter, when it lands publicly in the Federal Register between August and December, as a working document for the company’s regulatory program rather than as a policy update someone else will summarize for them later. Dave’s take When I look at this from a Fractional CPO seat, the “not grantable” letter is the single change I would have a first-time De Novo applicant building their pre-submission strategy around right now. The De Novo I shepherded through to clearance at Galen Robotics in 2023 would have been structurally different if a second-bite-at-the-apple letter had existed as a real outcome, because the design of the pre-sub program would have been written to optimize for that letter as the realistic worst case rather than for clean approval as the only outcome the capital plan could survive. The founders who land in that letter window in 2028 and 2029 with a coherent next-pass plan already drafted will use it as runway, not as a setback. From Dave’s video library Dave on the altitude mismatch that has founders doing the CPO job themselves, and what a fractional engagement actually looks like. The MDUFA VI mechanics are exactly the kind of product-level regulatory strategy that seat exists to own. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # The Neurosurgical Supply Shortage Reveals Supplier Qualification Risk | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-05-28-neurosurgical-supply-shortage-supplier-qualification [← All Field Notes](/field-notes) May 28, 2026 Medical Devices On May 6 the FDA added neurosurgical patties, sponges, and strips to the Medical Device Shortages List, citing a Medline Industries recall that began on March 13 after the company found higher-than-expected endotoxin levels on its branded neurosurgical pattie products, per the FDA letter to health care providers and MedTech Dive. The recall is classified as Class 2. The agency expects the disruption to extend through the end of 2026. Medline has not committed to a market re-entry date, and the HCP letter asks clinicians to conserve usage by reserving products for intracranial operations, opening packages only when necessary, and diversifying supply sources. On its surface the story reads as a hospital-operations problem. Underneath it is a structural lesson in supplier qualification, which is the part of the design history file that most hardware founders treat as procurement work and not as design work, and which lands on the wrong end of the Class 2 list when the design assumption underneath it turns out to be wrong. ## An Endotoxin Excursion Is a Failure of the Sterility Promise the Whole Category Is Built On Endotoxin is the marker. The actual failure is a process drift somewhere in the supplier’s manufacturing or sterilization chain, and the result is a recall of every SKU in the affected product family because the validation envelope no longer holds. For a hospital, the consequence is a year of triaged neurosurgical cases and substituted materials. For a founder building any single-use or implantable hardware product, the read is sharper. The sterility claim on the device is an inherited promise from the supplier’s validated process. The day that process drifts is the day the device’s own design history file becomes obsolete in the only place that matters to the FDA, which is the place where the company has to show that the product was manufactured exactly the way it was qualified to be manufactured. A Class 2 recall on a supplier’s product becomes a Class 2 problem inside any downstream company whose own clearance package leaned on that supplier’s data. The founders who finish the longest in MedTech are the ones who treated the supplier’s validation work as part of their own product, not as an upstream gift they received and could forget about. ## A Single-Supplier Concentration Took Out a Whole Class of Brain Surgery Procedures Medline holds dominant share inside the neurosurgical pattie and sponge category, which is how a single Class 2 recall translated into a year-long shortage at the procedural level rather than a temporary irritation at the procurement level. The FDA does not normally add a single-supplier recall to the Medical Device Shortages List. The fact that this one landed there says the agency assessed the substitution paths inside the rest of the supply base and concluded there was no near-term alternative source at sufficient scale, per the FDA HCP letter and Supply Chain Dive coverage. For a hardware founder in any regulated category, that is the part to read twice. Single-source qualification inside the DHF is permissible and often unavoidable on the path to clearance, because the validation cost of qualifying redundant suppliers is real and the timeline is rarely friendly. Single-source qualification on commercial day one is a different decision, with different consequences, and a different cost of being wrong. The structural question for the founder is not whether the regulatory program permits single-sourcing. It is whether the company’s commercial life can survive the same kind of process drift that just took the neurosurgical pattie category out for twelve months. ## The FDA Just Told the Whole Category to Diversify Supply Sources The agency’s letter to health care providers includes an explicit recommendation that the field as a whole diversify supply sources to reduce future exposure. The language is procedural and aimed at hospitals, but the precedent it sets is broader. The FDA is publicly putting supplier concentration on the list of structural risks it expects the device community to actively manage, which means the next 510(k) or De Novo reviewer who looks at a finished-device submission with a single qualified supplier on a critical component now has an external citation to point at when the reviewer asks why redundancy was not built into the post-clearance plan. For founders, the operating consequence is that supplier diversification has just been reframed from a cost-of-quality conversation into a quality-system input. The team that walks into the next pre-sub with a redundant-supplier roadmap is responding to the agency’s stated direction. The team that has not thought about it is starting that conversation behind the precedent the FDA itself created in May 2026. ## The Read-Across to Diagnostics, Defense, and Climate Hardware The substrate is different in every other regulated category, but the geometry is the same. A diagnostics founder running an IVD through 510(k) or De Novo carries the same sterility-and-validation chain on consumables and reagent kits, and the same Class 2 mechanism applies if the supplier’s endotoxin or bioburden envelope drifts. A defense hardware founder working on counter-drone or autonomy systems carries a structurally identical exposure on semiconductors, magnetics, and battery cells where a single supplier’s qualification anchors the whole bill of materials and a foreign-supplier shock translates directly into mission risk. A climate hardware founder building electrolyzers, grid hardware, or large-format batteries carries it on rare-earth-bearing components, ion-exchange membranes, and specialty polymers where the qualified supplier count is often one and the regulatory weight is in IRA tax-credit transferability rather than CDRH. The lesson the neurosurgical pattie shortage forces into the open is general. The bill of materials inside a regulated hardware company is a list of single points of failure that the founder either designed against or did not, and the cost of getting it wrong is paid in commercial months, not in procurement dollars. The relevant European parallel landed on the same day, when the EU MDR and IVDR registration deadline of May 28 went live for new devices and the legacy-device clock began ticking toward November, per BSI and Pure Global coverage. The whole regulated hardware world is being asked to make its supplier records legible and traceable at the same moment one of the most visible single-supplier failures in the category is playing out in front of the FDA. Dave’s take At Galen Robotics, the suppliers I qualified into the design history file before our De Novo were the suppliers we shipped with on commercial day one, because the validation work behind each one had a six-figure cost and a multi-month timeline that no Series A budget can absorb twice. When I see a category-wide shortage triggered by a single-supplier endotoxin excursion, what I read is that the design history files for the affected products treated supplier qualification as procurement homework rather than as design work. The founders who finish, in MedTech and in every adjacent hard-tech category, are the ones who treat the supplier list as a product spec from the day the prototype freezes. From Dave’s video library Dave on why the hard part of building did not disappear, it moved into execution. Supplier qualification is that hard part in device land: six-figure validation work nobody celebrates until a single-supplier recall empties the market. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Surgical Robotics in 2026: Knowing When Not to Launch | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-05-29-surgical-robotics-when-to-launch-cmr-versius [← All Field Notes](/field-notes) May 29, 2026 Surgical Robotics CMR Surgical spent the past year doing the one thing most venture-backed hardware companies cannot bring themselves to do. It held its US launch. CEO Massimiliano Colella, who took over in late 2024, stopped the planned rollout of the original Versius robot and waited roughly a year for the second-generation Versius Plus, telling MedTech Dive he did not want to launch the first generation into the US market. CMR has raised close to $1.5 billion, the most of any surgical device startup, and Versius is the second most-adopted soft-tissue robot outside the US, with more than 40,000 procedures across 30-plus countries since its 2019 CE mark. The decision to wait is the story. In a year when Medtronic, Johnson & Johnson, and a field of smaller challengers are all moving on Intuitive Surgical’s lead, the companies pulling ahead are the ones treating launch timing and launch position as design decisions, not as the reflex of shipping whatever is ready the day the round closes. ## Waiting a Year for the Right Generation Is a Sequencing Decision, Not a Delay Versius Plus received FDA clearance in December 2024 for gallbladder removal, with further submissions across gastrointestinal, gynecology, urology, and general surgery indications working through the agency during 2026, per MedTech Dive. CMR is running a soft US launch, partnering selectively with hospitals and surgery centers before any broad rollout. Colella has been blunt about why: the market does not forgive a single mistake. If you are building a surgical robotics or interventional device platform, the gravitational pull is always toward launching the generation you have, because the early revenue clock and the reference-site flywheel both start the moment the first system goes live. CMR’s read was that the first commercial impression in a market this unforgiving is worth more than the months of early revenue a generation-one launch would have bought. Which generation earns your first reference site is one of the highest-leverage calls a hardware CEO makes, and it is almost never the one the cap table wants you to make. ## The Challengers Are Competing Where the Incumbent Cannot Go Cheaply The smaller entrants are not trying to out-da-Vinci the da Vinci. Distalmotion and Moon Surgical are aiming squarely at ambulatory surgery centers, with Virtual Incision and CMR also pursuing outpatient-focused facilities, per MedTech Dive. Versius is portable and modular, designed to move between operating rooms, and Colella has framed it as complementary to whatever a hospital already runs rather than as a rip-and-replace. That is the right instinct against a 20-year incumbent, and it generalizes well past surgical robotics. A founder facing an entrenched leader, whether that is Intuitive in soft-tissue surgery, a defense prime in autonomy, or an established name in industrial automation, rarely wins with a feature-for-feature assault on the incumbent’s home turf. The opening is the segment the leader’s product architecture cannot serve economically. ASCs are exactly that kind of structural gap, because the capital cost and physical footprint of the dominant system were never designed to fit an outpatient setting. ## Medtronic and J&J Are Entering on Very Different Clocks The two best-capitalized entrants are pacing themselves years apart. Medtronic’s Hugo system cleared in urology in December 2024 and the company is now pursuing general surgery and gynecology, with expected early uptake among loyal Medtronic accounts and surgeons not already on Intuitive systems, per MedTech Dive. Johnson & Johnson submitted an FDA De Novo request for its Ottava robot on January 7, 2026, targeting general surgery procedures in the upper abdomen including gastric bypass and sleeve, backed by an investigational device exemption study in Roux-en-Y gastric bypass. Neither is launching broad. Both are sequencing one indication at a time. For a founder watching deep-pocketed competitors crowd into the category, the read is that capital does not collapse the timeline. Regulatory and clinical evidence still gate each indication individually, and a balance sheet the size of Medtronic’s or J&J’s buys reach, not the ability to skip the queue. ## Intuitive’s Answer Is Cadence, and the Read-Across Is General The incumbent is defending with release velocity. Intuitive has added video replay, tissue-force measurement, and new instruments to the da Vinci platform, and in January 2026 it secured FDA clearance for nine new cardiac indications including mitral and tricuspid valve repair, per MedTech Dive. The company forecasts 13 to 15 percent worldwide da Vinci procedure growth for 2026, down from 18 percent the prior year, and analysts put the point where competition meaningfully bites two to three years out. Cadence on a large installed base is the same moat playbook a climate-hardware or industrial-robotics founder will eventually face from an entrenched leader. You do not beat it by matching the release schedule. You beat it by owning a position the leader’s roadmap does not prioritize, which is precisely why the cardiac and ASC openings exist at all. Even a 20-year leader cannot be everywhere first, and the gap between where it is and where it is going next is where a challenger gets to build. Dave’s take The hardest call I ever made at Galen Robotics was not technical. It was deciding what we put in front of the first surgeons, and when, because in surgical robotics the first impression is the market. If you are building against an entrenched leader in any hard-tech category, your edge is rarely a better version of their product; it is a clear read of where they cannot go and the patience to launch there on your own clock. From Dave’s video library Dave on figuring out the right first thing to build before you spend months on the wrong one. CMR held a US launch for a year because what you put in front of the first surgeons, and when, is the whole market. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # The Climate-Hardware IPO Wave Is a Demand Story, Not a Tech Story | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-05-30-climate-hardware-ipo-wave-demand-not-technology [← All Field Notes](/field-notes) May 30, 2026 Climate Hardware The easy read on this spring is that the IPO window for climate tech swung back open. Fervo Energy, the enhanced-geothermal company, went public in mid-May at roughly a $12.4 billion market cap. X-energy, which builds small modular nuclear reactors, listed in April and its shares surged on the first day to a market value near $11.5 billion. Solv Energy, a solar and battery developer, went public in February at around $6 billion, per MIT Technology Review. Three multi-billion-dollar hard-tech listings in a single quarter, and the temptation is to file it under sentiment swinging back toward atoms. I think that read misses the actual mechanism. What carried these companies to the bell was not a friendlier market or a better technology. It was demand that had already been contracted before anyone rang it. ## The Companies Going Public Already Had Their Customers Look at who is standing behind each name. Google backs Fervo. Amazon owns roughly 20 percent of X-energy and is also one of its clients, per MIT Technology Review. These are not science projects priced off a roadmap and a TAM slide. They are companies whose largest strategic investors are also the buyers signing up to take the output, which is a very different posture to bring into a public offering. The hyperscalers are racing to secure electricity in an era of rising demand, much of it driven by data centers, and they are willing to anchor the supply side with both capital and offtake. For a founder building in energy, grid hardware, nuclear, or any category where the unit economics only close at industrial scale, that is the pattern worth studying. The listing is the visible event. The contracts that made it underwriteable were signed quarters earlier, out of view. ## A Hot End-Market Is the Tailwind, Not the Proof Rising electricity demand is real, and it helps. So does the political weather: geothermal and nuclear have kept federal support while renewables face headwinds under the current administration, per MIT Technology Review. But a hot end-market is not the same thing as demand for your specific machine, and conflating the two is how capital-intensive hardware companies talk themselves into raising against a story instead of a customer. The IPO-ready names did the harder thing. They converted a macro tailwind everyone can see into named buyers who had committed in writing. If you are a hardware founder watching this wave and reaching for the same narrative, the question is not whether the sector is hot. It is whether anyone has agreed to pay you for what comes off the line, at a price and volume that survive contact with your real cost structure. ## A Multi-Billion Valuation Is the Next Bar, Not a Finish Line There is a second lesson hiding inside the celebration, and it is the one I would press hardest on a founder reading these headlines. The number you print at an IPO, or at any round, becomes the number you then have to defend. Venture capital and public capital are both expensive money, and the post-money mark you accept today is the bar you have to clear at the next milestone. I watched Lucent go from $80 a share to about 52 cents, and the Ascend options I had converted into it, struck around $3, might as well have been worth zero. A multi-billion-dollar valuation is a claim about a future you still have to deliver, not a verdict on the present. Cheering the headline number while the deployment plan is mush is exactly how a company walks into a down round, and capital-intensive hardware gives you fewer cheap chances to recover than software ever did. Dave’s take For a capital-intensive hardware founder, the question this IPO wave should raise is not whether the window is open. It is who has already agreed to buy what you build, and whether you can defend the number you just printed when the next quarter or the next round arrives. Fervo and X-energy did not get public on a story. They got there on customers. Demand you can name beats a market you can merely describe, every single time. From Dave’s video library Dave on why polite interest proves nothing and a costly commitment proves everything. Fervo and X-energy went public on named customers and signed demand, which is the same proof at IPO scale. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Hard-Tech Mega-Rounds Land on the Physical Layer of the AI Stack →](/field-notes/2026-05-15-hard-tech-mega-rounds-physical-layer-capital) [Physical AI’s Big Checks Are Landing Above the Robot →](/field-notes/2026-05-26-physical-ai-foundation-models-platform-bet) [Rocket Lab and RTX Confirm the Subcontractor Playbook for Defense Hardware →](/field-notes/2026-05-16-rocket-lab-rtx-defense-hardware-subcontractor-playbook) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # A Narrow Indication Is the Regulatory Strategy for AI Medical Devices | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-05-31-narrow-indication-ai-medical-device-clearance [← All Field Notes](/field-notes) May 31, 2026 Medical Devices Picture the pitch a hardware founder wants to make to the FDA: our AI makes the procedure safer. It is the line every investor wants to hear, and it is almost always the wrong thing to file. On March 26, Philips earned FDA 510(k) clearance for EchoNavigator R5.0 with DeviceGuide, an AI tool that tracks a repair device in real time during minimally invasive mitral valve work by fusing live ultrasound and X-ray into one view. The part worth a founder’s attention is not the AI. It is how tightly Philips drew the box around it. The tool is cleared for one procedure, mitral transcatheter edge-to-edge repair, and for use with one device, the Edwards PASCAL Ace system it was co-developed alongside, per Philips. That narrow scope is not a constraint Philips settled for. It is the regulatory strategy. ## The Claim You File Is a Product Decision, Not a Regulatory One The instinct in most companies is to treat the FDA submission as something the regulatory team cleans up after the product is built. The order is backwards. The most consequential choice in the whole program is what you say the device is, in one sentence a reviewer will accept as moderate risk. I learned this at Galen the expensive way. We first framed our surgical platform as a remote-control telemanipulator, substantially equivalent to the da Vinci. The FDA rejected the predicate logic. The device did not change one screw. We reframed it as a stabilizer that improves a surgeon’s access to anatomy at the end of a long instrument, and that sentence carried us to a De Novo authorization. Same hardware, a different claim. The DeviceGuide clearance is that move run on an AI tool. “Track the PASCAL Ace device accurately during this one repair” is a workflow-and-accuracy claim a 510(k) can carry. “Our AI improves valve-repair outcomes” is a clinical claim that buys you a trial. ## Narrow Is Faster Because Narrow Is Provable A patient-outcome framing invites the agency to demand clinical proof, the slowest and most expensive evidence a startup can be handed. A workflow-or-access framing invites engineering and usability evidence, which you control and can generate on your own schedule. Pairing the tool to a single named device does something else founders underrate: it makes the performance claim testable. Does it track the PASCAL Ace accurately, yes or no. That is a question a reviewer can close out, which is exactly why it clears. This is where AI gets dangerous for founders, because the pressure is always to claim the model does more than you can show. Machine learning now sits inside a large and growing share of cleared software devices, and investors keep pushing portfolio companies for a bigger AI story. The regulator is asking for the opposite: a claim narrow enough to falsify. When those two pulls collide, the sober answer wins clearances. ## The Indication and the Investor Pitch Are the Same Sentence Here is the part that bites people later. The claim you make to the FDA and the claim you make to your board are not two stories you get to keep in separate rooms. Once you tell investors the AI delivers a clinical outcome, that language has a way of leaking into the design history file, the labeling, the website copy after launch. A single document that drifts from the cleared indication can sink the submission, or surface afterward as an off-label problem you did not budget for. The current marketing climate makes the discipline harder, because the grand, hard-to-check claim is the house style of the moment, and it tugs founders toward saying more than the evidence supports. Scoping the claim to exactly what you can prove is not timidity. In a regulated category it is the thing that gets you to market at all, and it is a product decision, which means it belongs to whoever owns the product. Dave’s take DeviceGuide reads like a story about AI arriving in the cath lab. The lesson underneath it is older than AI. What you say your device is decides what the FDA asks you to prove, and that sentence belongs to the person who owns the product, not the regulatory consultant you hand the file to at the end. Pick the narrowest true claim you can stand behind, then make every document, every slide, and every line of marketing agree with it. That is most of the game, and too many founders decide it by accident. From Dave’s video library Dave on what AI really is under the label, using the FDA's own list of cleared machine-learning devices to make the case. Philips just showed the practical move: scope the AI claim narrowly enough and the pathway opens. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [MDUFA VI Just Reset the De Novo and Pre-Sub Mechanics for Device Founders →](/field-notes/2026-05-27-mdufa-vi-de-novo-presub-credits-founder-mechanics) [AI-Embodied Surgical Robots Force a Regulatory Framework Reset →](/field-notes/2026-05-08-ai-embodied-surgical-robots-regulatory-framework-reset) [AI Imaging and Robotic Navigation Converge in the Interventional Stack →](/field-notes/2026-05-14-ai-imaging-robotic-navigation-interventional-stack) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Spectral AI’s De Novo Clearance: A Lesson in Non-Dilutive Capital | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-06-01-spectral-ai-de-novo-barda-non-dilutive-capital [← All Field Notes](/field-notes) June 1, 2026 Diagnostics When my team at Galen Robotics got the signal to pivot to an FDA De Novo, the room read it as a setback. I read it as the agency telling us we sat inside the bounds of acceptable risk, then handing us the pen to write our own standard. That memory came back the week Spectral AI announced, on May 26, that its DeepView system had earned De Novo authorization for assessing burn wounds. DeepView pairs multispectral imaging with an AI model that flags tissue unlikely to heal within 21 days, in about a fifth of a second per image. The headline writes itself: another AI device clears the FDA. The part a hardware founder should sit with is how it got funded, and what the clearance actually unlocks. ## The Money Arrived Before the Clearance, and It Came From Washington Spectral did not finance this device the way a software startup finances a feature. In March the company was awarded another $31.7 million from BARDA, the federal biomedical preparedness authority, on top of the $54.9 million already committed under a contract that can run up to $150 million in total. Spectral put in $9.7 million of its own. The reason the government is paying is not bedside convenience. DeepView is being built as a medical countermeasure for mass-casualty burn incidents, the kind a disaster or attack produces, alongside everyday use in emergency departments, trauma centers, and burn units. That is a real and underused financing path for safety-critical hardware. Non-dilutive capital carried a diagnostic through development and clearance without the founders selling the company off in pieces to do it. But non-dilutive is not the same as unconditional. The money comes attached to a mission, and the mission shapes the product: which indication you chase first, how you validate it, which settings of use you have to prove out. A founder who takes that money should be clear-eyed that the funder's priorities will steer the roadmap as surely as any board seat would. ## A De Novo Says the FDA Thinks You Are Moderate Risk, and Hands You the Standard DeepView went through De Novo because there was no predicate to point at. That is the whole tell, and most founders read it backwards. A De Novo route is not a punishment for being difficult. If the FDA considered the device high risk, it would be sitting in a PMA. Routing you to De Novo is the agency saying you fall inside the envelope of acceptable risk with new controls. And what you win at the end is leverage. When you finish a De Novo, the FDA writes a new regulation describing your device, because none exists yet. At Galen, they asked us what that language should say. We drafted a paragraph, sent it in, and the regulation that entered the books came back almost word for word what we had written. Spectral is now in that position for AI burn assessment. The next company into this category has to be substantially equivalent to them, and Spectral's own next feature becomes a clean comparison against itself. The 340-billion-pixel training set and the sub-second capture are the engineering proof. The lasting advantage is becoming the thing everyone else has to match. ## Authorization to Sell Is a Door, Not a Destination Here is the line in the announcement I would underline for any founder: the De Novo authorizes commercial distribution. That is the right to sell. It is not evidence that anyone will buy at the price and volume your model needs to work. For a diagnostic, the walls after clearance are the hard ones: reimbursement, fitting into a clinician's existing workflow, and standing up a commercial organization that can actually place and support systems in the field. Federal money can carry you to the clearance. It does not carry you across the adoption gap. Spectral, now a public company answerable to a quarterly clock, has to convert an authorization into installed devices and repeat use, and that conversion is where most cleared hardware quietly stalls. Dave’s take The reflex on a clearance day is to treat it as the finish line. I have stood in that room, and the honest version is different. A De Novo is the FDA inviting you to define a category and then go prove there is a business inside it. Non-dilutive money got Spectral to the door without selling the company a slice at a time, and that deserves real respect. What happens next is the part no contract pays for: getting clinicians to pick the device up, and getting paid for it, by name. From Dave’s video library Dave on why a milestone is not proof and the only proof is a buyer. A De Novo hands you a category, and non-dilutive capital gets you to the door, but the business still has to be proven the same way as everyone else's. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [A Narrow Indication Is the Regulatory Strategy for AI Medical Devices →](/field-notes/2026-05-31-narrow-indication-ai-medical-device-clearance) [MDUFA VI Just Reset the De Novo and Pre-Sub Mechanics →](/field-notes/2026-05-27-mdufa-vi-de-novo-presub-credits-founder-mechanics) [Guardant360 Liquid CDx and the Companion Diagnostic Platform Moat →](/field-notes/2026-05-22-guardant-liquid-cdx-companion-diagnostic-platform-moat) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # The Pentagon’s Drone Gauntlet Is a Manufacturing Test | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-06-02-pentagon-drone-gauntlet-manufacturing-readiness-test [← All Field Notes](/field-notes) June 2, 2026 Commercialization Somewhere a founder is perfecting the drone that will win a shoot-off at Fort Benning and still lose the contract. That is not a knock on the airframe. It is the design of the program. The Pentagon’s Drone Dominance Program, nicknamed the Gauntlet, runs in phases, and the phase that matters most for any hardware founder is not the one where the prototype flies. It is the one where you have to prove you can build the thing at volume. The Phase II qualifier window is estimated for this month, and what it tests is the part most founders defer until it is too late. ## The Gauntlet Separates the Demo From the Line Phase I looked like a demo day with weapons. The military invited 25 vendors to compete, started evaluations at Fort Benning on February 18, and put roughly $150 million behind prototype deliveries over the months that followed. That is the part everyone photographs. Phase II is where the program shows its real intent. According to the program’s own structure, it opens with a production and delivery test to prove manufacturing readiness, then runs Gauntlet II to find the systems worth scaling, then ends with actual delivery orders going to the winners. Winners have to show they can produce capable, low-cost units with a secure supply chain at scale. The whole effort runs past a billion dollars across four phases, aimed at fielding hundreds of thousands of one-way attack drones by 2027. Read the sequence and the message is plain. The government is not buying the best drone. It is buying the best drone someone can actually build, again and again, on a line. Those are two different competitions, and plenty of companies that win the first one lose the second. ## Why the Buyer Moved the Gate The reason is stated openly in how the program is framed. The United States is behind China on manufacturing and fielding drones, not on designing clever ones. When the bottleneck is production, the buyer stops paying for the prototype and starts paying for the factory. This matches a signal I have watched the defense world send for a while now. The procurement language keeps pulling toward narrow, direct, producible systems and away from the general-purpose moonshot, because a narrow form factor is far easier to build, secure, and field than a do-everything platform. General-purpose is a fundraising story. Narrow and producible is a delivery story. A program organized as a manufacturing gauntlet is the buyer saying out loud which story it intends to fund. The founders who file manufacturing under “solve after the raise” are optimizing for the phase that no longer decides anything. ## The Same Gate Shows Up in Every Regulated Market None of this is a defense quirk. The buyer who gates on the production system, not the prototype, appears in every hard-tech market that matters. In MedTech, the FDA does not let you sell on a working bench unit either. Before a cleared device ships, you complete design transfer and validate the manufacturing process under 21 CFR 820, proving the line turns out the same device every time and not just the golden sample you demoed for the review team. A strategic acquirer runs the identical play in diligence. So does a serious first enterprise customer. The gates carry different names, but they ask one question: can you make this, repeatably, at the cost and volume your plan assumes? For a founder building drones, diagnostics, grid hardware, or surgical robots, the takeaway does not change. The prototype earns the meeting. The production system earns the order. If you are still raising on a hero demo and treating the factory as a line item for some later quarter, you are deferring the exact thing your buyer has decided to judge you on. Dave’s take I have watched more than one team celebrate a flawless demo and then learn the hard way that the demo and the product were two different machines. The Pentagon has now turned that lesson into a published phase, with money attached to clearing it. If your roadmap puts manufacturing readiness after the next milestone, you have the order backwards, because the milestone that counts is the one asking whether you can build it at all. Prove the line, not the prototype. From Dave’s video library Dave reads the claim that AI removed the hard part and finds the hard part just moved. The Pentagon reached the same conclusion about drones: the gate is no longer the prototype, it is whether you can manufacture the thing at rate. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [The Neurosurgical Supply Shortage Reveals the Hidden Risk in Your DHF →](/field-notes/2026-05-28-neurosurgical-supply-shortage-supplier-qualification) [Deal Team Six Rewrites the Defense Hardware Founder’s Path to Market →](/field-notes/2026-05-25-pentagon-deal-team-six-defense-hardware-founder-path) [Physical AI’s Big Checks Are Landing Above the Robot →](/field-notes/2026-05-26-physical-ai-foundation-models-platform-bet) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # The Humanoid Robot Is Physical AI’s Hardest Bet to Ship | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-06-04-humanoid-robots-narrow-form-factor-edge-ai [← All Field Notes](/field-notes) June 4, 2026 Industrial Robotics Every robotics headline this week points at the same silhouette: a humanoid, roughly six feet tall, walking across a stage. Nvidia and Unitree just turned that silhouette into a product any university lab can order, and Nvidia’s Jensen Huang called it the opening of a multitrillion-dollar opportunity. I do not doubt the demo. I doubt the form factor. For most of the hard-tech founders I work with, the human-shaped robot is the most expensive way to lose, and now is a good moment to say why. ## What Nvidia and Unitree Actually Put on Sale The announcement is the Isaac GR00T reference humanoid robot, a design Nvidia and Unitree built together for labs to study. It pairs Unitree’s H2 Plus chassis, which stands nearly six feet tall, weighs about 150 pounds, and carries 31 degrees of freedom in the body, with a pair of Sharpa Wave tactile hands that bring the total to 75 degrees of freedom across the machine. All of it runs on Nvidia’s Jetson AGX Thor onboard compute, a Blackwell GPU rated at 2,070 FP4 teraflops with 128GB of memory, and it goes on sale through Unitree late this year, per Nvidia. Now look at who the launch customers are. Not a warehouse operator or an automaker. The names on the list are Ai2, ETH Zurich, the Stanford Robotics Center, and a robotics lab at UC San Diego. This is a research platform sold to research institutions, which is a very different object from a worker you drop onto a line next quarter. The number to sit with is 75. Seventy-five degrees of freedom is seventy-five things a controller has to sense, plan, and command in real time, in an unstructured world, often with no reliable network to lean on. That is close to the hardest standing problem in autonomy, and the platform’s own spec sheet is the evidence. The hands alone, the part that makes the robot look capable on stage, are where most of the difficulty hides. ## The Buyer With the Hardest Requirements Wants the Opposite Shape While the consumer story celebrates the generalist humanoid, the customer with the most demanding specifications in hardware has been asking for something narrower. The Defense Department keeps writing requests for untethered systems that function without an internet connection, and that single requirement reshapes everything downstream. If the robot has to work with no network, the intelligence has to run on the device itself, which rules out most of the hyperscale-dependent stack the splashy humanoid demos quietly rely on. Edge compute and a narrow form factor travel together. A drone or a four-legged platform has far fewer degrees of freedom and far more predictable contact with the world than a human-shaped machine, which is exactly what makes it something you can actually make autonomous on the hardware it carries into the field. General-purpose is a fundraising story. Narrow and producible is a delivery story. When the most exacting buyer in hardware keeps funding the second one, that signal outweighs a keynote. ## Surgical Robotics Already Proved the Narrow Path Works This is not a theory waiting for a test case. Computer vision and machine learning have been running on edge processors inside surgical-robotics systems for years, doing real perception and control inside a tightly bounded task. The lesson generalist robotics keeps skipping is that the narrow envelope is what made the autonomy tractable in the first place. So for a founder building robotics or autonomy, the question is not whether the humanoid is impressive. It is whether your form factor matches the procurement signal you are actually selling into. If your buyer is a defense or industrial customer, a machine that does one job without a network will beat a machine that could theoretically do everything with a data center behind it. Pick the form factor you can make autonomous and manufacture at volume, not the one that fills the last slide of the deck. That choice applies most directly to anyone building drones, quadrupeds, autonomous inspection, or edge-AI perception for regulated and safety-critical settings, where the buyer is going to ask what the device does the moment the network drops. Dave’s take The humanoid is the most photogenic object in technology right now, and that is precisely the problem. Photogenic and shippable are not the same machine. I would rather back a founder who picked a boring quadruped they can build a thousand of and run on edge compute than one chasing the seventy-five-degree-of-freedom dream that needs a research lab and a server farm just to stand up. The physical-AI decade gets won by whoever fielded a form factor they could actually deliver, not the one that filmed best. From Dave’s video library Dave on Barnum's methods and how to separate the show from the substance. The humanoid robot is the most photogenic object in technology right now, which is exactly why the test matters: photogenic and shippable are not the same machine. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Physical AI’s Big Checks Are Landing Above the Robot →](/field-notes/2026-05-26-physical-ai-foundation-models-platform-bet) [The Pentagon’s Drone Gauntlet Is Really a Manufacturing Test →](/field-notes/2026-06-02-pentagon-drone-gauntlet-manufacturing-readiness-test) [Deal Team Six Rewrites the Defense Hardware Founder’s Path to Market →](/field-notes/2026-05-25-pentagon-deal-team-six-defense-hardware-founder-path) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Closed-Loop Insulin Delivery: The Algorithm Is the Easy Part | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-06-05-closed-loop-insulin-delivery-validation [← All Field Notes](/field-notes) June 5, 2026 Medical Devices Somewhere inside Insulet right now, a product team has agreed to delete a button. The meal-announcement button, the one a person with diabetes presses to tell their pump a meal is coming, is on its way out. Insulet, MiniMed, and Tandem are all chasing the same finish line: an insulin system that doses correctly without ever being told to expect food. The trade press is calling it the arrival of the fully closed loop, and the algorithm gets all the attention. It is the easy part. Deleting that button is a regulatory and validation decision in a UX costume, and the company that understands that is the one that finishes. ## The Button They Are Deleting Here is the state of the race, per MedTech Dive. Insulet started a pivotal study in May for an automated insulin delivery system with no meal announcements, scoped to people with Type 2 diabetes who take insulin, with a 510(k) submission planned for 2027 and a launch targeted for 2028. MiniMed is studying a system called Vivera, where the meal announcement becomes optional rather than gone, in both Type 1 and Type 2 patients, and running a lower insulin target than its current 780G system. Tandem says it will start its own pivotal trial later this year. Three companies, three different answers to the same product question: how much of the human do you design out, and for whom? The trade-offs are the whole story. Take away the meal announcement and the algorithm now has to detect the meal and deliver the right dose on its own, without tipping a patient into hypoglycemia. Even the systems chasing a fully closed loop are not hands-off: users still adjust for exercise and illness, a sensor failure drops the device into a backup mode that asks for manual input, and performance shifts with the meal, the type of carbohydrate, and the size of the plate. None of that is a problem you solve in the algorithm once. It is a validation burden you carry through every one of those conditions, in front of the FDA, before you are allowed to ship. Watch where the population lines fall, because that is the real decision. When I was at Galen, we learned the expensive way that your study population is itself a regulatory claim, not a recruiting detail. We told the FDA we would run our cadaver study with surgical residents, got a written yes in the pre-submission, then watched the substantive review reopen it: were we planning to market to residents? We re-ran the entire study with sixteen attending surgeons in two weeks to keep the claim clean. Insulet scoping its first closed-loop system to Type 2 insulin users is that same lesson applied early and on purpose. A narrower population is a smaller, more defensible thing to prove. MiniMed studying both Type 1 and Type 2 is making a broader claim, and broader claims cost more to validate. ## Who Owns the Sensor Two days before that race got fresh coverage, MiniMed answered a different product question, and it is the one founders tend to underrate. On June 3 it expanded its partnership with Abbott to put a dual glucose-ketone sensor on its dosing systems, a sensor that watches for rising ketones to catch diabetic ketoacidosis before it becomes an emergency, per MiniMed. The sensor is built by Abbott, matches the size of the Instinct sensor the two companies already ship together, and is designed to plug exclusively into MiniMed's systems. Abbott holds a CE mark in Europe; the dual sensor is still under FDA review and not yet sold in the United States. In a closed-loop system, the sensor is not a component. It is the safety system. Everything the algorithm decides rests on whether the number coming off that sensor is true. MiniMed, a standalone diabetes company, has chosen to rent that input from Abbott, a company that sells sensing technology of its own. There are good reasons to do it: Abbott makes one of the smallest, most refined sensors in the world, and standing up a rival in-house would cost years MiniMed may not have. But it is a product-architecture decision dressed as a supply agreement, and it sets who controls the validation, the roadmap, and the failure modes of the one part the whole system has to trust. Adding a second analyte, ketones, widens what the system claims to do and widens what has to be validated next to it. The question worth pushing on is not the price of the sensor. It is what happens to the product the day the supplier's roadmap and yours stop pointing the same way. Dave’s take The headlines will keep framing this as an AI race, the smartest algorithm taking the most work off the patient. I have sat in the room where the demo wins applause and the validation plan is still a placeholder, and the demo is the part that matters least. Whoever wins the closed loop will be the company that scoped its intended use to something it could actually prove and kept control of the inputs that proof depends on, because automation is not the product. The validated claim is. From Dave’s video library Dave on companies burning enormous money on AI because the demo impressed the wrong people. The insulin-delivery race has the same shape: the algorithm wins the applause, and the validation plan decides who actually ships. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [A Narrow Indication Is the Regulatory Strategy for AI Medical Devices →](/field-notes/2026-05-31-narrow-indication-ai-medical-device-clearance) [The Neurosurgical Supply Shortage Reveals the Hidden Risk in Your DHF →](/field-notes/2026-05-28-neurosurgical-supply-shortage-supplier-qualification) [MDUFA VI Just Reset the De Novo and Pre-Sub Mechanics for Device Founders →](/field-notes/2026-05-27-mdufa-vi-de-novo-presub-credits-founder-mechanics) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # FDA 510(k) Exemptions Are Expanding. They Can Erase Your Moat | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-06-06-fda-510k-exemptions-predicate-moat [← All Field Notes](/field-notes) June 6, 2026 Regulatory A founder forwarded me the FDA notice this week with a one-line note: “Less paperwork, finally.” I read the same notice and went the other way. On June 5 the FDA published guidance saying it intends to exempt a set of unclassified medical devices from premarket notification, the 510(k) step, and that it will not enforce the requirement while the exemption is finalized. Most founders read a sentence like that as time and money handed back to them. Sometimes it is. But a 510(k) requirement is not only a cost you pay. For the company already on the market, it is a wall standing between them and the next entrant, and the FDA just started taking some of those walls down. ## What the FDA Actually Did The June 5 guidance is narrow on its face. It covers unclassified devices, the older products that predate the agency’s modern classification system and were never formally sorted into Class I, II, or III. The FDA now believes a short list of them are low enough risk to skip premarket review, so it is moving them toward exemption and holding off on enforcement in the meantime, which means it does not expect a 510(k) for those products today. On its own, a housekeeping notice. It is not on its own. Back in February the FDA published a separate request for comments naming a list of Class II devices it intends to exempt from premarket notification once finalized, per the Federal Register. Two notices in four months, one for legacy unclassified products and one for moderate-risk Class II devices, point the same way: the agency is pulling categories of hardware out of the 510(k) process rather than into it. For a founder mapping a pathway, the fact that matters is not which specific products moved this week. It is that the line of what needs a 510(k) at all is shifting, and it is shifting while you build. ## A Filing Requirement Is Also a Wall Here is the part the relief reaction misses. The 510(k) is a gate, and a gate works in both directions. When I was at Galen, our pathway turned into a de novo, and the lesson I took from it is one I now repeat to every device founder: clearing a hard regulatory bar costs you, but it also hands you something. A barrier you get to stand behind once you are through it. The de novo let us write the regulation and become the predicate every later entrant had to match. A 510(k) does a quieter version of the same thing. As long as a category requires premarket notification, a new competitor has to assemble a submission, find a predicate, and convince a reviewer before they can sell against you. That friction does free work for the incumbent every single day. Exempt the category and the friction is gone. The next entrant can build and ship without ever filing. If part of your product’s edge was that the regulatory path was hard, the exemption just lowered the wall for everyone behind you. The savings came off your filing budget and out of your barrier to entry at the same time. ## Who Should Cheer, and Who Should Re-Plan This is not a reason to mourn every exemption. If you are pre-market and the device you are building just landed on an exempt list, you got a real gift: faster, cheaper entry into a category you were about to spend a year and six figures clearing into. Take it. The founders who should re-plan are the ones whose defensibility was leaning on the filing itself. Single-product device companies in a newly exempt class. Anyone whose deck cites “regulatory barrier to entry” as a moat. If a reviewer no longer stands between your category and a fast follower, that line is now decoration, and your defensibility has to come from somewhere the FDA does not hand out: a validated claim a competitor cannot copy from the outside, a clinical dataset, a manufacturing process, switching costs that are real. The exemption will not build any of that for you. It just takes away the excuse to keep putting the question off. Dave’s take Deregulation gets sold to founders as pure tailwind, and that framing is lazy. Every requirement the FDA drops is a cost off your books and a door held open for the next person who wants your customers, and which of those two matters more comes down to whether you are the incumbent or the insurgent in that category. I would rather a founder know which one they are this quarter than find out the morning a competitor ships without filing anything at all. From Dave’s video library Dave on reading a big government move for its real consequences instead of the headline. FDA's expanding 510(k) exemptions deserve that read: the same deregulation is a cost saving for the incumbent and an open door for the insurgent. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Spectral AI’s De Novo Clearance: A Lesson in Non-Dilutive Capital →](/field-notes/2026-06-01-spectral-ai-de-novo-barda-non-dilutive-capital) [A Narrow Indication Is the Regulatory Strategy for AI Medical Devices →](/field-notes/2026-05-31-narrow-indication-ai-medical-device-clearance) [MDUFA VI Just Reset the De Novo and Pre-Sub Mechanics for Device Founders →](/field-notes/2026-05-27-mdufa-vi-de-novo-presub-credits-founder-mechanics) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Medicare's RAPID Coverage Pathway Moves the Work Earlier for Founders | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-06-07-rapid-coverage-pathway-reimbursement-cliff [← All Field Notes](/field-notes) June 7, 2026 Medical Devices FDA clearance is the moment most device founders treat as the finish line. It is not even the starting line for the thing that actually pays the bills, which is whether Medicare will cover the device and pay a hospital enough to want to buy it. A new pathway from CMS and the FDA just moved the timeline on that question, and the change is both better and more demanding than the headlines suggest. The headline says two months instead of a year. The fine print says the real work happens before you ever enroll your first patient. ## What RAPID Actually Changes On April 23 CMS and the FDA jointly announced the RAPID Coverage Pathway, a way to compress the gap between FDA authorization and Medicare coverage. Today a device can clear the FDA and then wait a year or more for a national coverage decision, selling into a void while it waits. Under RAPID, per MedTech Dive, CMS issues a proposed national coverage determination the same day the FDA authorizes the device, opens a thirty-day comment window, and can have national coverage in place as soon as two months after market authorization. It is not open to everyone. The pathway is for breakthrough-designated devices that address an unmet need for Medicare patients. Class III devices qualify; Class II devices have to be in the FDA’s Total Product Life Cycle Advisory Program first. CMS says roughly forty devices qualify right now, with about twenty more potentially eligible, and the older Transitional Coverage for Emerging Technologies route pauses for new candidates. FDA Commissioner Marty Makary framed it as cutting red tape; CMS Administrator Mehmet Oz framed it as the two agencies getting in the room together earlier. That second framing is the one founders should sit with. ## The Evidence Work Moves Years Earlier Here is what “earlier” means in practice. To earn same-day coverage, a device has to have run an investigational study that enrolled Medicare beneficiaries and measured the clinical outcomes both the FDA and CMS agreed on up front, according to a Latham & Watkins analysis of the program. One trial, designed before a single patient is enrolled, now has to satisfy two agencies that have historically asked for different things. The FDA wants to know the device is safe and does what you say. CMS wants to know it improves outcomes for a seventy-year-old in a way worth paying for. RAPID asks you to answer both in the same protocol. This is a product decision, not a reimbursement-consultant decision, and it lands squarely on whoever owns the framing of the product. When I was at Galen, the call that mattered was how we defined the device: we framed it as a tool that improves a surgeon’s access to anatomy, not as a remote manipulator competing with the incumbent, and that framing was what made the regulatory pathway work. RAPID adds a second axis to that same decision. It is no longer only “what is this device such that the FDA accepts it.” It is also “what outcome, measured how, will make a CMS actuary cover it.” Both answers now have to be designed into the study at the IDE stage, which is years before launch. Bolt coverage strategy on after clearance and you have already missed the window RAPID opened. ## Covered Is Not the Same as Paid There is a second catch, and it has a live deadline. Alongside RAPID, CMS has proposed repealing the New Technology Add-on Payment alternative pathway in its FY2027 inpatient payment rule, with the comment period closing June 9. That add-on payment is the extra Medicare reimbursement that helped hospitals afford brand-new devices in the years before standard rates catch up. Strip it out, and Latham & Watkins flags the obvious risk: a device can be covered on paper yet pay a hospital so little that no one adopts it. Faster coverage with thinner payment is not automatically a win. For a founder, the lesson is to model the payment, not just the coverage. A national coverage determination is permission to be reimbursed. The payment rate is whether the hospital makes or loses money every time it uses your device, and that number decides adoption. If your plan stops at “we will be covered,” you have planned for half of the problem. Dave’s take I have watched founders treat FDA clearance as the day the company becomes real, then spend the next eighteen months learning that a covered device a hospital loses money on is not a product. RAPID is a genuine gift if you design your evidence for both agencies from the start, and a quiet trap if you treat coverage as something that happens to you after the FDA says yes. If you are scoping an IDE study right now and no one is asking which outcomes a Medicare actuary will pay for, you are building next year’s reimbursement problem this year. From Dave’s video library Dave on testing the assumption that decides everything before you build around it. RAPID makes reimbursement that assumption: design the evidence for both agencies from the start or discover the gap two years too late. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [FDA 510(k) Exemptions Are Expanding. They Can Erase Your Moat →](/field-notes/2026-06-06-fda-510k-exemptions-predicate-moat) [Spectral AI’s De Novo Clearance: A Lesson in Non-Dilutive Capital →](/field-notes/2026-06-01-spectral-ai-de-novo-barda-non-dilutive-capital) [MDUFA VI Just Reset the De Novo and Pre-Sub Mechanics for Device Founders →](/field-notes/2026-05-27-mdufa-vi-de-novo-presub-credits-founder-mechanics) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # The 2026 Hardware Funding Records Hide a Higher Bar, Not Easier Money | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-06-08-hardware-funding-records-higher-bar [← All Field Notes](/field-notes) June 8, 2026 Commercialization A founder sent me the Crunchbase headline last week with one line on top: the money is back, raise now. Defense-tech venture funding had already broken its all-time record and we were barely into June. He read it as a green light. I read the same number and saw the opposite for most of the founders I talk to. The record is real. So is the part that never makes the headline. This capital is piling into a small number of companies, and the bar to be one of them has moved into work most hardware founders keep putting off. ## What the records actually say The defense figure is not hype. Crunchbase reported that companies in its military, national security, and law enforcement categories have taken in $14.6 billion in venture funding so far this year, already past the $9.6 billion the whole sector raised in all of 2025, and it happened across roughly 107 rounds in five months. The names doing the lifting are the usual ones: Anduril closed a $5 billion Series H at a $30.5 billion valuation, Shield AI a $2 billion Series G, Saronic a $1.75 billion Series D. Medtech is running warm too, on pace for one of its strongest years since 2021. Look under the totals, though, and the shape changes. Across med-device and health tech the count of deals has been shrinking while the average check per deal keeps climbing. More dollars going to fewer companies. That is not a rising tide lifting every boat, it is a sorting. The capital that looks abundant in aggregate is getting harder to reach one founder at a time, because each surviving check is bigger and the list of who clears the bar to receive it is getting shorter. ## Concentration is not abundance I say the same thing to every founder who comes to me in the middle of a raise: venture capital is not free, it is expensive money, and you had better know the exact milestone you are buying with it. A concentrating market makes that rule harder to honor, not easier. The post-money valuation you accept this round becomes the bar you have to clear at the next one, and when money stacks into fewer companies at bigger numbers, the distance between the price you take and the traction you can show stretches wider. The down round on the far side gets more likely, not less. When a founder shows me a billion-dollar raise and a deployment plan that is still mush, the number is not the interesting part. What I want to know is what they are going to do with the cash and what they will have to put on the table at the next round to defend the price they just set. Most of the celebrating I watch skips straight past that question, and the founders skipping it are the ones who walk into the next raise unable to answer it. ## What actually unlocks a check now So what separates the companies getting funded from the ones getting passed over? Less of it is the technology than founders want to believe. Read what device investors now expect on the table by a Series A and it is blunt: a specific FDA pathway, named and defended, 510(k) or De Novo or PMA, sitting next to early reimbursement work, the CPT or HCPCS codes the product would bill under and the budget case for the hospital that has to pay. Not a prototype that runs. A defended route from the prototype to a paying customer. Change the vertical and the words change, not the shape. A defense founder needs a believable procurement road, not a demo reel. A climate-hardware founder needs an offtake agreement, not a better cell chemistry. The capital is chasing the same thing in each case: proof that someone already did the work of turning a working device into a fundable path to market. That work is product, not engineering and not the pitch. It is the framing of the pathway itself, and it is the piece founders most often defer and then cannot produce when the term sheet conversation finally arrives. Dave’s take The 2026 funding records are not telling hardware founders that money got easier. They are telling them money got pickier, and it is sorting on the one thing a working prototype cannot prove on its own: that there is a credible, defended path from the bench to a buyer. I would rather a founder spend this quarter writing their regulatory and reimbursement pathway down in plain language than spend it polishing a deck around a number someone else raised. The check follows the path now. It stopped following the demo a while ago. From Dave’s video library Dave on the proof that cannot lie when everyone else will. Record defense-tech funding is concentrating into fewer companies for the same reason: a working prototype proves nothing about the path to a buyer, and investors have stopped pretending otherwise. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Spectral AI’s De Novo Clearance: A Lesson in Non-Dilutive Capital →](/field-notes/2026-06-01-spectral-ai-de-novo-barda-non-dilutive-capital) [The Pentagon’s Drone Gauntlet Is a Manufacturing Test →](/field-notes/2026-06-02-pentagon-drone-gauntlet-manufacturing-readiness-test) [The Pentagon’s Deal Team Six and the Defense-Hardware Founder Path →](/field-notes/2026-05-25-pentagon-deal-team-six-defense-hardware-founder-path) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # When a Quality System Catches the Problem and Still Can’t Fix It | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-06-09-quality-system-capa-effectiveness [← All Field Notes](/field-notes) June 9, 2026 Medical Devices For too long at one of my early companies, every engineering change order needed my signature before it could ship. I told myself the signature kept product and engineering aligned. The honest version is that it was mostly ego, and underneath the ego sat a mistake I now watch founders make at every size: I had confused my attention with a system. A signature is not a quality system. A quality system is the thing that catches a problem, fixes it, and proves the fix worked, whether or not the founder is in the room. I thought about that signature again this week reading the FDA’s warning letter to Medline, because it is the same mistake written at industrial scale. ## What the FDA actually found The device at the center of it is unglamorous: a NAMIC angiographic control syringe, the kind used to push contrast dye during imaging. In a warning letter dated March 25, the FDA said excess silicone was migrating to the connector and letting the syringe unwind off the manifold during a procedure, which can pull air into the line. Air in an arterial line is an embolism. The agency was blunt about how the company had scored that risk, writing that Medline’s “declaration of a low risk to health was inconsistent given the potential for air embolisms.” Here is the part that should stop a founder. This was not a problem nobody saw. Medline opened a corrective action, CAPA-01872, back in June 2023 as disconnection complaints climbed. The response was to scrap affected inventory and clean the machines more often. By the FDA’s account that never amounted to a real root-cause investigation, and more than two years later the agency still found the corrective action inadequate, with 177 medical device reports, hundreds of complaints, and four serious injuries logged in the meantime. The detection worked. The closing of the loop did not. The same letter flagged the housekeeping too. An investigator observed particulate accumulating on top of manufacturing equipment in the cleanrooms, and the last documented cleaning of that equipment was logged back in November 2023. A quality system was running the whole time. It was generating records, opening actions, and filing reports. What it was not doing was the one job that matters, which is making the defect stop. ## Detection is the easy part Almost every hardware founder I work with has the detection half handled. Complaints come in. A log fills up. Somebody opens a ticket. The half that fails is the second one: turning a found problem into a verified fix and proving the verification held. That is what a corrective action is for, and it is the single most common place I watch young device companies fake it. They open the action because the quality manual says to, they write down a containment step that looks like progress, and they never go find the cause. A year later the same defect is still shipping. For a founder who is still pre-market, this is not abstract. When you file, and later if an acquirer runs diligence, your corrective-action history gets read like a credit report. An open action that never closed is worse than a clean record, because it is written proof that you knew about a problem and could not fix it. The FDA does not have to build a case that you were careless. Your own quality system already documented it for them. This is why I now treat the quality system as product work and push founders to staff it as product work, not as a binder a consultant assembles before an audit. The signature I clung to for years felt like control, but it was one human standing in a doorway dressed up as a process. Medline had the process and still could not close the loop, which is the more expensive version of the same failure, because it scales. A bottleneck slows one company down. A quality system that detects and does not fix ships the defect by the case. Dave’s take The temptation, at five people and at five thousand, is to treat the quality system as the document you hand the auditor. It is not. It is the machinery that decides whether a problem you already found actually gets fixed before it reaches a patient. I would rather a founder spend one week making a single corrective action genuinely close, cause named, fix verified, evidence attached, than spend a quarter making the manual look complete. The binder protects no one. The closed loop does. From Dave’s video library Dave on why delegation fails on trust, not process, and the difference between delegating and abdicating. Medline's quality system caught the problem in 2023 and still never closed it, which is abdication with paperwork. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [The Neurosurgical Supply Shortage Reveals Supplier-Qualification Risk →](/field-notes/2026-05-28-neurosurgical-supply-shortage-supplier-qualification) [Closed-Loop Insulin Delivery: The Algorithm Is the Easy Part →](/field-notes/2026-06-05-closed-loop-insulin-delivery-validation) [FDA 510(k) Exemptions Are Expanding. They Can Erase Your Moat →](/field-notes/2026-06-06-fda-510k-exemptions-predicate-moat) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # GM’s Sodium-Ion Bet: Why It’s Selling Cells, Not Systems | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-06-10-gm-sodium-ion-grid-storage-product-boundary [← All Field Notes](/field-notes) June 10, 2026 Climate Hardware The most interesting thing General Motors said this week was not about a battery. It was about where to stand. GM is moving into grid and data-center energy storage, one of the most contested hardware markets going right now, and it has decided not to build the storage systems. It is going to make the cells and sell them to the company that does. That reads like a smaller ambition than the headlines suggest. I think it is the more disciplined one, and it is the call I spend a good part of my week pushing founders to make. ## What GM committed to Start with the facts, because the shape of the deal matters more than the headline number. GM has put $900 million toward commercializing new battery chemistries, including its own development center, and is building sodium-ion cells with trial production targeted for 2028. It does not plan to sell finished storage systems. It plans to sell those cells to Peak Energy, a startup that integrates them into grid and data-center installations. In parallel it sells lithium iron phosphate cells to LG Energy Solution for interim systems, and through Redwood Materials it is routing retired EV battery packs into stationary microgrids, including a 12-megawatt system at a Crusoe data center in Nevada and a smaller install at one of GM’s own plants in Michigan. The chemistry choice is the tell. TechCrunch reports the sodium-ion cells are cheaper, longer lasting, and less prone to overheating than the lithium that goes into an electric car, at the cost of being bigger and heavier for the same amount of stored energy. Peak says systems built on them can drop the cooling and fire-suppression hardware entirely. None of that helps you in a vehicle, where every kilogram counts. All of it helps you in a box that sits bolted to a concrete pad and never moves. ## Why the chemistry is a product call, not a science one Sodium-ion is not a better battery than lithium in the way a car buyer would measure. It holds less energy per kilogram, which is exactly why nobody is putting it in a sports car. For a battery parked next to a data center, weight is free. What costs you there is dollars per kilowatt-hour, how many charge cycles the cell survives, and whether it can sit safely next to a building full of servers without a fire-suppression budget. GM picked the chemistry whose weakness does not apply to the job and whose strengths do. That is the whole move. Founders get this backwards constantly. They chase the best number on every line of the spec sheet, because a high spec feels like progress and a traded-away spec feels like a concession. But the application only rewards two or three of those numbers, and it quietly penalizes the rest. A battery optimized for energy density is the wrong battery for a stationary site, even though it looks better on paper. Decide the duty cycle first and it tells you which specs to fight for and which to give up without apology. The team that defines the job wins the spec argument before it starts. ## Deciding which layer is the product The harder discipline is the second decision: GM chose to be the cell, and let Peak own the system built around it. I learned how hard that call is to force years ago at Ascend Communications. We had something like 30 million lines of code running most of the world’s dial-up access, and the company, which thought of itself as a hardware company, could not see that the software was the product until somebody named it and drew a boundary around it. I pushed to call it an operating system and sell it as one. I did not write a line of that code; the engineering teams did. My job was to recognize which layer we actually sold and say so out loud. That is the move GM is making here. It is not trying to own the cell, the enclosure, the controls software, and the field install all at once. It is going to be very good at one layer and sell into the companies that own the others. For a founder burning runway, that kind of narrowing is survival, not timidity. The instinct runs the other way, because owning more of the stack feels like capturing more of the value. Most of the time it just means you are mediocre at five things instead of excellent at one, and the customer can tell. Dave’s take Two decisions are buried in this announcement, and both are product decisions wearing a chemistry costume. One is matching the spec you chase to the job the product actually does. The other is choosing which layer of the stack you will own and refusing the rest. GM is worth more than a hundred billion dollars and still had to make both calls on purpose. A founder with eighteen months of runway has less room to get them wrong, not more. Pick the duty cycle, pick the layer, and let the parts you were never going to win belong to someone else. From Dave’s video library Dave on the three jobs a founder keeps on purpose and everything else they refuse. GM just made the same call at industrial scale: sell the cell, refuse the system, and own only the layer where you are actually best. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [The Climate-Hardware IPO Wave Is a Demand Story, Not a Tech Story →](/field-notes/2026-05-30-climate-hardware-ipo-wave-demand-not-technology) [The Humanoid Robot Is Physical AI’s Hardest Bet to Ship →](/field-notes/2026-06-04-humanoid-robots-narrow-form-factor-edge-ai) [The 2026 Hardware Funding Records Hide a Higher Bar, Not Easier Money →](/field-notes/2026-06-08-hardware-funding-records-higher-bar) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Dual-Use Hardware: When Battlefield Drone Tech Comes Home | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-06-11-dual-use-counter-drone-comes-home [← All Field Notes](/field-notes) June 11, 2026 Defense Hardware This month the Pentagon took a system built to knock drones out of the sky, ran it for two days at a Marine air station on the Arizona border, and then cleared it for use across the entire U.S. military. The system is CACI’s SkyValor, and the task force behind it has been quietly moving hundreds of millions of dollars into counter-drone hardware pointed, in part, at the homeland. Most of the coverage filed this under defense procurement. I read it as the clearest example this year of something I press on every hard-tech founder with a defense-adjacent platform: the line between the battlefield version of your product and the civilian version is thinner than you want it to be, and which road you build for is a choice you make on purpose or by drift. ## The twelve-lines version of dual-use Years ago I toured a research group working on a satellite robot. The idea was to launch it into orbit, have it chase down satellites that had run out of fuel, couple with them, drill into the tank, top them off, and patch the hole. Clever piece of engineering. I turned to the person next to me and said the difference between a robot that heals a satellite and one that kills satellites is probably twelve lines of code. He gave me a knowing look and said, that may be, but we don’t talk about that. That is the whole dual-use problem in one exchange. The propulsion, the rendezvous, the autonomy, the hardware all transfer. Only the target and the intent change. Look at what SkyValor actually does and the same thinness shows up. It senses autonomously around the clock, picks up threats more than forty miles out in some cases, and defeats them without firing a shot, jamming their signals or snaring them in nets from almost four miles off. Strip the word “threat” off that description and you have airport perimeter security, prison airspace control, stadium and critical-infrastructure protection, wildfire and search-and-rescue tracking. The radar does not know whether the drone overhead is a smuggler, a hobbyist, or hostile. The rules of engagement do. ## What does not transfer is the validation history The same week, the same task force handed Perennial Autonomy a $500 million three-year contract for its Merops interceptor. The number that matters there is not the contract. By the company’s account the Merops has shot down more than four thousand Russian drones over Ukraine since 2024, and the Army has bought roughly thirteen thousand of them at about fifteen thousand dollars each. That combat record is the reason it costs a fraction of a missile and can field after a short trial instead of a decade-long program. The airframe is replicable. Four thousand real intercepts against an adversary who adapts every week are not. For a founder, that is the part worth sitting with. The thing a competitor cannot lift off your spec sheet is the environment where your system has piled up real, adversarial failure data. I have watched founders burn a year chasing a benchmark in a clean lab when a quarter of that effort aimed at messy field deployment would have bought them something no rival could fake. If your platform has two roads, the question is which one is generating that hard-won data, and whether it is the road you actually want to be your larger market. Dave’s take I will analyze the defense-procurement map all day. I am more careful about who I will help build on it. When a founder brings me a platform like this, the first thing I want is to see the two roads it can travel, the kinetic one and the one that comes home to an airport, a hospital, a power plant. If the only honest answer is the war road aimed at people, I am not their guy. If there is a real civilian road, it is usually the bigger market anyway, and the battlefield is just the customer rich enough to pay for the hard parts first. From Dave’s video library Dave on Barnum's tools as neutral and intent as the divider, including the dual-use detector. A counter-drone system moving from Ukraine to the US border is the two-roads question in physical form. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [The Pentagon’s Drone Gauntlet Is Really a Manufacturing Test →](/field-notes/2026-06-02-pentagon-drone-gauntlet-manufacturing-readiness-test) [Pentagon Deal Team Six and the New Path to Market for Defense Hardware Founders →](/field-notes/2026-05-25-pentagon-deal-team-six-defense-hardware-founder-path) [The Humanoid Robot Is Physical AI’s Hardest Bet to Ship →](/field-notes/2026-06-04-humanoid-robots-narrow-form-factor-edge-ai) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Penumbra Thunderbolt Stroke Clearance: A Platform Play, Not a New Device | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-06-12-penumbra-thunderbolt-stroke-platform-extension [← All Field Notes](/field-notes) June 12, 2026 Medical Devices On June 11 Penumbra said the FDA had cleared THUNDERBOLT, the first computer-assisted vacuum thrombectomy device authorized to pull a clot out of the brain during an acute stroke. Read the announcement and it lands as a product launch, a faster way to clear a blocked vessel. I read it as something quieter and more useful to a founder. Penumbra did not invent a new way to treat stroke. It took the aspiration engine that already runs its vascular business, added one new wrinkle, pointed it at the hardest indication it serves, and shipped it inside the catheters it already sells more of than anyone. The interesting decision was not the engineering. It was the discipline to extend instead of build. ## A re-indication, not a reinvention Strip THUNDERBOLT down to what is actually new and the list is short. The core is Penumbra’s ENGINE, the same vacuum aspiration platform behind the company’s existing thrombectomy line. The catheters it ships with, the RED 62 through RED 72 family, are already in the field and already lead their category. The genuinely new content is modulated aspiration, a pulsing rather than constant pull that is meant to detect a clot, fatigue it, and ingest it in one pass instead of grinding at it. That is a real advance in the working end of the device. Everything around it rides an installed base Penumbra had already built. For a hardware founder, that ratio is the lesson. Your second product should travel the regulatory ground and the sales channel your first product already cleared, not start from a blank page. This holds for anyone whose system has a core engine that can be re-pointed at an adjacent problem: an aspiration pump, a sensing module, an actuation stage, a power-conversion block. Penumbra defined its engine years ago in a way that let stroke become a new indication rather than a new company, and because the catheter franchise was already there, the device walked into the cath lab pre-sold. Stroke is also an enormous indication. Someone in the U.S. has one roughly every forty seconds, so a faster, cleaner clot grab has a population behind it worth the years of platform work. ## Why “is it better?” is the wrong question The reflex around any new thrombectomy device is a head-to-head: does modulated aspiration pull clot more completely than the last catheter on the tray? I think that is the wrong question, and I say a version of it in almost every talk I give about surgical robotics. People want to argue whether the robot makes an individual surgeon better. It does not matter. What mattered about the da Vinci is that it grew the population of surgeons who could reliably perform a minimally invasive procedure by orders of magnitude. More capable operators means more patients get the option. That is how we all win. Mechanical thrombectomy has the same shape. Its value was never about one neurointerventionalist at one flagship center getting marginally better at a single occlusion. The shift was that it moved stroke from a drug-and-wait window into a procedure a growing number of stroke centers can run quickly, with a repeatable result. Penumbra’s real leverage with THUNDERBOLT is making a clean first-pass removal more repeatable across that whole population, not winning a bench comparison. If you are commercializing a device like this, position it against the realistic counterfactual, the incomplete result at the center that does not have the best hands in the country, not against the best operator on his best day. That is the comparison the buying hospital is actually making. Dave’s take Penumbra drew its product boundaries well enough, and early enough, that stroke could arrive as a packaging decision rather than a ground-up program. That is the win, not a cleverer way to grab a clot. When a founder walks me through their architecture, that is the call I push on hardest. Not what is our first device, but what is the engine, and how many indications can it carry before we have quietly become three companies. Get that boundary right and your second clearance is mostly logistics. Get it wrong and every new indication is a fresh build with a fresh burn rate. From Dave’s video library Dave on the foundational work that keeps compounding long after the first win. Penumbra drew its product boundaries years ago so that stroke could arrive as a packaging decision, which is what that early unglamorous work buys you. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [A Narrow Indication Is the Regulatory Strategy for AI Medical Devices →](/field-notes/2026-05-31-narrow-indication-ai-medical-device-clearance) [FDA 510(k) Exemptions Are Expanding. They Can Erase Your Moat →](/field-notes/2026-06-06-fda-510k-exemptions-predicate-moat) [Surgical Robotics in 2026: Knowing When Not to Launch →](/field-notes/2026-05-29-surgical-robotics-when-to-launch-cmr-versius) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # NEURA’s $1.4 Billion Humanoid Bet, and Why I’d Bet Narrow | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-06-13-neura-1-4b-physical-ai-humanoid-form-factor [← All Field Notes](/field-notes) June 13, 2026 Industrial Robotics When a robotics company raises up to $1.4 billion, the reflex is to treat the number as the story. NEURA Robotics, a German firm most founders were not talking about a year ago, announced on June 10 what its lead investor Tether is calling one of the largest physical-AI rounds on record. The money is pointed at cognitive humanoids, a network of training environments the company calls NEURA Gyms, and serial production toward millions of robots by 2030. The consensus take is that this is what the future of physical AI looks like, and that the humanoid is the form factor that gets there first. I think that framing is backwards, and the reason has little to do with whether NEURA is a good company. ## The form factor the buyers are actually pulling toward Watch where the most demanding hardware buyers are spending, and they are not asking for a robot that can do everything. The clearest signal I track sits in defense procurement, where recent requests have leaned on two requirements that quietly exclude the humanoid-plus-hyperscale architecture: the machine has to run untethered, and it cannot depend on a live internet connection to think. Both push the intelligence onto the device, onto edge compute, and away from the trillion-dollar data centers that the consumer-AI story assumes will keep doing the heavy lifting. A robot that stops working when the network drops is a liability the Pentagon has decided it will not buy. That constraint also dictates shape. A human-simulated body has far more degrees of freedom and far messier contact with the world than a quadruped or a fixed arm, which makes it the hardest thing to run autonomously on a processor it can carry. Surgical robotics is the existence proof that the narrow path works. I watched computer vision and machine learning run on edge processors inside constrained surgical systems years before “physical AI” was a fundraising category. The pattern was real then, and it scaled because the form factor was disciplined. For a founder building robots or automation, the move is simple to state and hard to hold to: match your form factor to the procurement signal, untethered, narrow, direct, not to the venture narrative that rewards a general-purpose humanoid because it demos well on a stage. ## What “up to $1.4 billion” actually obligates you to do Notice the phrasing in NEURA’s own announcement: up to $1.4 billion, conditional, not a closed round wired to the bank. That is not a knock on the company, it is a reminder that the headline and the cash are different things. Where NEURA earns some credit is the part most mega-rounds cannot show, an order book and deployment pipeline the company says already exceeds $1 billion. At least there is a stated plan for the money, which is more than I can say for a lot of the wave. But the discipline that should travel home to founders watching this is the one nobody celebrates. Venture capital is expensive money, and the post-money valuation you accept today becomes the bar you have to clear at the next round. Take more than your milestones can justify and you set yourself up for the down round that dilutes you into irrelevance. Before a physical-AI or automation founder takes a fraction of a round like this, I want a specific answer to two questions: what exactly does this capital buy that you cannot earn another way, and what operating profile do you have to deliver next time to defend the number you just accepted. Cheering the headline while the deployment plan is mush is how you walk into that down round. Dave’s take I am not betting against NEURA, I am betting against the idea that one general-purpose body wins physical AI. There is no one robot to rule them all. Humanoids earn their place where the job genuinely needs a human shape, the cobot working in a space built for people, and task-engineered machines take everything else, usually cheaper and sooner. When a robot’s main credential is that it gives a really good demo, I get cautious. The durable companies in this wave will be the ones that knew exactly what their money was for. From Dave’s video library Dave on why a giant valuation is not the same as a real business, and why fabricated value is the thing to watch. A $1.4 billion humanoid raise is an obligation before it is an achievement. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [The Humanoid Robot Is Physical AI’s Hardest Bet to Ship →](/field-notes/2026-06-04-humanoid-robots-narrow-form-factor-edge-ai) [The 2026 Hardware Funding Records Hide a Higher Bar, Not Easier Money →](/field-notes/2026-06-08-hardware-funding-records-higher-bar) [Battlefield Drone Tech Came Home. The Dual-Use Line Is the Lesson. →](/field-notes/2026-06-11-dual-use-counter-drone-comes-home) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # The Robot Is the Giveaway, the Instruments Are the Business | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-06-14-surgical-robot-instruments-are-the-business [← All Field Notes](/field-notes) June 14, 2026 Surgical Robotics Early in my years in surgical robotics, I learned the part of the business that never makes the pitch slide. The robot is the giveaway. The instruments are the company. A hospital can amortize a capital system over years, but it cannot escape the steady invoice for the cautery hooks, the graspers, and the scissors that get used up case after case, and that recurring line is where the margin actually lives. So when Restore Robotics announced in early April that the FDA had cleared it to remanufacture two more of Intuitive Surgical’s da Vinci instruments, I did not file it as a small aftermarket item. I read it as someone testing, in public, how well defended that recurring line really is. ## The business hiding inside the capital equipment Surgical robotics runs on a razor and blades model, and the da Vinci is the textbook case. The system itself is a large one time sale. The instruments are the annuity. Intuitive treats each instrument as usage limited, so once a unit reaches its cap the hospital buys another, and across a high volume program that adds up to far more revenue than the robots ever did. That is not an accident of the technology. It is product architecture, a deliberate decision about where the recurring money sits, made long before the first unit shipped. I have argued for years that the da Vinci’s contribution was never about whether a robot assisted surgeon beats the rare expert who can do the procedure by hand. It was that the platform grew the population of surgeons who can perform a minimally invasive operation by orders of magnitude. More surgeons, more patients, and that is how everyone wins. But the same population scale that makes the clinical case also makes the instrument stream enormous, and a large, predictable revenue line is exactly the kind of thing a competitor eventually decides is worth attacking. ## Know what actually defends your recurring revenue Restore now holds four FDA 510(k) clearances to remanufacture da Vinci instruments: the two new cautery tools, a permanent cautery hook and a permanent cautery spatula, on top of monopolar curved scissors for the Xi and the older Si system, per MedTech Dive. Its chief executive described the clearances as a step toward transforming the economics of robotic surgery. Strip the framing and the claim is direct. Take the most dependable margin in the category and hand a piece of it back to the hospital. The lesson I want founders to take is underneath it. Every company that sells hardware plus a consumable rests on a question most never ask out loud. What actually defends the recurring revenue? There are three answers, and they age very differently. Sometimes it is physics, because the consumable is genuinely consumed, a reagent that is spent or a filter that clogs. Sometimes it is regulation or intellectual property, because the item is durable but the maker has wrapped it in a usage limit, a clearance, or a patent. And sometimes it is nothing but habit. Intuitive’s instrument moat leans hard on the second kind. The instruments are durable enough to remanufacture, which tells you the limit on reuse is a design and regulatory construct, not a property of the steel. A construct can be matched. Restore did not build a better instrument. It used the same 510(k) predicate logic that Intuitive’s own clearances rely on and aimed it back at the company that built the category. If your recurring revenue is defended by a rule rather than by physics, assume a well funded competitor will eventually clear a substitute through your own predicate, and price and plan for that day before it lands. Dave’s take I am not rooting for or against Intuitive here. I am pointing at the thing I wish more hardware founders internalized at the prototype stage. The recurring revenue is the business, and how well that revenue is defended is a product decision you make years before anyone comes for it. Build the razor so the blades are protected by something real, or accept that someone will eventually sell those blades back to your customer at a discount. Either path can work. Choosing it by accident cannot. From Dave’s video library Dave on finding the handful of numbers your business actually runs on. For a surgical robotics company that number is instrument revenue per system, and Restore's remanufacturing clearance just showed what happens when it is undefended. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Penumbra’s Thunderbolt Stroke Clearance Is a Platform Play →](/field-notes/2026-06-12-penumbra-thunderbolt-stroke-platform-extension) [FDA 510(k) Exemptions Are Expanding. They Can Erase Your Moat →](/field-notes/2026-06-06-fda-510k-exemptions-predicate-moat) [When a Quality System Catches the Problem and Still Can’t Fix It →](/field-notes/2026-06-09-quality-system-capa-effectiveness) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # GT Medical's $100M GammaTile Raise: Evidence, Not Clearance | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-06-15-gt-medical-gammatile-evidence-unlocks-capital [← All Field Notes](/field-notes) June 15, 2026 Medical Devices A medical device company can hold an FDA clearance for nearly a decade and still not own the thing investors actually pay for. GT Medical Technologies is the clean example. Its GammaTile implant has been cleared since 2018, expanded to newly diagnosed tumors in 2020, and selling the entire time. Yet the hundred million dollars did not arrive until last week, after a randomized trial put the device against the standard of care and won. The clearance was years old. What changed was the evidence, and the capital followed it. ## What the money was actually buying GT Medical closed an oversubscribed $100 million Series E on June 9, led by new investor Viking Global, per the company and MedTech Dive. GammaTile is a bioresorbable implant placed in the cavity at the moment of brain tumor surgery, so radiation begins immediately instead of weeks later in a separate course. It has been cleared and on the market for years. What moved the round was the ROADS data shared at this year’s ASCO meeting: a Phase 3 randomized trial in newly diagnosed brain metastases where the twelve month surgical bed recurrence rate was 1.3 percent with GammaTile against 15.4 percent for surgery followed by stereotactic radiation, the current standard of care. Chief executive Per Langoe called the financing validation of the device’s potential to become a standard of care itself. ## Clearance and evidence are two different games A clearance answers one question: may you sell this. Comparative evidence answers a different one: should anyone switch to it. I have argued for years that framing the indication is the regulatory strategy, because how you define a device determines how much evidence the FDA will demand to clear it. At Galen we framed our system as a surgeon access stabilizer rather than a remote manipulator, which kept the bar on engineering and usability instead of a full clinical outcomes trial. That is the right first move, because it gets you to market cheaply and fast. But the outcomes trial you deliberately skipped is the same one that later moves adoption, reimbursement, and growth capital. GT Medical sequenced it well: clear on the cheaper framing, get on the market, then fund the randomized comparison once there is revenue and a reason to run it. There is also a reason the data is believable, and it comes down to workflow. GammaTile closes a gap. The patient is already in the operating room for the resection, and the radiation source goes in during that same procedure, so there is no three to four week wait between surgery and the start of radiation and no second appointment to book. A device that slots into the operation already happening carries a real adoption advantage over one that asks the surgeon and the patient to add a step. The pattern reaches well past MedTech. A defense hardware startup can pass the qualification testing that lets it sell to a program office and still be waiting on the field data that proves it beats the incumbent it wants to displace. A diagnostics company can clear its assay and still need the head-to-head accuracy study before a lab will switch vendors. Permission to sell and proof that you are worth switching to are different milestones, funded at different times, and treating them as one thing is how a cleared product can sit flat for years. Dave’s take The trap I watch hardware founders walk into is treating the clearance as the finish line when it is the starting line. Clearance buys you the right to go generate the evidence that actually builds the business, and the founders who stall are usually the ones who threw the party for the regulatory milestone and then never funded the comparison that would make a buyer change behavior. GT Medical spent years and a randomized trial proving its implant beats the alternative, and that is the work that turned a cleared product into a hundred-million-dollar one. From Dave’s video library Dave on selling by changing the buyer's math rather than their mind. GT Medical raised $100 million years after clearance because a randomized trial finally gave buyers a number that changed their behavior. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Penumbra’s Thunderbolt Stroke Clearance Is a Platform Play →](/field-notes/2026-06-12-penumbra-thunderbolt-stroke-platform-extension) [Medicare’s RAPID Pathway Moves the Work Earlier for Founders →](/field-notes/2026-06-07-rapid-coverage-pathway-reimbursement-cliff) [A Narrow Indication Is the Regulatory Strategy for AI Devices →](/field-notes/2026-05-31-narrow-indication-ai-medical-device-clearance) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Why OraSure Built a Collection Kit, Not a Better STI Test | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-06-16-orasure-sti-collection-kit-access-bottleneck [← All Field Notes](/field-notes) June 16, 2026 Diagnostics A diagnostics company can spend years chasing a more sensitive assay and still watch the test go unused, because the chemistry in the lab was never the part that decided whether anyone benefited. The part that decides is whether a person ever produces a sample to run. OraSure took that seriously. Last week it cleared a product most people will skim right past, a urine collection kit, and I think that was a sharper product call than another assay would have been. ## What OraSure actually cleared On June 11 the FDA cleared OraSure’s Colli-Pee Dx kit, developed through its DNA Genotek unit, for at-home collection of a first-void urine sample that screens for four sexually transmitted infections: chlamydia, gonorrhea, trichomonas, and Mycoplasma genitalium. A person collects at home, drops the kit in the mail, and the sample runs on Roche’s cobas 5800, 6800, and 8800 systems, the molecular analyzers many high-volume labs already own. OraSure did not build a new assay or a new analyzer. It built the step in front of both, with a stabilization chemistry it calls NucleoPrecision that holds the DNA and RNA at room temperature so a mailed sample is still good when it arrives. Chief executive Carrie Eglinton Manner framed the clearance as connecting people to care that is more accessible and private. That is marketing language, but it names the actual constraint. ## The bottleneck was never the test Screening assays for these infections are not the weak link. Labs have run accurate molecular STI panels for years. The weak link is human. A clinic visit, a clinician-collected swab, the time off work, the discomfort, the loss of privacy, all of it sits between a person and a result, and a lot of people who should be tested simply are not. A urine sample you collect at home and drop in the mail removes most of that. The win is not a better number on a sensitivity chart. The win is more people getting tested at all. I make a version of this argument constantly about surgical robots. The debate people want to have is whether a robot-assisted surgeon gets better outcomes than the small group who can do the procedure by hand. My answer is that it does not matter, because it is the wrong question. The point of the robot is to grow the population of surgeons who can do the procedure by orders of magnitude, which grows the population of patients who get a minimally invasive option. More surgeons, more patients. The same shape holds here. Is the assay better? Does not matter. More people will actually get tested, and that is where the value is. ## Where founders point the effort For a diagnostics or hard-tech founder, the lesson is about where you spend scarce product and regulatory effort. The impressive work, the novel assay, the cleaner algorithm, the more dexterous machine, is usually not the step that gates real adoption. The dull human step in front of it usually is, and it is the one founders skip because it does not demo well. OraSure pointed its clearance at the collection step and let Roche’s installed analyzers do the rest, which is a faster road than convincing labs to swap hardware. If you are building a diagnostic, ask which step actually loses patients between intent and result, and put the regulatory dollars there. If you are selling into hospitals, ask the same question about the workflow gap that keeps a clinically sound product sitting on the shelf. The accuracy you are proud of is table stakes once someone is in the room. Getting them in the room is the harder product, and the less glamorous one. Dave’s take The founders I watch stall are usually the ones who fell in love with the hardest, most impressive part of their own product and assumed adoption would follow from being good. It rarely does. Something dull and human sits between your technology and the person it is meant to help, and closing that gap is the job. OraSure cleared a collection kit and may reach more people than a marginally better test ever would. Find your version of that kit. From Dave’s video library Dave on the five questions that surface why customers really leave, which is rarely the reason you assume. OraSure asked the right question and found the bottleneck was never assay sensitivity, it was getting a sample to run at all. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Closed-Loop Insulin Delivery: The Algorithm Is the Easy Part →](/field-notes/2026-06-05-closed-loop-insulin-delivery-validation) [The Humanoid Robot Is the Most Photogenic Bet in Physical AI, and the Hardest to Ship →](/field-notes/2026-06-04-humanoid-robots-narrow-form-factor-edge-ai) [Spectral AI’s De Novo Clearance Is a Lesson in Non-Dilutive Capital →](/field-notes/2026-06-01-spectral-ai-de-novo-barda-non-dilutive-capital) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # How a 2018 De Novo Let Dexcom Put a Glucose Monitor on a Toddler | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-06-17-dexcom-otc-cgm-kids-de-novo-dividend [← All Field Notes](/field-notes) June 17, 2026 Medical Devices Sometime this summer a parent will pick a glucose sensor off a pharmacy shelf, near the reading glasses and the pregnancy tests, and stick it on the arm of a two-year-old. No prescription. No clinic visit. That is the part of this clearance the coverage is leading with, and it is the right hook. The part worth studying is that Dexcom could pull it off cheaply, and the reason traces back to a regulatory decision the company made eight years ago. ## What Dexcom actually cleared This month the FDA cleared the Dexcom Stelo Glucose Biosensor System for children as young as two who do not use insulin, the first time a continuous glucose monitor has been sold over the counter to anyone under eighteen. Stelo has been available without a prescription to adults since March 2024. The hardware did not change. The sensor still reads glucose every fifteen minutes, lasts up to fifteen days, and pairs with a phone. What changed is the labeled population, and the careful fence around it. Stelo is not for people who take insulin, not for anyone with a history of dangerous low blood sugar, not for dialysis patients, and it does not alert you to a crisis. The FDA was explicit that it is a supplement to medical care, not a replacement. That fence is not fine print. It is the whole reason a clinical-grade sensor gets to sit on a retail shelf. ## The decision that made it cheap Back in 2018, Dexcom's G6 became the first integrated continuous glucose monitor the FDA ever authorized, through a De Novo. A De Novo is what you file when there is no predicate to point at, when nothing like your device has been cleared before. Founders treat that as a punishment, a sentence to extra time and extra evidence. I think that is exactly backward. When the FDA routes you to a De Novo, it is telling you it considers your device moderate risk and inside the bounds it will accept. And if you get through, you do not just get cleared. You write the classification. You become the predicate every later entrant has to match, including your own next product. I watched this from the inside at Galen Robotics. Our pre-submission routed us to De Novo, we drafted the proposed regulation, and the FDA took our language almost word for word in a couple of days. We did not squeak through someone else's standard. We set it. That is what Dexcom did for the entire CGM category in 2018, and it is why this week's pediatric clearance was a label expansion riding on real estate the company already owned, not a fresh fight. ## Where founders should point the regulatory dollars If you are building a device with no obvious predicate, a novel diagnostic, a new class of wearable, a piece of software that does something the FDA has not classified yet, a De Novo is the asset, not the detour. Every expansion afterward, a new age group, an over-the-counter version, an added claim, becomes a cleaner, cheaper filing against the standard you authored. Dexcom has now run that play twice off one 2018 decision: first to over-the-counter for adults, now to children. The exclusions do the same work in the other direction. By drawing the fence to leave out insulin users, Dexcom kept Stelo out of the high-risk envelope where dosing decisions live, which is what keeps it over the counter at all. The population you deliberately leave out is as much a product decision as the one you chase. Spend the regulatory budget early, on the framing and the standard, not late, on defending a sloppy one. Dave’s take A De Novo is the cheapest long-term regulatory real estate a hardware founder can buy, if you treat the drafting as product work and do it deliberately. Dexcom made one moderate-risk call in 2018 and has been collecting on it ever since, all the way down to a sensor a parent can buy for a toddler. Most founders see the De Novo letter and feel dread. I see an invitation to write the rules of your own category. Take it, and write them like you will have to live inside them for a decade, because you will. From Dave’s video library Dave on the product leadership work founders end up doing themselves, and what it looks like done deliberately. Dexcom's 2018 De Novo is that work in regulatory form: one product decision that has paid rent for eight years. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [A Narrow Indication Is the Regulatory Strategy for AI Medical Devices →](/field-notes/2026-05-31-narrow-indication-ai-medical-device-clearance) [FDA 510(k) Exemptions Are Expanding. They Can Erase Your Moat →](/field-notes/2026-06-06-fda-510k-exemptions-predicate-moat) [Closed-Loop Insulin Delivery: The Algorithm Is the Easy Part →](/field-notes/2026-06-05-closed-loop-insulin-delivery-validation) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # How to Read the Pentagon's Drone-Budget Surge as a Hardware Founder | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-06-18-how-to-read-pentagon-drone-budget [← All Field Notes](/field-notes) June 18, 2026 Defense Hardware Picture a defense-hardware founder opening this week’s budget news and seeing one number: the Pentagon’s Defense Autonomous Warfare Group goes from $225.9 million this year to a requested $54.6 billion next year. The instinct is to treat that as a buy order and start sketching a bigger drone line. That instinct is wrong, and acting on it is how a good company builds the wrong thing. The budget is not asking for more airframes. It is a fairly precise description of how the Pentagon intends to acquire autonomy, and a founder who studies it learns what to build and, just as usefully, what to skip. ## What the $54.6 billion is actually pointed at The money follows a change in approach. The Defense Autonomous Warfare Group, or DAWG, replaced the Replicator Initiative late last year. Where Replicator was about getting hardware into the field fast, DAWG is built around software. The acting Pentagon comptroller, Jules Hurst, calls it a “pathfinder” that live-tests orchestration tools for autonomy. Of the requested $54.6 billion, about $1 billion is normal base budget and roughly $53 billion is flexible reconciliation money with a five-year window to spend it. The clearest tell is who is getting picked: Shield AI was selected to put its Hivemind autonomy software into the LUCAS program, the Low-Cost Uncrewed Combat Attack System. Retired General David Petraeus called DAWG the largest single commitment to autonomous warfare in history. Across the whole 2027 request, drones and counter-drone systems draw roughly $70 billion, and the biggest slice, $53.6 billion, goes to autonomy, platforms, and contested logistics. ## The org chart is part of the spec How the Pentagon plans to buy tells you as much as the dollars. On June 12, the Senate Armed Services Committee voted 18 to 9 to advance a $1.15 trillion defense bill that urges the Pentagon to stand up a Robotic and Autonomous Systems Combatant Command. It would be the twelfth such command, led by a four-star general, focused on acquiring small attack drones, getting them to troops, counter-drone defense, and unmanned surface vessels. Senator Roger Wicker framed the bill around artificial intelligence, autonomous weapons, and low-cost munitions. A combatant command exists to move procurement and delivery at scale. For a founder, that points at what actually gets bought: things you can make in volume and push to the field, not one beautiful prototype. It also fits a signal I have been watching for a while. The DoD’s recent requests keep asking for untethered, narrow, edge-autonomous machines, not internet-dependent general-purpose ones. General-purpose is a fundraising story. Direct and narrow is a delivery story, and the budget rewards delivery. If a founder shows me a general-purpose humanoid built to chase this kind of money, I get skeptical fast. The autonomy wins that hold up are narrow and engineered around one task, and a human silhouette usually drags back the very limits the machine existed to remove. ## The human-judgment line is a design constraint, not a footnote There is a catch buried in the enthusiasm. Lawmakers are already asking whether DoD Directive 3000.09, which requires “appropriate levels of human judgment” over the use of force, can hold up when you are orchestrating thousands of autonomous systems at once. That tension will not be settled in a press release. If you are building autonomy, treat it as a hard boundary you design to now, the way a medical-device team designs to a regulation before the product exists rather than after. Decide up front where human judgment enters, how it is logged, and how it survives scale. The teams that bake that in will still be standing when the rules get specific. The ones who bolt it on later will be redesigning under audit. Dave’s take A budget this size pulls every robotics founder toward the biggest, broadest version of the idea, because that looks like the version that matches the money. I would push the other way. The Pentagon is saying, in dollars and in org charts, that it wants narrow autonomy it can buy by the thousand and trust around people. The founders who win this cycle are the ones already building exactly that while everyone else is demoing a humanoid. Match the spec in front of you, not the headline number. From Dave’s video library Dave on extracting a confident decision from one big ambiguous number. A $54.6 billion autonomy request is exactly that kind of number: read as a spec it tells you what to build, read as a buy order it tells you nothing. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Dual-Use Hardware: When Battlefield Drone Tech Comes Home →](/field-notes/2026-06-11-dual-use-counter-drone-comes-home) [The Humanoid Robot Is Physical AI’s Hardest Bet to Ship →](/field-notes/2026-06-04-humanoid-robots-narrow-form-factor-edge-ai) [The Pentagon’s Drone Gauntlet Is a Manufacturing Test →](/field-notes/2026-06-02-pentagon-drone-gauntlet-manufacturing-readiness-test) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Sanctuary AI's Physical AI Claim Is One You Can Actually Check | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-06-19-sanctuary-ai-validated-number-not-demo [← All Field Notes](/field-notes) June 19, 2026 Industrial Robotics Most physical-AI announcements are built so you cannot check them. A humanoid crosses a stage, lifts a block, and a press release tells you the future of labor has arrived. None of it is verifiable, and by the time anyone could verify it, the funding round has closed. So when Sanctuary AI published a number this week that a customer can check every day, I paid attention. Not because the robot is special. Because the claim is falsifiable. ## What Sanctuary actually published On June 17, Sanctuary AI, the Vancouver company best known for its humanoid, said it had hit a 99.5 percent success rate on a wire-plugging task at a Tier 1 automotive supplier, at a cycle time of 2.54 seconds. The task is the unglamorous kind that has kept robots off production lines for years: guide a flexible wire into a moving target on a live conveyor, where the material shifts in your grip as you handle it. The percentage matters less than the bar it was measured against. Sanctuary ran the task against the customer’s own live production benchmark, not a lab rig of its own design, and it did the work on existing industrial robot arms instead of waiting for its humanoid to be ready. Olivia Norton, the co-founder and chief technology officer, put the logic in one line: physical-AI adoption is gated by AI that meets both the performance and the cycle-time requirement. That is the whole game. ## Why a number you can check is not a demo I spend a lot of energy telling founders that most of what gets sold as AI is rebranded machine learning wrapped in P.T. Barnum showmanship. The tool is genuinely useful. The marketing around it is built to be unfalsifiable: too dangerous to release, most powerful model ever, claims shaped so nobody can check them before the news cycle moves on. My test on any of it is two questions. Can the specific claim be disproven, and has anyone outside the company actually checked it? A humanoid demo fails both. Sanctuary’s number passes both. There is a specific figure, tied to a named task, measured against a benchmark the customer already runs on its own floor. If it were wrong, the customer would know by lunch. That is the line between a product claim and a hype claim, and it has nothing to do with how impressive the robot looks on video. ## Where this lands for hard-tech founders If you are building anything physical, sit with the cycle-time point. A robot that does the task but runs slower than the worker it replaces is not a product. It is a science project with good PR. The number that decides whether you have a business is not your demo metric. It is your performance against the workflow the buyer runs today. Sanctuary made two calls that founders routinely get wrong. It picked a narrow, real, contact-rich task instead of a broad capability story, and it ran that task on hardware its customer could already buy and support instead of betting adoption on a brand-new platform. Both choices trade spectacle for a number someone can sign off on. If you cannot state your claim as a figure a customer could disprove against their own operation, you do not have validation yet. You have a demo, and demos do not close. Dave’s take The fastest way to tell whether a hard-tech company is real is to ask what number it will let an outsider check. The teams I watch stall are usually the ones still polishing the demo, because the demo is where they feel strongest and safest. Sanctuary did the harder thing. It got measured on someone else’s floor, against someone else’s benchmark, and published the result. Build the version of your product that survives that, and stop selling the version that only works in your own lab. From Dave’s video library Dave on turning your operation into one honest number an outsider can check. Sanctuary did the industrial version: measured on a customer's floor, against the customer's benchmark, with the result published. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [NEURA’s $1.4 Billion Humanoid Bet, and Why I’d Bet Narrow →](/field-notes/2026-06-13-neura-1-4b-physical-ai-humanoid-form-factor) [The Humanoid Robot Is Physical AI’s Hardest Bet to Ship →](/field-notes/2026-06-04-humanoid-robots-narrow-form-factor-edge-ai) [How to Read the Pentagon’s Drone-Budget Surge as a Hardware Founder →](/field-notes/2026-06-18-how-to-read-pentagon-drone-budget) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # The FDA's Former Device Chief Is Advising Startups. What It Buys You | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-06-20-fda-device-chief-advising-medtech-startups [← All Field Notes](/field-notes) June 20, 2026 Regulatory The most senior medical-device regulator in the country spent most of fifteen years deciding which products reached patients and which ones went back for more data. This week he took a seat on the other side of the table. Dr. Jeffrey Shuren, who ran the FDA’s device center, joined the accelerator MedTech Innovator as an advisor to early- and mid-stage startups. My read is that this is genuinely good for founders, and the reason is narrower than the announcement makes it sound. An insider sharpens how you read the agency’s signal. He does not take the hardest part of the work off your desk. ## What an insider changes MedTech Innovator runs the largest accelerator for device, diagnostic, and digital-health companies, and it brought Shuren on to advise on regulatory strategy, policy trends, and market access. Its CEO, Paul Grand, said the point was to help portfolio companies avoid costly delays. That is the right word. Delays. Most founders treat an FDA interaction as a question-and-answer session. You ask, the agency answers, you proceed. It does not work that way. A pre-submission meeting is closer to a strategy presentation. You are showing the agency how you intend to define the product, and in how it responds, the agency is showing you where it thinks the risk sits. The skill is reading that signal correctly. Misread it and you can spend a year building tests around a premise the agency was quietly waving you off of. Someone who sat in that chair for years reads it the way you read your own native language. That is the genuine saving here, and most of it is the money you would have burned walking down the wrong path. ## What it can’t change There is a line an advisor cannot cross for you, and I learned where it sits the hard way. At Galen, we ran a pre-sub for our de novo. We told the FDA we would do a 16-subject cadaver study with residents and show that the robot let them nail the maneuvers that are hard to do by hand. The agency wrote back: sounds great, go for it. We ran the study, submitted it, and in substantive review the first question was whether we planned to market to residents. The thing they had approved in writing two months earlier was now the thing they were questioning. So I went to Johns Hopkins with my hat in my hand. I got a free OR and cadaver heads, and I re-ran the whole study with 16 attending surgeons in two weeks. One was a last-minute fill-in, an otolaryngologist who does balloons and tonsils and will never do microsurgery freehand. She nailed a microsurgery skills test she had failed as a resident. That one data point was the strongest evidence we had that the device did what we claimed. But notice what no advisor could have done for me. The study population was our claim to own, and a pre-sub yes did not bind the team that reviewed the file months later. ## The part you still own Here is the distinction that matters whenever you put a regulatory heavyweight on your board or your cap table. The framing of the product is product strategy, and it belongs to you, not to the regulatory team. Galen could have been defined two ways: as a remote-control telemanipulator competing with da Vinci, or as a stabilizer that improves a surgeon’s access at the end of a 40-centimeter instrument. Same hardware. The first framing invites a clinical-outcomes trial and a high evidence bar. The second invites an engineering and usability bar that is cheaper, faster, and far easier to control. And that one decision sets everything downstream: the indication language, the study population, the predicate logic, the marketing claims you can make after clearance. Every one of those has to stay consistent with the framing, or the file falls apart. An advisor like Shuren can tell you how a reviewer will read your framing. He cannot pick it for you. If a consultant is picking it, you have the wrong person in the product seat. Dave’s take The founders who get the most from a regulator-turned-advisor are the ones who walk in already owning their framing and use the access to test how the agency will read it. The ones who get the least are hoping to hand the whole problem to a famous résumé. Access reads the signal. It does not write your product definition, and the product definition is the document that decides whether the rest of the file holds together. From Dave’s video library Dave on the test for which work you own yourself and which you hand off. A former FDA device chief can read the agency for you, but the product definition is work a founder cannot rent out. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [How a 2018 De Novo Let Dexcom Put a Glucose Monitor on a Toddler →](/field-notes/2026-06-17-dexcom-otc-cgm-kids-de-novo-dividend) [FDA 510(k) Exemptions Are Expanding. They Can Erase Your Moat →](/field-notes/2026-06-06-fda-510k-exemptions-predicate-moat) [The Robot Is the Giveaway, the Instruments Are the Business →](/field-notes/2026-06-14-surgical-robot-instruments-are-the-business) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # When Beating the Predicate Breaks Your 510(k): Substantial Equivalence | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-06-21-when-better-specs-break-510k-equivalence [← All Field Notes](/field-notes) June 21, 2026 Regulatory A regulatory professional posted a question on a forum that hard-tech founders get backwards all the time. Their device puts out a higher voltage than the predicate it wants to clear against. Same intended use, similar build everywhere else. Would the FDA still call it substantially equivalent? The instinct behind the question is that a stronger spec helps the case. Under the way a 510(k) actually gets read, the number that beats the predicate is often the number that knocks you out of the lane. ## What substantial equivalence actually tests A 510(k) clears your device by showing it is substantially equivalent to one already on the market. The law that governs the program, section 513 of the Food, Drug, and Cosmetic Act, sets the bar in two parts. Same intended use, and either the same technological characteristics or different ones that do not raise different questions of safety and effectiveness. The phrase doing the work is “different questions.” The test is not whether your device is as good as the predicate or better. It is whether the difference introduces a safety or effectiveness question the predicate’s data never had to answer. Higher output voltage is a change in how the device delivers energy, and energy source is one of the features the statute names. So the question a reviewer asks is not whether more voltage is better. It is whether more voltage creates a risk, such as thermal injury or unintended stimulation, that the cleared predicate was never evaluated against. If it does, you owe performance data to close that question, and the data may have to be clinical. ## Where the airtight predicate argument falls apart I have watched a predicate argument that looked airtight collapse on exactly this seam. At Galen we built a cooperative-control robot for skull-base surgery, and we were sure it was a slam-dunk 510(k) against the da Vinci. The case was stronger than category overlap. We shared software lineage with da Vinci, and Russ Taylor had written the original code. From where we sat, the predicate logic was clean. The FDA did not see it our way. They told us the indication did not match, that ours was a different kind of system, and that it would be advisable to consider another path. They never ordered the De Novo. In agency language, “advisable to consider” is the answer, so we took the hint. The lesson sits right under the voltage question. Predicate logic that looks binding to you is not binding on the FDA. What ends a 510(k) is not a worse device. It is a difference the agency decides raises a new question, and your shared code or your stronger number does not settle that for them. ## The two honest roads, decided early Here is the choice when your differentiator is the difference that breaks equivalence. You have two honest roads, and the work is picking between them before you write a line of the submission. Keep the 510(k) and generate the data that closes the new question, paying the time and cost to do it. Or stop forcing the predicate and run a De Novo. The same law that defines equivalence lets a device with no fair predicate, or one found not substantially equivalent, ask the FDA to classify it as low or moderate risk on its own terms. Founders hear De Novo as a demotion. It is closer to the opposite. You write the safety story the agency adopts, and when it is done you become the predicate that everyone after you has to clear against. This applies to any hard-tech founder whose pitch leans on a number that beats what is already cleared. Higher power, faster cycle, finer resolution. The improvement you sell to investors is the same improvement a reviewer reads as a new question. The earlier you decide which road that number puts you on, the cheaper the whole file gets. Dave’s take Better than the predicate and substantially equivalent to the predicate pull against each other, and plenty of founders do not feel the tension until a reviewer names it. Pick the claim you are making before you pick a predicate, not after. The number that wins your demo can be the number that costs you a year, and the founders who clear fastest are the ones who saw that coming and built the file around the right road from the start. From Dave’s video library Dave on why better working beats better looking every time a buyer decides. Substantial equivalence enforces the same rule: the spec that wins your demo can be the spec that breaks your 510(k). I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [FDA 510(k) Exemptions Are Expanding. They Can Erase Your Moat →](/field-notes/2026-06-06-fda-510k-exemptions-predicate-moat) [How a 2018 De Novo Let Dexcom Put a Glucose Monitor on a Toddler →](/field-notes/2026-06-17-dexcom-otc-cgm-kids-de-novo-dividend) [The FDA’s Former Device Chief Is Advising Startups. What That Buys You →](/field-notes/2026-06-20-fda-device-chief-advising-medtech-startups) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Medtronic's Pulse Oximeter Fix Began in the Cohort | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-06-22-medtronic-pulse-oximeter-validation-cohort [← All Field Notes](/field-notes) June 22, 2026 Medical Devices A nurse reads 95 percent off the monitor while the patient in the bed is quietly losing oxygen. The pulse oximeter clipped to the finger is doing exactly what it was built to do. The catch is what it was built against. For years the reference data behind these sensors leaned on lighter-skinned volunteers, so the device learned to read one kind of body well and other bodies less well. This month Medtronic cleared a version meant to narrow that gap. It turns on a decision most hardware founders never write down: the population you validate on is a product call, and the product owner owns it. ## What Medtronic actually cleared Medtronic disclosed on June 9 that the FDA cleared its Nellcor pulse oximetry system with a new component it calls the Nell-EQ processor, through the 510(k) pathway. The processor uses patient- and sensor-specific signal characteristics to turn the raw optical reading into a blood-oxygen number more consistently across different patients, skin tones, and care settings. This was not a rush job. The agency had already granted the device its Safer Technologies Program designation in 2025, and in January Medtronic published a verification study it says covered a full range of skin tones, reporting accuracy that beat the threshold in the FDA’s draft guidance for pulse oximeters. ## The gap lived upstream None of this was a physics surprise. A pulse oximeter infers oxygen from how light passes through tissue, and skin pigment changes that absorption. The FDA wrote its draft guidance after years of concern that the readings run high more often in patients with darker skin, which can hide a drop steep enough that a clinician would otherwise act on it. That failure mode was knowable from the requirement, not discovered in the field. When your validation cohort skews toward one population, your real intended-use claim quietly narrows, even while the label still says patients. Deciding who the device is for, and who you prove it on, is product strategy, the same kind of call as how you frame the indication itself. I tell founders not to hand that one to the regulatory team. The regulatory team executes against the framing. Choosing the body the product has to hold up for belongs to whoever owns the product, and getting it wrong does not show up as a regulatory problem until much later, when it is also a clinical one. ## What skipping it costs For a founder, the validation cohort is cheap to scope at the protocol stage and brutal to fix after launch. Medtronic can absorb this. It is a large incumbent retrofitting a processor and re-running studies on a product it has sold for decades. A seed-stage company measuring something on a human body does not get that runway. If your device reads a signal off a patient and makes a call, whether it is a wearable, a diagnostic, an imaging system, a monitor, or software that counts as a medical device, the body you test on is part of the spec, not a recruiting detail you settle later. The commercial side is moving the same way. Danaher agreed in February to buy Masimo, the other major name in this market, for $9.9 billion. Accuracy across every patient is turning into a place companies compete, not only a box they check for the FDA. Dave’s take The requirement was always that the thing works on patients, and patients were never a single body. What reads in the press as an equity story is, from the product seat, a plain validation miss nobody priced in early. Scope the cohort before you write the protocol. It is the cheapest version of this fix you will ever get, and almost nobody reaches for it until a reviewer, a journal, or a lawsuit reaches for it first. From Dave’s video library Dave on deciding what to measure before you measure it. Medtronic's pulse oximeter fix started in the cohort, because the cheapest version of that correction was always the scoping decision at the start. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [How a 2018 De Novo Let Dexcom Put a Glucose Monitor on a Toddler →](/field-notes/2026-06-17-dexcom-otc-cgm-kids-de-novo-dividend) [Closed-Loop Insulin Delivery: The Algorithm Is the Easy Part →](/field-notes/2026-06-05-closed-loop-insulin-delivery-validation) [When a Quality System Catches the Problem and Still Can’t Fix It →](/field-notes/2026-06-09-quality-system-capa-effectiveness) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Karl Storz Is Retiring a Surgical Robot That Worked | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-06-23-karl-storz-retires-senhance-surgical-robot [← All Field Notes](/field-notes) June 23, 2026 Surgical Robotics Karl Storz made a quiet decision this month that every founder building a better surgical robot should sit with. The company is keeping Asensus Surgical’s software, its clinical data, and its patents, folding all of it into its own R&D. And it is throwing the robot itself away. The Senhance system is being phased out, the Asensus brand retired, and 108 people in Morrisville, North Carolina, about 30 of them engineers, are losing their jobs. The hardware was not broken. That is the part worth slowing down for. ## What Karl Storz is actually retiring Karl Storz filed a layoff notice with North Carolina on June 16, according to MedTech Dive. The cuts begin August 15 and run over six months, landing on 108 of the more than 200 people Asensus employs across a dozen countries, including roughly 30 engineers and a layer of senior managers. The Senhance robot is being phased out, and development on the newer Luna platform is stopping. What survives is not the machine. Karl Storz is pulling Asensus’s software engineering, clinical data, and intellectual property into its global R&D group and running any future robotics under its own name. It bought Asensus back in August 2024. Most coverage frames this as an acquirer trimming a redundant product line, which is fair enough. It also skips the more useful question: why a working laparoscopic robot was worth more to Karl Storz dead than alive. ## What actually kills a surgical robot Surgeons and investors love to argue about whether a given robot gets better outcomes than a skilled surgeon operating by hand. In my experience that is the wrong argument. The da Vinci did not win because it beat the best laparoscopic surgeons. It won because it grew the number of surgeons who could perform a hard minimally invasive procedure at all, from a few dozen in the world to tens of thousands. More capable surgeons means more patients offered something other than open surgery. That is the win, and it has little to do with whether any one surgeon is sharper on the console. Senhance never moved that number the way it had to. It put a second laparoscopic console in a room where the incumbent had already trained a generation of surgeons and locked in their muscle memory. Asking a surgeon to relearn a workflow is one of the hardest sells in medicine, and you only win it by handing back something big: a procedure they could not do before, or real friction taken out of one they already do. A comparable robot that asks for the same relearning without that payoff is not a cheaper da Vinci. It is a reason to stay with the da Vinci. ## What this means if you are building hard tech This is not only a surgical-robotics lesson, though it lands hardest there. Any founder putting a better version of an incumbent’s machine in front of a trained operator is making the same bet, whether that operator is a surgeon, a sonographer, a lab tech, or a line worker. The buyer is not measuring your device against the textbook ideal. They are measuring it against the workflow they already trust, and your improvement has to clear the cost of changing. Look at what Karl Storz actually kept: the engineers, the data, the patents. What it valued was the accumulated know-how, not the chassis around it. If your pitch is mostly that your hardware is better, you are competing on the part that turned out to be the easiest to walk away from. Dave’s take A surgical robot lives or dies on one number: how many patients end up getting the better procedure because of it. Not whether it outperforms a great surgeon, which is the fight everyone wants to have. Senhance never moved that number enough to matter, and no amount of engineering polish was going to change that. If you are building hardware against an entrenched incumbent, answer one question before you raise another dollar. Who can do something with your machine that they genuinely could not do before? If you cannot name them, you do not have a wedge. You have a second option nobody asked for. From Dave’s video library Dave on the math that says when it is actually safe to quit, which is a number, not a feeling. Karl Storz just ran that calculation on a robot that worked fine and retired it anyway, because working and worth continuing are different tests. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [The Robot Is the Giveaway, the Instruments Are the Business →](/field-notes/2026-06-14-surgical-robot-instruments-are-the-business) [NEURA’s $1.4 Billion Humanoid Bet, and Why I’d Bet Narrow →](/field-notes/2026-06-13-neura-1-4b-physical-ai-humanoid-form-factor) [Sanctuary AI’s Physical AI Claim Is One You Can Actually Check →](/field-notes/2026-06-19-sanctuary-ai-validated-number-not-demo) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # What FDA's QMSR Transition to ISO 13485 Means for Founders | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-06-24-qmsr-iso-13485-quality-system [← All Field Notes](/field-notes) June 24, 2026 Regulatory Most hardware founders found out the rulebook changed from a consultant’s invoice. As of February 2, 2026, the quality system every FDA device maker answers to is governed by a different regulation than it was a year ago, and the usual reaction is to hand the whole transition to a brand-name firm and pay whatever it quotes. The change is real and worth understanding. Paying someone else to understand it for you is the expensive mistake. ## What changed on February 2 The FDA retired the Quality System Regulation, the Part 820 framework that governed device manufacturing for nearly thirty years, and replaced it with the Quality Management System Regulation. The QMSR keeps the 21 CFR Part 820 citation but rewrites what sits underneath it. It folds in ISO 13485:2016, the international quality standard, by reference, so the structure US device makers run on is now the one the rest of the world was already using. The agency kept its own provisions on top. Where an FDA definition conflicts with an ISO one, the FDA definition wins, and the US requirements for device registration, listing, and distribution records stay in force. It applies to every finished device manufacturer marketing in the United States, foreign firms included. The enforcement side moved too. The FDA dropped the inspection technique it had used for years and adopted a new compliance program, numbered 7382.850, retiring the two older inspection programs that came before it. An investigator walking your floor this year reads your quality system against ISO 13485, not the old checklist. ## Why founders overpay to get ready This is where founders bleed cash. A quality system is not exotic. ISO 13485 is a published standard you can read, and mapping your processes to it is work your own team understands better than any outside firm, because they are the ones running those processes. The costly pattern is paying a specialist to build the whole thing end to end, when the leverage is in using them to tell you what to figure out and to refine what you produce. I learned the price of the other approach on Galen’s De Novo. Our bill with a brand-name law firm came to close to half a million dollars. Our own people wrote the documentation. The regulatory work product was produced inside the company. What the board wanted was the firm’s signature on the file, because it changes how the next investor reads it. That signature had real value. It was also a separate value from the thinking, and we paid premium rates for both when we only needed to buy one. A former FDA reviewer later put it to me plainly: that firm was the right lawyer and the wrong regulatory consultant. The same split applies to a quality system. Buy the polish and the sign-off if your board demands it. Do not buy someone else’s understanding of how your own company makes things. ## What the transition actually asks of you The QMSR is not a new discipline to learn. If you already built to ISO 13485, which most teams with any ambition outside the US did, the gap is small and the work is documentation and definitions. If you built only to the old regulation, the structure is familiar but the emphasis shifts, and the honest move is to read the standard yourself before anyone quotes you to interpret it. A quality system has also stopped being a back-office artifact. Investigators read it differently now, and so do the diligence teams on the far side of a financing or an acquisition. A device maker who can walk an inspector or an investor through its own system without a consultant in the room is signaling something a binder never will. This is device law, so it lands hardest on the MedTech and diagnostics founders who are my deepest vertical and the people this regulation actually governs. The wider lesson still travels. For any founder building regulated or safety-critical hardware, the quality and design-control system is becoming the first thing a serious buyer or regulator inspects, and the founder who owns it rather than rents it holds an edge that grows over time. Dave’s take The founders who handle this well are not the ones who spend the most on it. They are the ones who read the standard, mapped it to how they already work, and brought in a specialist only to pressure-test the result and put a name on it. A rule change is a poor reason to outsource your understanding of your own company, and a good reason to finally learn the part of it you have been avoiding. From Dave’s video library Dave on the value-of-your-time test for what to own and what to hand off. QMSR rewards the founders who ran it honestly: learn your own quality system, and rent a specialist only to pressure-test the result. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [When a Quality System Catches the Problem and Still Can’t Fix It →](/field-notes/2026-06-09-quality-system-capa-effectiveness) [The FDA’s Former Device Chief Is Advising Startups. What That Buys You →](/field-notes/2026-06-20-fda-device-chief-advising-medtech-startups) [When Beating the Predicate Breaks Your 510(k) →](/field-notes/2026-06-21-when-better-specs-break-510k-equivalence) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Why a 3-Hour-Faster Blood Culture Test Is a Real Product | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-06-25-faster-blood-culture-is-a-real-product [← All Field Notes](/field-notes) June 25, 2026 Diagnostics Founders are trained to chase the tenfold improvement. The device that crossed the FDA finish line this month did something less glamorous and more instructive: it got about fifteen percent faster at a job people are already desperate to do better. On June 3 the FDA cleared the BD BACTEC FXI Culture System, a blood-culture platform that detects bloodstream infections roughly three hours sooner than the generation before it. Three hours sounds like a rounding error until you remember what the clock is measuring. ## An incremental number sitting on an urgent problem The headline figures are modest. Mean time to detection drops from about twenty hours to seventeen, the fifteen percent the announcement leads with. The system loads up to sixty vials at once, a fifty percent jump in capacity, and ships in 480- and 960-vial configurations. BD builds it and Waters is bringing it to the US market. It already cleared in Europe in April and is licensed in Japan. None of that reads like a breakthrough. The pain underneath it is not modest at all. Sepsis is the textbook case of a problem where every hour counts, and a blood culture is the test standing between a clinician guessing at a broad-spectrum antibiotic and knowing what they are actually treating. Speed there is not a vanity metric. It moves the decision a doctor makes next. A small improvement on a genuinely painful problem is worth more than a spectacular improvement on a problem nobody is losing sleep over. ## The mistake I have watched founders make I have watched a lot of brilliant teams fall in love with their solution instead of someone’s pain, and I have done it myself. The cautionary tale I keep in my pocket is the Magic Link from 1994: real ex-Apple engineers, email and messaging from anywhere thirteen years before the iPhone, a data deal with AT&T, manufacturing by Sony. Technically gorgeous. Nobody wanted it. The vision was there and it was still a fancy science project, because it answered a question users were not yet asking. The opposite of that is the discovery method I learned in surgical robotics. You watch the procedure and ask what the hardest part is. When the surgeon describes the awkward workaround they have just learned to tolerate, that is your opening, because every surgeon trained the same way is tolerating the same thing. Tolerated workarounds are concentrated pain. The blood-culture lab has its own version: the wait, and the antibiotic chosen blind during it. BD did not win here by inventing a category. They shaved hours off a wait that hospitals were already counting and added capacity for labs already running flat out. ## What this means for diagnostics founders For IVD and diagnostics founders specifically, the lesson is to anchor every roadmap decision to a clinical decision that changes. A faster result that does not change what the clinician does next is a feature. A faster result that lets them stop guessing is a product. Notice the split, too. BD makes the system and Waters carries it to market. Getting cleared and getting distributed are two different problems, and the second one decides whether anyone actually uses what you built. The read travels past diagnostics. Any hard-tech founder choosing between a bold new architecture and a boring improvement to something people already depend on should weigh the boring option more seriously than instinct allows. Dave’s take The flashy demo and the urgent, unglamorous fix compete for the same engineering hours, and founders almost always overweight the demo. I would rather ship a fifteen percent improvement a clinician feels on a Tuesday than a tenfold improvement to a problem no one was bleeding over. Pain is the signal. Find the workaround people have stopped complaining about because they have given up expecting better, and build there. From Dave’s video library Dave on the questions that find the pain customers stopped mentioning. BD's three-hour-faster blood culture wins because someone noticed the workaround clinicians had given up complaining about. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Why OraSure Built a Collection Kit, Not a Better STI Test →](/field-notes/2026-06-16-orasure-sti-collection-kit-access-bottleneck) [How a 2018 De Novo Let Dexcom Put a Glucose Monitor on a Toddler →](/field-notes/2026-06-17-dexcom-otc-cgm-kids-de-novo-dividend) [The Robot Is the Giveaway, the Instruments Are the Business →](/field-notes/2026-06-14-surgical-robot-instruments-are-the-business) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Energy Dome's CO2 Battery Wins Where Lithium Can't | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-06-26-google-energy-dome-co2-battery-long-duration [← All Field Notes](/field-notes) June 26, 2026 Climate Hardware Somewhere inside Google’s energy procurement group, someone signed a ten-year contract for a technology that did not exist at commercial scale eighteen months ago. On June 23 the company announced its first bilateral deal with Energy Dome: a 23-megawatt, 200-megawatt-hour CO2 battery in County Offaly, Ireland, due online in 2028 and backed by a ten-year capacity contract with the grid operator EirGrid. The easy way to read that is corporate climate theater. The more useful way is to ask why a buyer with the pick of every storage technology on earth would underwrite a category that lithium does not occupy. ## A buyer underwrote it, not a lab The Ireland deal is not a one-off. Earlier in June, Energy Dome, Google, and the Arizona utility SRP announced a 19-megawatt, 200-megawatt-hour project, and Energy Dome’s first full-scale plant has been running in Ottana, Sardinia since July 2025 at 20 megawatts and 200 megawatt-hours, good for about ten hours of storage. The technology itself is almost dull to describe. You compress carbon dioxide into a liquid when power is cheap, then let it expand back through a turbine when the grid needs the power returned. What is not dull is who is paying for it. A hyperscaler signing a decade of offtake is a different kind of signal than a pilot grant or a Series C. A pilot proves a machine runs. A ten-year contract proves a buyer will bet its own balance sheet on the machine still running in 2038. That second kind of proof is the one founders actually need, and the one that is hardest to manufacture. ## Energy Dome did not build a cheaper lithium cell Lithium is very good at short bursts, the one-to-four-hour band that smooths a cloudy afternoon or an evening demand peak. Its economics get worse the longer you stretch the duration, because every extra hour means more cells. Energy Dome did not pick that fight. It went after the eight-to-twenty-four-hour band where lithium gets expensive, with a machine that runs roughly thirty percent cheaper than lithium for that job, lasts about three times as long, fits on five hectares, and gets built in under two years. This is the trap I warn surgical-robot founders about, and it moves over to grid hardware without losing much. The losing move is to pitch a new robot as a smaller, cheaper version of the market leader. The moment you line up against the incumbent on its own terms you lose, on installed base, on budget, and on the buyer’s mental model of what the thing even is. The winning move is to find the work the leader serves poorly and own that ground instead. Energy Dome is not a cheaper lithium battery. It is the thing lithium cannot affordably be. ## What this means for climate-hardware founders If you are building long-duration storage, grid hardware, or any capital-heavy machine a utility or a hyperscaler has to underwrite, your spec sheet is not the product. The economic case is. Nobody writes a ten-year contract because your round-trip efficiency is elegant. They write it because the duration you serve is one they cannot buy cheaper anywhere else, and because the payback math holds over the life of the deal. Build the buyer’s economic story before you optimize the engineering, and aim at the duration or the geography where the incumbent’s economics fall apart, not the spot where they are strongest. The category you can defend is the one the leader does not want. Dave’s take Energy Dome won by refusing to be measured against lithium, and by serving a duration lithium cannot reach at a price that works. That is a positioning decision before it is an engineering one. When a buyer signs ten years of offtake, they are not paying for your cleverness. They are paying for a number that closes, and they are telling you exactly which fight was worth picking. From Dave’s video library Dave on pricing from the value delivered, not the cost to build. Energy Dome's ten-year Google contract is that principle signed in ink: the buyer is paying for a number that closes, in a duration lithium cannot serve. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [GM’s Sodium-Ion Bet: Why It’s Selling Cells, Not Systems →](/field-notes/2026-06-10-gm-sodium-ion-grid-storage-product-boundary) [The 2026 Hardware Funding Records Hide a Higher Bar, Not Easier Money →](/field-notes/2026-06-08-hardware-funding-records-higher-bar) [Dual-Use Hardware: When Battlefield Drone Tech Comes Home →](/field-notes/2026-06-11-dual-use-counter-drone-comes-home) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Tissium's $68M Raise and the Power of a Narrow De Novo | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-06-27-tissium-narrow-de-novo-nerve-repair [← All Field Notes](/field-notes) June 27, 2026 Medical Devices Somewhere in Tissium’s regulatory file is a sentence most founders would read as a limitation. The company’s first commercial product, Coaptium Connect, is cleared by the FDA to do one thing: repair a peripheral nerve, without sutures, when the gap is a centimeter or less. Not nerves in general. Not the whole platform of light-activated polymers that might one day seal a hernia or a blood vessel. One injury, one centimeter. On June 25, MedTech Dive reported that Tissium had pulled together 60 million euros, about 68 million dollars, to carry that narrow product into the U.S. operating room. The narrow scope is not the weakness in this story. It is why the money showed up. ## The indication is the product Coaptium Connect pairs a bioresorbable, light-activated surgical polymer with a 3D-printed chamber, applied with a device Tissium calls its Light. The FDA granted it a De Novo classification, the pathway for a device novel enough that no predicate exists to match it against. The broader platform reaches toward hernia repair, in the clinic under the name Eclipsium, and cardiovascular sealing, still in the pipeline, and the company has raised roughly 200 million euros over its life. But the FDA did not clear a platform. It cleared one indication with a boundary you can measure with a ruler. That precision is the regulatory strategy, not a footnote to it. ## What a De Novo actually buys you At Galen, we did not choose the De Novo. The FDA routed us to it, and most founders miss that signal when it lands. We had turned in a pre-submission, which a lot of people misread as a chance to ask the agency a list of questions. It is not. It is a strategy presentation. You state your plan, and they nod or they do not. The agency came back and told us our predicates were not appropriate for a 510(k) with the device as we had described it, and that we might be better suited to a De Novo. We took the hint. What surprised me is that the De Novo is an advantage. The FDA suggesting it means the agency already sees you as moderate risk, because a genuinely high-risk device would be in a PMA by then. You also get to define the safety story yourself, instead of inferring a competitor’s test plan telepathically from a clearance you are not allowed to read. The best part came at the end. When you win a De Novo, the FDA has to write a new regulation for your device, because none exists, and they asked us what it should say. We sent a paragraph. They asked for it in Word. A day or two later the regulation describing our device read word for word what we had written. We became the predicate. Tissium just did the same thing for sutureless nerve repair, and the next company that wants to do this has to prove it is substantially equivalent to Tissium. ## The capital followed the cleared product, not the platform Look at how the money is built. About 30 million euros is a Series D2 that closed at the end of 2025, and up to another 30 million comes from the European Investment Bank in three 10-million tranches, each released against commercial, clinical, and financing milestones. The institutional half of the raise is not a bet on the platform vision. It is metered against the commercialization of one cleared product. The investors are underwriting what the FDA has already de-risked. If you are a founder holding a platform technology, a biomaterial or a novel mechanism that could plausibly do ten things, that structure is worth studying. Pick the one indication you can define tightly enough to clear, clear it, and let the rest of the platform compound off a product that exists rather than a deck that promises five. Dave’s take A platform is a story you tell investors. A cleared indication is a fact you can build on. Tissium has both, and it raised on the fact. If you are holding a technology that could do ten things, the operator move is to find the one you can define down to a centimeter, get the FDA to write your regulation, and become the thing everyone after you has to match. From Dave’s video library Dave on real value versus the story told to raise on. Tissium has both a platform story and a cleared indication, and it raised on the fact, not the story. That is the ordering worth copying. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [How a 2018 De Novo Let Dexcom Put a Glucose Monitor on a Toddler →](/field-notes/2026-06-17-dexcom-otc-cgm-kids-de-novo-dividend) [When Beating the Predicate Breaks Your 510(k): Substantial Equivalence →](/field-notes/2026-06-21-when-better-specs-break-510k-equivalence) [GT Medical’s $100M GammaTile Raise: Evidence, Not Clearance →](/field-notes/2026-06-15-gt-medical-gammatile-evidence-unlocks-capital) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Why the FDA's 2,200 New Hires Won't Speed Up Your Review | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-06-28-fda-rehiring-wont-speed-your-review [← All Field Notes](/field-notes) June 28, 2026 Regulatory A founder messaged me the morning the news broke, relieved. The FDA is staffing back up, he figured, so the backlog will clear and his submission will move faster. I get the instinct. I also think it is the wrong thing to bank on. At the BIO convention in late June, the agency said it is hiring around 2,200 people to refill the ranks it cut a year ago. That is good news for the FDA. It is not a number you get to put in your investor deck. ## What the FDA actually said At a town hall during the BIO International Convention on June 24, FDA officials walked through the rebuild. The agency is hiring roughly 2,200 staff after losing more than 3,000 to last year’s DOGE cuts. About 600 are onboarding now, a couple hundred have already started, and recent orientations have been bringing in several dozen people per pay period. Attrition at the drug center is back down to its historical norm. BioSpace reported the details from the convention floor. The honest way to read it is recovery in progress, not recovery finished. Notice which parts of the agency were named: the drug and biologics centers. The device founders I work with answer to a different center entirely, so “the FDA is hiring” tells you almost nothing about who will read your specific submission, or how many reviews they will have under their belt when they do. ## Why a bigger FDA does not speed up your review Here is the part founders miss. Swapping out 3,000 experienced reviewers for 2,200 new ones is not a wash, and it is not quick. Institutional memory does not show up in an orientation packet. A reviewer who is three months in and unsure of the precedent tends to be more cautious, not less. They ask for more. They lean on the letter of the standard instead of judgment. The published backlog can shrink while your particular file draws a more conservative read than it would have two years ago. So the staffing number is not a lever you control. What you do control is how strong your submission is before it reaches anyone, green or seasoned. That starts with treating the pre-submission meeting as the real event. A pre-sub is not a Q&A where you ask the FDA what to do. It is a strategy presentation. You assert your predicate, your indication, and your testing plan, and the agency signals back whether it agrees. “You may be better suited as a De Novo” is not a suggestion. It is an answer. Read it as policy and you spare yourself a rejected submission. Treat it as a hallway chat and a churning review staff will find every soft edge you left. ## Plan for the reviewer you have not met Even a clean pre-sub does not insulate you, and this is where my own scars are. On Galen’s path, the FDA reviewed our study design through a Q-sub and told us it was acceptable. Months later the substantive review team, different people with a different mandate, came after us over the fact that our data came from residents rather than attending surgeons, even though that same design had been blessed in the Q-sub. We re-ran a sixteen-subject cadaver study with attendings in two weeks, and we only pulled it off because Johns Hopkins’ neurosurgery department handed us a free operating room and cadaver heads. That kind of rescue does not exist without the relationships behind it. The lesson stuck: a yes from one reviewer does not bind the next one. That is more true in a year when the agency is cycling through new staff, not less. The founders who clear a turbulent FDA are the ones who show up with a narrow indication, a De Novo safety story they wrote themselves, and a protocol built to survive a cold read by someone who was never in the room for the friendly conversation. Dave’s take I would not build a single week of runway on the FDA hiring back up. The recovery is real and I am glad to see it, but it is the agency’s project, not your strategy. The variable you own is the clarity of what you submit, and that has paid off in every staffing climate I have worked through, the fat years and the gutted ones. Control the file, and let the agency fix the agency. From Dave’s video library Dave on separating what a government shift actually changes from what it appears to change. The FDA hiring 2,200 people back is real recovery, and still not a strategy: the variable a founder owns is the clarity of the file they submit. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [The FDA’s Former Device Chief Is Advising Startups. What That Buys You →](/field-notes/2026-06-20-fda-device-chief-advising-medtech-startups) [What FDA’s QMSR Transition to ISO 13485 Means for Founders →](/field-notes/2026-06-24-qmsr-iso-13485-quality-system) [When Beating the Predicate Breaks Your 510(k) →](/field-notes/2026-06-21-when-better-specs-break-510k-equivalence) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Aidoc's Radiology AI Got a Breakthrough Tag, Not a Clearance | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-06-29-breakthrough-designation-is-not-clearance [← All Field Notes](/field-notes) June 29, 2026 Medical Devices The most important line in Aidoc’s announcement last week was not the word breakthrough. It was a quieter sentence further down: a radiologist still has to sign every report the AI drafts. That one design choice is what lets a generative model write radiology text at all, without triggering the kind of clinical trial that would push the product out by years. Aidoc framed it correctly. Most founders chasing the same headline will fixate on the wrong word. ## What Aidoc actually got On June 25, Aidoc said the FDA had granted Breakthrough Device Designation to First Read, a tool that drafts preliminary report text for more than 100 pre-specified findings on chest X-rays, with a radiologist signing off on every one. The company’s clinical AI platform already runs in more than 2,000 hospitals, and it raised $150 million in April to push models like this forward. MedTech Dive covered the designation. Here is the part worth slowing down on. A designation is not a clearance. The breakthrough program moves a device to the front of the review queue and opens up more feedback from the agency along the way. To qualify, the device has to promise more effective diagnosis or treatment of a serious or life-threatening condition. That is a scheduling decision, not a verdict on the product. You still have to run the review and win it on the merits. Founders see breakthrough and hear validated. The FDA is only saying prioritized. ## The sentence that does the regulatory work Go back to the radiologist signing every report. That is not a usability footnote. It is the indication, and the indication is the whole game. A tool that drafts text for a radiologist to approve is a different product from one that reports findings on its own, even if the underlying model is identical. The first is an assistant inside the clinician’s existing workflow, judged on accuracy and usability with a human in the loop. The second claims to stand in for clinical judgment, which invites the expensive, outcome-driven evidence bar and the trial that comes with it. I lived the cost of getting that choice wrong. At Galen we first framed our surgical robot as a telemanipulator, substantially equivalent to da Vinci. The FDA rejected it. We reframed the identical machine as a device that improved a surgeon’s access to anatomy at the end of a long instrument, and that version went through as a De Novo. Same hardware. The framing changed everything. Aidoc made its version of that call up front, and even the count of 100 pre-specified findings is part of it: bounded and named, not an open-ended promise to read the whole scan. ## What this buys a founder building clinical AI If you are building diagnostic AI, software as a medical device, or anything adjacent to clinical autonomy, the instinct to claim the biggest capability is the instinct that lands you in the slowest, most expensive review. “Our AI reads the scan” is a sentence that invites a trial. “Our AI drafts a read your radiologist signs” is a sentence that invites an engineering and usability file. Pick the second one, then carry it consistently through the deck, the indication, the study design, and the demo. A single document that drifts to the bigger claim can undo the framing. The honest engineering read points the same way. The capability to actually replace the radiologist is not there yet, so selling autonomy is selling the Barnum version. Keep the human in the driver’s seat because the technology is not ready to take the wheel, and because that happens to be the same framing the agency will reward. Dave’s take What Aidoc earned here is product framing wearing a regulatory costume. They decided what the product is, a drafting assistant a radiologist signs, before they decided how to clear it, and the breakthrough program rewarded that discipline. When a founder asks me how to get an AI device through the FDA faster, the answer is almost never a different pathway. It is a narrower, more honest claim. Enhance the clinician, keep them in the driver’s seat, and most of the regulatory story writes itself. From Dave’s video library Dave on AI working behind the professional, where it raises the floor instead of replacing the judgment. Aidoc built its whole regulatory strategy on that same frame: a drafting assistant whose every output a radiologist signs. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Tissium’s $68M Raise and the Power of a Narrow De Novo →](/field-notes/2026-06-27-tissium-narrow-de-novo-nerve-repair) [When Beating the Predicate Breaks Your 510(k) →](/field-notes/2026-06-21-when-better-specs-break-510k-equivalence) [The FDA’s Former Device Chief Is Advising Startups. What That Buys You →](/field-notes/2026-06-20-fda-device-chief-advising-medtech-startups) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Medtronic's Hugo Filing Is an Access Play, Not a Specs Fight | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-06-30-medtronic-hugo-general-surgery-access-play [← All Field Notes](/field-notes) June 30, 2026 Surgical Robotics When Medtronic said in early June that it had filed to put its Hugo robot into general and gynecologic surgery, the reaction across the industry was instant: the company is finally coming for Intuitive’s home turf. That part is true, and it is also the least interesting thing about the filing. The argument everyone is about to have is whether Hugo is as good as da Vinci. Having helped build one of these systems myself, I think that is the wrong fight, and the way Medtronic packaged the announcement tells me the company knows it too. ## What Medtronic actually filed On June 3, Medtronic announced 510(k) submissions to clear Hugo for general surgery, including hernia repair, and for gynecologic procedures. Those two specialties are where most robotic surgery happens. By Intuitive’s own 2024 figures, da Vinci ran about 1.25 million general surgery cases and 468,000 gynecologic ones in the United States, against roughly 201,000 in urology. Hugo is cleared here only for urology so far, from December 2025, and is in commercial use at U.S. medical centers. So this is Medtronic reaching for a slice of the market that is an order of magnitude larger than where it competes today. MedTech Dive reported the filings and the volume breakdown. ## The question that actually decides this Whether Hugo wins or loses a head-to-head against da Vinci on any given case is not what settles this market. A second credible platform widens the population of surgeons and hospitals that can offer robotic surgery at all. The hard part of a complex laparoscopic procedure lives in the surgeon’s wrists, and a robot moves that difficulty into the machine. That is how robotics grew the number of surgeons who can perform a minimally invasive prostatectomy from a few dozen worldwide to tens of thousands. The honest comparison was never robot versus the small expert group who can do it by hand. It was robot versus open surgery, because open surgery is what most of those patients would otherwise have gotten. A second platform pushes that access curve further still: more installed sites, real price competition, more surgeons trained on a system, more procedures that move from open to minimally invasive. If you are building any robotic or capital surgical platform, this is the positioning lesson. Hospitals are not weighing your robot against a world-class manual surgeon they do not employ. They are weighing whether they can offer a minimally invasive option to patients who would otherwise go under the knife the old way. Position against that counterfactual, not against the category leader’s best day. ## The tell is everything filed alongside the robot Look at what shared the press release. Medtronic did not just file for two indications. It also won 510(k) clearance for its ProGrip self-gripping mesh for robotic ventral hernia repair, a procedure performed nearly 470,000 times a year in the United States, and it filed a separate 510(k) for a LigaSure energy instrument built to run on Hugo. The robot is the capital purchase a hospital makes once. The mesh, the energy device, the staplers, the sealing instruments are what get consumed on every case and reordered for as long as the system runs. Medtronic has sold surgical consumables for decades. Filing the indication and the disposable stack in the same breath is a company telling you the robot is the entry point and the workflow around it is the business. Founders selling expensive boxes outside the operating room face the same arithmetic. Diagnostics learned this years ago: place the analyzer cheaply, make the margin on the reagents that run through it. The instrument opens the door; the consumable pays the rent. Whether you are shipping a surgical robot, a sequencer, a tissue sealer, or an industrial cell, the number that decides the business is rarely the sticker price. It is what gets consumed, reordered, and serviced after the thing is installed, and whether you control that stream or hand it to someone else. Dave’s take When I was building a surgical robot at Galen, the questions that mattered were never about beating a competitor on a spec sheet. They were about which surgeons could realistically adopt the thing, and what each procedure would cost to run. Medtronic’s filing is built for exactly those two questions. If you are raising on the elegance of your hardware, you are pitching the part a buyer weighs last. Pitch the access you create, and the recurring revenue that follows each case you make possible. From Dave’s video library Dave on why the price conversation is really a value conversation about the buyer's economics. Medtronic filed Hugo's expansion on access and per-procedure cost, not specs, because that is the math hospitals actually buy on. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [The Robot Is the Giveaway, the Instruments Are the Business →](/field-notes/2026-06-14-surgical-robot-instruments-are-the-business) [Karl Storz Is Retiring a Surgical Robot That Worked →](/field-notes/2026-06-23-karl-storz-retires-senhance-surgical-robot) [When Beating the Predicate Breaks Your 510(k) →](/field-notes/2026-06-21-when-better-specs-break-510k-equivalence) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Physical AI's Hard Problem Isn't Vision, It's State Recovery | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-07-01-physical-ai-state-recovery-not-vision [← All Field Notes](/field-notes) July 1, 2026 Industrial Robotics Dr. Behrooz Rezvani has built and sold two semiconductor companies, and last week he published a warning aimed at everyone rushing to put intelligence into robots. The consensus in physical AI is that scale solves the hard problems: bigger world models, more data, more convincing demos. Rezvani argues the part that actually breaks deployments is the one nobody funds, the robot knowing, moment to moment, where it really is. I spent years building a robot that had to be right about that inside a human body, and I think he is naming the most underpriced problem in physical AI today. ## The layer that doesn’t demo Writing in The Robot Report on June 23, Rezvani, the founder and CEO of the physical-AI startup Atomathic, splits the robotics stack into four layers he calls physical AI 2.0: world models, physical state recovery, reasoning, and action. Vision and language live in the layers that give good demos. State recovery is the unglamorous middle. A camera gets blinded by glare, an object hides in shadow, two sensors disagree, and the machine’s estimate of its own situation quietly goes wrong. A machine that starts from a wrong idea of where it is cannot reason its way back to reality. His fix is not more data. It is a dedicated recovery layer built on physics-based constraints and richer sensing, so the system corrects its own state before the clever reasoning ever runs. Atomathic is backed by RTX Ventures and GM Ventures, which tells you defense and automotive already treat this as a real gap, not an academic one. ## Why the boring layer is where the companies get built This is the pattern I keep seeing in robotics, and it is why I bet narrow. The parts that photograph well, a humanoid folding laundry or a world model generating a synthetic kitchen, get the funding rounds and the keynote slots. The parts that decide whether the machine survives contact with a messy real environment do not. NVIDIA’s Cosmos platform can generate endless simulated worlds, and that is genuinely useful, but no volume of simulation fixes a sensor that cannot tell you the truth about the last inch of a grasp. The robots that already win bounded physical tasks earned it the hard way. Surgical robots run computer vision and machine learning on edge processors, on the device, because a robot in an operating room cannot wait on a data center or a clean camera frame. Da Vinci did not beat human hands by being more human-shaped. It won by being engineered around the specific task of controlling more instruments than a surgeon has hands. ## What this means if you are building one If you are a robotics or physical-AI founder, the useful question is not how general your platform sounds in the deck. It is what your machine can still do when the network is gone, the lighting is bad, and the object is not where the model expected it. Defense buyers already write that into their RFIs, asking for untethered robots that run without a connection, because a robot that needs the internet to function is a liability the moment someone cuts the link. That constraint also picks your form factor: a quadruped or a drone is far easier to make autonomous on edge compute than a general-purpose humanoid. Warehouse and industrial-automation founders face a softer version of the same test every shift. Borrow Rezvani’s four layers as a checklist. If your roadmap spends everything on perception and reasoning and nothing on knowing where you are and recovering when you are wrong, you have built a great demo and a fragile product. Dave’s take Physical AI is not overhyped because the tools are fake. The tools are useful. It is overhyped because the money follows the layer that looks impressive on stage instead of the layer that fails silently in the field. Pick the bounded task your robot is actually for, engineer the form factor and the sensing around that task, and prove the machine can recover when it is wrong. That is a slower story to tell an investor than one robot that does everything, and it is the one that ships. From Dave’s video library Dave's decision method for what to hand to AI: bounded tasks, with a human holding the context. Physical AI is relearning it in the field, where the unglamorous state-recovery layer decides whether a robot deploys at all. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [NEURA’s $1.4 Billion Humanoid Bet, and Why I’d Bet Narrow →](/field-notes/2026-06-13-neura-1-4b-physical-ai-humanoid-form-factor) [Sanctuary AI’s Physical AI Claim Is One You Can Actually Check →](/field-notes/2026-06-19-sanctuary-ai-validated-number-not-demo) [How to Read the Pentagon’s Drone-Budget Surge as a Hardware Founder →](/field-notes/2026-06-18-how-to-read-pentagon-drone-budget) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # A Cleared Device Is Not a Business: Zimmer Buys Pacira's iovera | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-07-02-cleared-device-is-not-a-business [← All Field Notes](/field-notes) July 2, 2026 Medical Devices Pacira BioSciences spent nearly a decade owning a pain-relief device that already carried FDA clearance. This week it agreed to sell that device to Zimmer Biomet for up to $140 million, and the terms give the game away: the money is not paying for the technology. The iovera system has worked, and been legal to sell, since 2017. What Zimmer is buying is the one thing a clearance certificate never hands you, which is a commercial engine that actually fits the product. I have watched a lot of founders mistake a regulatory milestone for a business, and that gap is the whole lesson sitting inside this deal. ## What Zimmer actually bought On June 30, Pacira announced it would divest its iovera business to Zimmer Biomet, the orthopedics-device maker, in a deal worth up to $140 million: $70 million up front and another $70 million tied to milestones running through the end of 2031, per the company’s release. Fierce Biotech reported the terms. The transaction is expected to close by the end of September. iovera is a handheld device that delivers controlled doses of cold to a targeted nerve and briefly interrupts the transmission of pain signals, a drug-free option for procedures where a clinician might otherwise reach for an opioid or a nerve block. It was developed by a company called Myoscience, cleared by the FDA in 2017, and picked up by Pacira when it acquired Myoscience in 2019. ## A cleared device is not a business So why does a nine-year-old cleared product change hands for that kind of money now? Read Pacira’s own explanation. Its chief executive said the sale advances the company’s shift toward being a biopharmaceutical business, and that Zimmer’s established expertise commercializing medical devices, plus its reach, could unlock the full potential of iovera. Put back into plain terms: iovera sat inside a company built to sell injectable drugs like Exparel and Zilretta through a pharmaceutical sales motion, and a handheld capital device sold into orthopedic and surgical suites is a different animal. The clearance was never the constraint. The commercial fit was. I have lived the version of this where the technology was not just cleared but dominant, and still was not a business. At Ascend Communications in the 1990s I came in as a software person to a hardware company, and inside our access concentrators sat thirty million lines of code that were quietly running roughly ninety percent of the world’s dial-up internet. I did not write a line of it. The engineering teams did. What that code lacked was a name and a commercial identity, so it lived as a feature nobody could point at until we branded it as an operating system, called it TAOS, and sold it as one. The engineering was already excellent, and the thing that moved the needle was the decision to treat it as a product with its own identity. iovera is that lesson in a different industry: a real device with a real clearance that landed in the wrong commercial home. ## What this means if you are about to clear a device If you are a hard-tech founder anywhere near a De Novo, a 510(k), or a CE mark, this deal is worth sitting with. Clearance answers one question honestly, which is whether it is safe and legal to put your device in front of customers. It says nothing about whether your company has the sales motion, the distribution channel, and the reorder relationships to actually move it. Those are separate builds, and the second one is usually harder and slower than the regulatory one. A cleared device in a company whose commercial instincts do not match it can sit for years, which is roughly what happened here before Zimmer arrived with a channel that fits. When you map your plan, do not let the clearance date quietly become your go-to-market date. They answer different questions, and only one of them pays the bills. Dave’s take The trap I watch founders fall into is treating a regulatory approval as the finish line, when it is closer to a permit than a business. Pacira held a working, cleared device for years and is now taking seventy million dollars up front because someone else owns the channel that fits it. Build the commercial engine with the same seriousness you bring to the submission, or you may end up selling the fruit of your work to whoever did. From Dave’s video library Dave on why building it is half the job and being found is the other half. Pacira held a cleared, working device for years, and Zimmer is buying it because Zimmer owns the channel that reaches its customers. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Aidoc’s Radiology AI Got a Breakthrough Tag, Not a Clearance →](/field-notes/2026-06-29-breakthrough-designation-is-not-clearance) [Medtronic’s Hugo Filing Is an Access Play, Not a Specs Fight →](/field-notes/2026-06-30-medtronic-hugo-general-surgery-access-play) [Why a 3-Hour-Faster Blood Culture Test Is a Real Product →](/field-notes/2026-06-25-faster-blood-culture-is-a-real-product) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Quantum Systems' $1.2B Round: The Number Defense Founders Should Study | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-07-03-quantum-systems-1-2b-drone-round [← All Field Notes](/field-notes) July 3, 2026 Defense Hardware When Airbus Defence and Space helped lead a $1.2 billion round this week in a company whose founder openly says he intends to disrupt the traditional prime contractors, a legacy defense giant effectively backed its own would-be replacement. Quantum Systems, the Munich drone maker, closed the largest private defense financing in European history, and the headline numbers are doing what headline numbers do. For a founder building anything in hard tech, though, the figure worth your attention is not the $1.2 billion or the $8 billion valuation. It is 19,000, the number of missions the company’s drones flew in Ukraine last year. The money is chasing that operational record, and earning a record like it is the part most founders never reach. ## What actually earned the round On July 2, Quantum Systems closed a $1.2 billion Series D at a post-money valuation of about $8 billion, co-led by Blackstone, Airbus Defence and Space, Advent International, and Noteus, with Fidelity, Wellington, and a roster of institutional names alongside. CNBC and Bloomberg both reported the figures. Eight months earlier the company was valued at roughly $3.5 billion, so this round more than doubled it. What sits underneath that jump is the part worth studying. The company’s systems flew more than 19,000 missions in Ukraine last year, it reports triple-digit revenue growth and double-digit profitability, and it manufactures across seven countries with signed procurement in Germany, the Netherlands, Romania, Australia, and the Baltic states. The profitability is self-reported and cannot be independently audited, so weigh it accordingly. But the shape is clear: this is a company that shipped a product, put it in the field, and got paid for it before it raised at this scale. That is the model the industry now calls a neo prime, the approach Anduril proved in the United States. You self-fund the R&D, build a commercial product first, get it battle-tested, and then sell it to governments already working, rather than winning a government requirement and starting the development clock from zero. For a hard-tech or defense founder, the procurement signal in that is the same one I keep pointing founders toward: the buyers are rewarding narrow, already-deployed systems, not the widest possible narrative. ## The ITAR-free choice was a product decision The center of gravity in this raise is not the drone. It is MOSAIC UXS, the company’s command-and-control software, and one design choice inside it: it is built to be ITAR-free. The US arms-export rules give Washington a veto over how allied nations use platforms that carry American-controlled technology. A control layer with no such dependency lets European governments deploy and modify it without asking permission, which is what those governments mean when they talk about technological sovereignty. That was a product and architecture decision made early, and it is a large part of the strategic value institutional capital is now paying for. One honest gap is worth naming: the platform is designed to plug into NATO’s command systems, but that integration has not been independently certified, and designed-for is not the same as proven. I have made the point for years that how you frame and architect the product is the strategy, not a step before it. Here is the version I usually reach for. I once toured a research group building a satellite robot meant to chase down satellites that had run out of fuel, couple with them, drill in, refuel, and patch the hole. I joked to the person next to me that the difference between a robot that heals a satellite and one that kills satellites is probably twelve lines of code. He nodded and said, that may be, but we don’t talk about that. The dual-use line in hard tech is thinner than people pretend. Quantum started life closer to a mapping-drone company, and the same platform that maps terrain flies reconnaissance in a war. Both roads are already sitting inside the hardware. Which one you build for, and which one you fund off of, is a decision the founder owns. ## Expensive money at an $8 billion bar There is a reason to hold the celebration at arm’s length. An $8 billion post-money valuation is not a trophy, it is the bar the company now has to clear at the next round, and it doubled in eight months. Anduril’s own chief executive spent part of late June warning that when a few companies succeed, a crowd of investors chases them into risky behavior. Early-stage defense startups were raising at 17 to 50 times revenue in the first quarter of this year, by PitchBook’s count. Quantum’s profitability and signed contracts put it in a different tier than a company chasing the same wave on a deck. As one PitchBook analyst put it, execution, not invention, will determine returns. If you are a founder without 19,000 missions of proof behind you, raising into those multiples sets a bar you then have to clear with a product that has not yet earned it. Dave’s take The headline invites you to raise like Quantum Systems. The more useful exercise is to study how they earned the right to. A narrow, deployable product, proven in the field, bought by real customers, plus one architecture decision that made the platform genuinely theirs, all of it came before the mega-round, not after. Take the money as the start of an obligation you now have to deliver on, not as the prize for having pitched well. From Dave’s video library Dave on why the raise is not the achievement and fabricated value is the disease. Quantum Systems earned its $1.2 billion on 19,000 real missions first, which is the order that makes a mega-round an asset instead of a liability. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [How to Read the Pentagon’s Drone-Budget Surge as a Hardware Founder →](/field-notes/2026-06-18-how-to-read-pentagon-drone-budget) [NEURA’s $1.4 Billion Humanoid Bet, and Why I’d Bet Narrow →](/field-notes/2026-06-13-neura-1-4b-physical-ai-humanoid-form-factor) [Sanctuary AI’s Physical-AI Claim Is One You Can Actually Check →](/field-notes/2026-06-19-sanctuary-ai-validated-number-not-demo) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Launch on the Narrowest Indication You Can Defend | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-07-04-launch-narrow-fda-indication-expansion [← All Field Notes](/field-notes) July 4, 2026 Medical Devices Most founders I meet want to launch on the biggest indication they can claim with a straight face. The math is seductive. A bigger label means a bigger patient population, so why aim small? This week offered a clean case against that instinct. Bioness cleared its PoNS device for stroke rehabilitation, a market of more than seven million Americans living with stroke-related gait disability. It reached that market by launching, five years earlier, on one of the narrowest indications it could have picked. ## The expansion is the story, not the debut PoNS is a home-use neurostimulation device. The patient places a mouthpiece on the tongue, which sends mild stimulation to branches of the trigeminal and facial cranial nerves, and uses it alongside a prescribed physical-therapy program. The stroke clearance rests on the Stroke Registrational Program, a three-study, 159-patient effort run across ten centers in the US and Canada. Active PoNS plus therapy produced a 45.5% higher response rate than therapy alone. By the study’s six-point gait-assessment threshold, 56.1% of treated patients responded against 11.1% of controls. It cleared as a 510(k), with Medicare coverage in place at launch. None of that was the device’s debut. PoNS first reached the US market in 2021 on FDA De Novo authorization, for a much smaller group: gait deficit from multiple sclerosis. In the years since, it built commercial payer coverage with names like UnitedHealthcare, Anthem, and Aetna. The stroke market was always the ambition. MS was the beachhead. ## Why launching narrow is a regulatory on-ramp, not a tax Here is the mechanic most founders miss. The stroke indication cleared as a 510(k), not a second De Novo. That was possible because the 2021 De Novo created a new device type, and PoNS became its own predicate. Its next indication could clear against itself. I make this point about De Novo constantly, because founders treat it as a punishment, a sentence to extra time and extra work. It is closer to the opposite. Being routed to De Novo means the FDA already views your device as moderate-risk. You write the safety story instead of inferring a competitor’s. And you become the predicate the rest of the category has to match. I took Galen through exactly that. We were routed to De Novo, drafted the proposed regulation ourselves, and the FDA adopted our wording almost verbatim. That language still governs the category. Launching on a clean, defensible first indication is what earns that predicate position. The broad label you skipped on day one becomes a cheaper filing later, built against your own device instead of a stranger’s. If you are building neurostimulation, diagnostics, or anything with a novel mechanism and several candidate indications, that sequence is worth more than the wider claim you could fight for today. ## The big market was the point all along Launching narrow is not lowering your sights. The seven-million-person stroke market was the reason to build PoNS in the first place. You earn your way into a market that size by proving the smaller one, clinically and commercially, before you point the machine at a harder target. The evidence design, the payer relationships, the home-use workflow, all of it was rehearsed on MS first. I say a version of this to first-time founders often. The billion-dollar ambition is not your opening move. You build the first small, real thing, and the appetite for the large outcome grows into you as you come to understand the product and the market. For a medical device, the narrow indication is that first real thing. One honest caveat. Becoming the predicate helps only when the category rewards it. In a commoditized class with a line of fast followers, defining the standard can cost you more in clarifications and defense than it protects. Narrow-first is a strong default, not a law. Dave’s take A bigger first indication feels like ambition. More often it is impatience wearing ambition’s coat. The founders who reach the seven-million-person market tend to be the ones who were willing to win a smaller one first, cleanly enough that the regulator, the payers, and their own evidence all pointed the same way. Pick the indication you can genuinely defend, and let the market you actually want be the thing you grow into. From Dave’s video library Dave on winning your first ten customers before you chase the market. Bioness ran the regulatory version: a narrow MS indication won cleanly in 2021 became the platform for reaching seven million stroke patients now. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Tissium’s $68M Raise and the Power of a Narrow De Novo →](/field-notes/2026-06-27-tissium-narrow-de-novo-nerve-repair) [When Beating the Predicate Breaks Your 510(k) →](/field-notes/2026-06-21-when-better-specs-break-510k-equivalence) [Aidoc’s Radiology AI Got a Breakthrough Tag, Not a Clearance →](/field-notes/2026-06-29-breakthrough-designation-is-not-clearance) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Merck's $11.3B Bio-Techne Deal and the Premium on Being Indispensable | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-07-05-merck-bio-techne-11b-tools-acquisition [← All Field Notes](/field-notes) July 5, 2026 Diagnostics Merck KGaA is about to spend $11.3 billion in cash on a company that sells antibodies and reagents, not a blockbuster drug or an AI platform with a slick demo. Bio-Techne is a half-century-old catalog of proteins and lab consumables that sits quietly inside almost every researcher’s freezer and a good share of diagnostics developers’ workflows. Merck agreed to pay well above the market price to get it. For a diagnostics or tools founder, the size of the check matters less than what the check is buying: a business almost nobody can rip out. ## What $11.3 billion actually buys Bio-Techne is a Minneapolis supplier with roughly 3,000 employees, more than $1.2 billion in fiscal 2025 sales, 34 locations, and 15 manufacturing sites. Its catalog runs to about 6,000 proteins and 425,000 antibodies, plus immunoassay kits, ProteinSimple analytical instruments, RNAscope spatial-biology tools, and cell-therapy materials. Merck is paying $73 a share, all cash, a 36% premium to Bio-Techne’s one-month average price, in its largest life-science acquisition since it bought Sigma-Aldrich in 2015. The deal is expected to close late this year or early next. None of that reads like a breakthrough, and that is the point. What a strategic buyer pays a premium for here is not novelty. It is that thousands of labs and assay developers have already built and validated their work on Bio-Techne’s specific reagents. Swap a reagent and you re-run the validation. That friction is worth billions, because it means next year’s revenue arrives whether or not anyone at Bio-Techne does something clever. Merck already runs one of the world’s largest life-science supply businesses, and a deep, cataloged reagent base bolts straight onto what it sells. A buyer pays up for that kind of certainty in a way it never will for a promising pipeline. ## Why the catalog is so hard to walk away from I think about defensibility in two flavors. There is the coercive kind, where you trap a customer with contracts and pre-sold consumables, and they leave the moment a competitor offers them a door. Then there is the good kind, where you have solved the customer’s problem so completely that switching makes no sense even with the door wide open. I usually explain it with da Vinci: surgeons are trained on it, the operating room is built around it, and Intuitive keeps it easy to buy and maintain. Nobody is trapped. Leaving is simply the worse choice. Bio-Techne has quietly built the same thing in the lab. Once a diagnostics developer has validated an assay on a specific antibody, ripping it out means re-running the validation, and that is a cost almost nobody volunteers for. That is the part worth internalizing if you are building in diagnostics or life-science tools. The unglamorous component sitting inside everyone else’s experiment does not look like much on a pitch slide, and it is easy to file under the stuff that supports the real work. But it is often the most durable thing you own, because a customer who would have to redo months of work to replace you is a customer who stays. It is worth treating that component as the business rather than the plumbing. That is not a reason to build something dull and wait for a bank to call. Plenty of suppliers that look indispensable are one substitution away from irrelevant. The Bio-Techne kind of durability comes from being wired into work other people cannot easily redo: regulated assays, validated pipelines, materials with a documented history behind them. If you are building in diagnostics or life-science tools, ask early whether a customer could swap you out over a weekend, or whether doing so would cost them a quarter of revalidation. The second answer is the one a strategic buyer pays for. Dave’s take A deal like this can look like luck from the outside, a founder who happened to build something a giant wanted. It rarely is. The companies strategics pay premiums for are usually the ones built years earlier to run without any one person in the room: predictable revenue, customers held in place by their own validated workflows, a catalog that keeps selling on a slow Tuesday. An exit is not a final act you fall into. It is a design choice, and the indispensable, unglamorous version of your company is often the one worth the most. From Dave’s video library Dave on the difference between building a job and building a business, which comes down to whether it runs without you. Merck paid an $11.3 billion premium for exactly that: a company built to sell on a slow Tuesday with nobody heroic in the room. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [A Cleared Device Is Not a Business: Zimmer Buys Pacira’s iovera →](/field-notes/2026-07-02-cleared-device-is-not-a-business) [Why a 3-Hour-Faster Blood Culture Test Is a Real Product →](/field-notes/2026-06-25-faster-blood-culture-is-a-real-product) [Quantum Systems’ $1.2B Round: The Number Defense Founders Should Study →](/field-notes/2026-07-03-quantum-systems-1-2b-drone-round) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # The Pentagon's New Drone Integrator and Your DoD Buyer Map | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-07-06-pentagon-drone-integrator-dod-buyer-map [← All Field Notes](/field-notes) July 6, 2026 Defense Hardware A founder building dual-use hardware can spend a year courting one service’s program office, learning its budget cycle and finding its champions, and then watch a reorganization move the purchase somewhere else. That is roughly what happened this month. On July 1, the Pentagon made public a memo from Defense Secretary Pete Hegseth creating a single office to run nearly every unmanned and autonomous program across the department, Breaking Defense reported. If you sell autonomy into the U.S. military, this is less a policy debate than a go-to-market problem: your buyer may have just moved, and the map you were selling against is out of date. ## What the new office absorbs The office is the Direct Reporting Portfolio Manager for Unmanned Systems, or DRPM-UxS. Per Breaking Defense, it is meant to be the single joint integrator for the department’s unmanned and autonomous programs, and it reports to the deputy secretary of defense rather than living inside any one service. The portfolio is wide. It covers small drones in groups 1 through 3, autonomous ground vehicles, most unmanned surface vessels, and the software behind autonomy and swarming. The Defense Autonomous Warfare Group and a counter-drone task force become deputy offices under it. Hegseth’s reasoning was blunt: adversaries build millions of unmanned systems a year, the United States has been slow to field them at scale, and drones are, in his words, the most consequential battlefield innovation of this generation. I have watched a founder spend a year building trust with a single program office, only to have a reorg move the buy out from under them. The deep-pocketed customer is real. The org chart around it is not stable, and this memo is a large version of that instability. ## What consolidation does to your go-to-market When a buyer consolidates, the number of doors drops and the weight behind each one goes up. Selling autonomy service by service used to mean many separate conversations, each with its own budget line and its own internal champion. One integrator means fewer conversations and much less room to route around a no. A program that used to be your named customer can become a line inside a department-wide marketplace, which is a different sale: you are competing to be stocked, not chosen. That changes what you build proof around. A signed relationship with one service office is worth less than a capability a central integrator can drop into many missions without re-integrating it each time. This is where the deep pockets get both attractive and risky. DoD money can fund technology no civilian customer would pay to build first, and that is genuinely useful. But I ask hard-tech founders to show me the two roads their technology can take before they take that money. If the only road is the kinetic one and the buyer just reorganized, you have bet the company on one customer’s org chart. If there is a credible civilian or human-saving second road, say austere-environment autonomy that also serves disaster response or surgery, or edge navigation that works anywhere, you have a business that outlasts a procurement reshuffle. ## The signal underneath the reorg Set the org chart aside and the procurement signal has been steady for a while, and this memo pushes on it again. The department keeps asking for systems that work untethered, without a dependable internet connection, and at the scale of millions of cheap units rather than a few exquisite ones. That points at edge computing rather than another dependency on hyperscale data centers, and at narrow, direct form factors rather than general-purpose humanoids. A quadruped or a small drone is far easier to make autonomous on edge compute than a machine trying to imitate a person. The autonomy warfare group Hegseth just folded into the new office is exactly where the funding will show whether that edge-first, narrow-form bet is real. If you are building general-purpose autonomy because it tells a bigger fundraising story, this reorg is a reminder that the customer is buying for delivery, not narrative scope. Dave’s take The founders who take defense money and come out intact are rarely the ones who out-lobbied a program office. They are the ones who built for the second road and for edge autonomy before an RFI asked them to, so a memo like this one lands as a tailwind instead of a fire drill. Watch what the new office actually funds rather than what the announcement says, and make sure your architecture matches the mission the Pentagon is describing, not the one on your slide. From Dave’s video library Dave on the altitude mismatch that has founders doing the CPO job themselves, what a fractional engagement costs, and why full-time is not the only fix. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [How to Read the Pentagon’s Drone-Budget Surge as a Hardware Founder →](/field-notes/2026-06-18-how-to-read-pentagon-drone-budget) [Quantum Systems’ $1.2B Round: The Number Defense Founders Should Study →](/field-notes/2026-07-03-quantum-systems-1-2b-drone-round) [Physical AI’s Hard Problem Isn’t Vision, It’s State Recovery →](/field-notes/2026-07-01-physical-ai-state-recovery-not-vision) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # The First PTSD Neurostimulation Device Cleared FDA. Why Veterans Get It First | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-07-07-ptsd-neurostim-de-novo-clearance-vs-market [← All Field Notes](/field-notes) July 7, 2026 Medical Devices When a founder finally holds an FDA clearance letter, the temptation is to treat it as the finish line. It is closer to a permission slip. Neurovalens got one of the good ones this month. On July 2, MedCity News reported that the FDA granted De Novo authorization to Modius Spero, the first neuromodulation device cleared to treat the symptoms of post-traumatic stress disorder. That is a genuine milestone. It also opens a gap most hardware founders underestimate, because a clearance tells you that you are allowed to sell, not that anyone is set up to pay you. ## What the De Novo actually buys Modius Spero is a headset worn for thirty minutes a day, usually in the evening, that sends low-level electrical pulses to the vestibular nerve behind each ear. The clearance rests on a US study of 383 adults with diagnosed PTSD, run over twelve weeks, where roughly two-thirds of participants reported clinically meaningful symptom improvement against an active sham group. Those are not soft numbers for a device treating a condition with few good options. The pathway matters as much as the data. The FDA used De Novo because there was no existing device type to match, which means Neurovalens did not squeeze into someone else’s category. It defined one. When the agency writes the regulation for a first-of-its-kind device, that regulation becomes the standard the next entrant has to meet. I have been on the writing end of that. At Galen, our De Novo ended with the FDA asking us what the new regulation should say, we sent them a paragraph, and the clearance came back with our words in it. Neurovalens is now the predicate for PTSD neuromodulation. Anyone who follows has to prove they are equivalent to Modius Spero, and Neurovalens’ own next version gets to be equivalent to itself. ## A clearance is a door, not a market The same reporting notes that broad reimbursement is not expected until later in 2027. So the device is cleared today, and the payment machinery that would let a typical clinic bill for it is more than a year out. For most companies that lag is where momentum dies. You have the letter, the press, and the LinkedIn victory lap, then eighteen months of burn while you wait for coverage codes to catch up to your clearance. This is the part I press hard on with medical-device founders, and it is the deepest vertical I work in. The economics are not a step you run after the product clears. They are the product decision. If the path from cleared to paid is not mapped before the clearance, the clearance funds a waiting room, not a business. ## Why veterans get it first Neurovalens closed that gap with its launch channel. Modius Spero reaches patients through the Department of Veterans Affairs starting this summer, ahead of the 2027 coverage timeline. That is not a consolation channel. The VA is a rare buyer that is both the payer and the provider, it controls its own budget, and it does not have to wait for a national coverage decision to adopt something. It also serves a population where PTSD is common and the need is urgent, so the mission and the math point the same direction. That is an access play, and access is usually where the value in a hard-tech device actually lives. The da Vinci did not win because any single surgeon got better with it. It won because it grew the number of surgeons who could offer a minimally invasive procedure, so far more patients got one. The same logic holds here. The first job is not universal coverage, it is getting the device to the people who need it through a buyer who can say yes today. If you are commercializing a novel device into a reimbursement lag, find your version of the VA, a single motivated buyer who owns their budget and does not need the broader payer system’s permission. Dave’s take A De Novo is one of the best positions in medtech, because you set the standard and you become your own predicate. But the clearance is the start of the work, not the payoff. The teams that turn a first-in-category clearance into a company are the ones who line up a buyer who can pay now while the rest of the market catches up. Neurovalens found that buyer before the coverage codes existed. That sequencing, not the press release, is the part worth studying. From Dave’s video library Dave on why a clearance, a demo, or a finished feature is not proof you have a market, and the only proof that counts is a costly yes from a real buyer. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour consultation →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Tissium’s $68M Raise and the Power of a Narrow De Novo →](/field-notes/2026-06-27-tissium-narrow-de-novo-nerve-repair) [A Cleared Device Is Not a Business: Zimmer Buys Pacira’s iovera →](/field-notes/2026-07-02-cleared-device-is-not-a-business) [Launch on the Narrowest Indication You Can Defend →](/field-notes/2026-07-04-launch-narrow-fda-indication-expansion) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # The First Patient-Facing Medical LLM Cleared FDA on a 510(k) | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-07-08-first-patient-facing-medical-llm-510k [← All Field Notes](/field-notes) July 8, 2026 Medical Devices There is a decision buried in UpDoc’s FDA clearance that matters more to a founder than the milestone everyone is celebrating. The company cleared the first Software as a Medical Device to put a large language model in front of patients. That is a real first: cleared late last year, debuted commercially last month with $18 million in seed funding, and read this week by regulatory lawyers as a new pathway. The reflexive take is that an AI can now manage patient care. What UpDoc actually cleared is far narrower, and the narrowness is why it worked. ## What actually cleared UpDoc’s platform, led by physician-founder Sharif Vakili, handles type 2 diabetes medication management between visits. A patient talks or types to an agent through voice or chat, and the software carries out a treatment plan a clinician has already specified: insulin titration within physician-approved parameters, lab ordering, and the follow-through a busy practice rarely has time for. It is already running at Cleveland Clinic, Allegheny Health Network, and UCSF Health. The seed round drew the American Diabetes Association, Eli Lilly, Mayo Clinic, and a stack of venture firms. This is a serious company with serious backers, and the product is real. The detail that keeps falling out of the coverage is the one that governs the whole thing. The language model does not decide. It carries out a decision a licensed human already made, inside limits that same human set. ## Why a 510(k) and not a De Novo Here is the move worth showing a founder. A first-of-its-kind LLM device had an easy argument for a De Novo, the pathway you use when nothing like your product exists yet. UpDoc went the other way and cleared a 510(k), the pathway that turns entirely on substantial equivalence. A 510(k) means you are telling the FDA your device is equivalent to something already on the market, doing a recognized job within recognized bounds. You cannot make that argument while also claiming your device is smarter, faster, or better than what came before. The most expensive word in a 510(k) is “better.” A superiority claim changes what you are asserting, from “treat me as an equivalent tool” into “I am different and superior,” and that invites exactly the harder review a 510(k) is meant to avoid. So the discipline is to frame every difference, including a brand-new language model sitting in the loop, as still equivalent and safely evaluated within the same limits. UpDoc scoped the indication so tightly, one condition, provider-written plans, human-set parameters, that a chatbot talking to a diabetic patient reads to a reviewer as a bounded medication-management tool rather than an autonomous clinician. The scoping was the strategy. ## The hype gap works in your favor There is a wide space between what the headlines say this clearance means and what the label says it means, and that space is useful to any founder being pushed for an AI story. Most of what gets sold as “AI” is machine learning wearing a bigger name, and the FDA’s own records make the point: well over a thousand Software as a Medical Device authorizations already involve machine learning. UpDoc is not the first “AI” medical device by a wide margin. It is the first with a patient-facing language model, which is a genuine and specific advance, and it earned the clearance by describing itself in checkable, bounded terms instead of grand ones. When an investor asks a hard-tech founder what the AI plan is, the answer that actually funds looks like UpDoc’s, not like a launch post: a bounded task, a real clinician holding the judgment, and an indication tight enough to clear. Dave’s take UpDoc’s clearance will get read as the day AI started practicing medicine. I read it the other way around. They got a language model past the FDA precisely by not letting it behave like one, scoping it to carry out a plan a licensed clinician already signed. The teams that win the AI-in-medtech decade will be the ones who can make their software sound modest to a regulator and still solve a real problem. Modesty on the label is a design skill, not a concession. From Dave’s video library Dave on why most of what gets sold as AI is rebranded machine learning, using the FDA’s own list of more than a thousand cleared machine-learning devices to make the case. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [When Better Specs Break Your 510(k) Substantial Equivalence →](/field-notes/2026-06-21-when-better-specs-break-510k-equivalence) [The Narrow Indication Behind an AI Medical Device Clearance →](/field-notes/2026-05-31-narrow-indication-ai-medical-device-clearance) [The First PTSD Neurostimulation Device, and Why Veterans Get It First →](/field-notes/2026-07-07-ptsd-neurostim-de-novo-clearance-vs-market) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # A Longer FDA Approval Clock Is a Runway Problem | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-07-09-fda-approval-clock-is-a-runway-problem [← All Field Notes](/field-notes) July 9, 2026 Commercialization Picture a founder sizing her Series A around a clearance date. The plan pencils out: submit in the spring, clear by year end, start booking revenue on a device the FDA all but waved through in a pre-submission meeting. Then the review runs months longer than the model assumed, and a runway that looked comfortable at eighteen months turns into a countdown. New numbers out this month say that founder is now the base case, not the unlucky exception. FDA device approvals are climbing and the time to earn them is stretching at the same time, and the space between those two facts is where hardware companies get hurt. ## The number that moved The data comes from a BTIG research note by analyst Ryan Zimmerman, reported by MedTech Dive. Through June, the FDA granted 23 premarket approvals, its most stringent device review, up from 13 in the same stretch last year. More approvals is the good news. The catch is timing. The average wait for a premarket approval reached almost 599 days in the first half of 2026, against roughly 402 days a year earlier. A few slow outliers pulled that average up, but the direction is not in doubt. The lighter pathways moved too, just less. The average 510(k) decision ran about 156 days, a week or so longer than last year, across 1,669 clearances. De Novo classifications held at 14, with review time up a little over eight percent. Zimmerman made the operator point himself: longer decisions matter most for capital runway and demand forecasting, and most of all for smaller companies bringing something new to market. Sit with that. A slower clock does more than push a milestone to the right. It rewrites the runway math every early-stage hardware company runs, and that clock belongs to someone else. ## Treat the clock as a product input For a Series A hardware founder, a stretching approval timeline is a product and finance decision before it is a regulatory one. Three habits keep a longer clock from becoming a fatal one. Model runway to the date you would hate, not the date you hope for. A pre-submission meeting that goes well is one reviewer’s read on one day, not a promise. I watched this at Galen. We told the FDA in a Q-submission that we would run a sixteen-subject cadaver study with residents, and the agency wrote back, sounds great, go for it. In substantive review they turned around and questioned the exact study design they had approved in writing, and I had two weeks to recruit sixteen attending surgeons and re-run the whole thing. The timeline moved because a clock I did not control moved. Budget for that version of events, not the friendly one. Keep your indication and your study population tight, because both are claims, and loose claims add months. The cleaner the thing you are asking the FDA to agree with, the faster it tends to go. This holds well past medtech. A defense supplier waits on a certification authority, a grid-hardware startup waits on interconnection approval, an aircraft company waits on the FAA. The name on the gate changes. The fact that you do not own it does not. Keep the commercial story alive so a delay does not zero you out. Revenue, costs, and runway are the three numbers that decide whether you reach the approval at all, and runway does not care about your ambition, it cares about math. The founders who survive a long clock tend to have a way to earn or conserve while they wait, rather than betting the company on a single date landing on schedule. Dave’s take A longer approval clock does not kill hardware companies. A runway modeled to the optimistic version of that clock does. Read the BTIG numbers as a planning input rather than a complaint and they are a gift: build to the date you would hate, keep your scope tight enough to be believed quickly, and hold enough cash that a six-month slip is an inconvenience instead of an obituary. The founders who cross the prototype-to-product gap treat the regulator’s calendar as a fixed cost they already budgeted for, not a surprise that shows up late. From Dave’s video library Dave on the handful of numbers a founder actually needs to watch, including the runway math that decides how long you can afford to wait on a slow approval. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Why FDA Rehiring Will Not Speed Up Your Review →](/field-notes/2026-06-28-fda-rehiring-wont-speed-your-review) [A Cleared Device Is Not Yet a Business →](/field-notes/2026-07-02-cleared-device-is-not-a-business) [Record Hardware Funding, and the Higher Bar It Buys →](/field-notes/2026-06-08-hardware-funding-records-higher-bar) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Why Eighty Percent of Factories Still Have No Robots | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-07-10-eighty-percent-of-factories-have-no-robots [← All Field Notes](/field-notes) July 10, 2026 Industrial Robotics A mid-sized plant manager reads the same robotics headlines you do. Physical AI raising billions, humanoids on every trade-show stage, a foundation model that supposedly generalizes across any task on the floor. Then he walks his own line, where scheduling still runs on a spreadsheet and not one machine moves on its own. The distance between what the vendors announce and what the floor actually runs is the defining tension in manufacturing right now, and a number back in circulation this month puts a hard edge on it. Eighty percent of United States manufacturing facilities operate with zero automation. Not lagging, not partial. Zero. For a hardware founder selling into that market, that is uncomfortable and useful at the same time: the capability race the industry keeps covering is not the race that decides who wins. ## Belief is high. Deployment is almost nothing. The figure comes from Intrinsic chief technology officer Brian Gerkey, whose company builds robotics software inside Google, and was reported by Manufacturing Dive. Set it against how manufacturers talk about their own future. Deloitte’s 2025 Smart Manufacturing survey found that ninety-two percent believe smart manufacturing will be the primary driver of competitiveness over the next three years. Yet only about twenty-nine percent said they already use AI or machine learning at the facility or network level, and just twenty-four percent have deployed generative AI. Jeff Burnstein, who runs the Association for Advancing Automation, put the gap plainly to Manufacturing Dive: interest is high across the board, but execution is where things get difficult. Belief is nearly universal. Deployment is a rounding error against it. ## The bottleneck is not the robot When a project stalls, founders assume the machine was not good enough. On a factory floor it is almost never the machine. The barriers manufacturers actually name are cost, integration complexity, a shortage of people who can deploy and maintain the system, and no clear return. Jasmeet Singh of Infosys described plants still running on fragmented legacy systems, with data that was never structured for a robot to use, and buyers who want proven payback before they commit real money. The hard step is not the demo. It is the jump from a pilot to a measurable business outcome, and that is exactly where most of these projects die. I have watched this failure mode well outside manufacturing, and it always rhymes. Physical AI has its own version, the graveyard of pizza-delivery robots and their cousins: impressive machines, real working demos, dead companies. They fell in love with their engineering, not with their customer. A working demo feels like progress, but it is not progress if nobody has proven they will pay for it at a margin that lets the company live. The plant manager is not buying a robot. He is buying lower cost or lower liability on one specific process, and if the machine does not deliver that, what are you actually selling him? ## The economics have to close on the first line, not the fiftieth This is where a lot of robotics founders get the sequence backward. Unit economics do not gate your hundredth deployment. They gate your first one. A plant will not fund a science experiment, so the payback math has to make sense on day one or you never get onto the floor at all. That discipline is why the more interesting news this month is not another billion-dollar physical-AI round. It is a wave of vendors meeting the eighty percent much closer to zero. No-code cobots that need no dedicated cell and no specialist programmer, from Hirebotics on an explosion-proof painter to Productive Robotics on machining. Ambi Robotics and Pickle Robot shipping what they call the first commercial physical-AI system that unloads a truck and palletizes with no person in the loop. Techman Robot naming the culprit directly, the hidden cost trap of long commissioning cycles and endless reprogramming that quietly erodes the payback. Every one of those moves lowers the cost of getting started, which is the number the buyer is actually weighing. Dave’s take Eighty percent of factories sitting at zero is not a story about robots that cannot do the work. It is a market waiting for someone to make the deployment and the day-one math close. When a physical-AI founder walks me through how the model benchmarks, I care less about that and more about whether a plant manager can run the payback in his head and say yes on the first line. Build the customer proof before you build the fleet. The companies that win this cycle will treat integration cost and unit economics as product decisions, not as someone else’s problem to solve after the demo lands. From Dave’s video library Dave on separating where automation actually returns money from where you are just feeding the meter, the same payback question a plant manager asks before a robot ever reaches the floor. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Physical AI Is a State-Recovery Problem, Not a Vision Demo →](/field-notes/2026-07-01-physical-ai-state-recovery-not-vision) [What Sanctuary AI Validated Was a Number, Not a Demo →](/field-notes/2026-06-19-sanctuary-ai-validated-number-not-demo) [A Cleared Device Is Not Yet a Business →](/field-notes/2026-07-02-cleared-device-is-not-a-business) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Non-Surgical Brain Interfaces Win on Access. That Is Not the Hard Part. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-07-11-non-surgical-brain-interfaces-access [← All Field Notes](/field-notes) July 11, 2026 Medical Devices A widely shared perspective made the rounds this month arguing that brain-computer interfaces no longer need surgery to be transformative, and that the choice between a surgical and a non-surgical implant has become a strategic question rather than a technical footnote. I think that read is right as far as it goes. It also stops one decision short of the thing that decides whether any of these companies turns into a business. The modality debate is real. It is also downstream of a product question most brain-interface teams have not answered out loud: what measurable improvement are you selling, and who pays for it? ## Access is the advantage, and it is a big one Start with what the non-surgical camp gets right. Synchron’s Stentrode is delivered through the jugular vein and seated in the vasculature next to the motor cortex, no craniotomy involved, and it has let people with ALS operate a computer, send messages, and bank online. A 2023 study in JAMA Neurology reported meaningful communication rates with a safety profile that does not compare to open neurosurgery, and the FDA granted the approach Breakthrough Device Designation back in 2021. The surgical path, by contrast, is slow by construction. The first BrainGate2 trial opened in 2009 and enrolled fewer than twenty participants across a decade, because a craniotomy gates who can enroll and how fast anyone learns. This is the same reframe I use for the da Vinci. The robot did not win the operating room because it made surgeons measurably better than their own hands. It won on access, by putting minimally invasive capability in front of far more surgeons, and therefore far more patients. Non-surgical brain interfaces are running the same play. Lower procedural risk means a wider eligible population, and the Lancet Commission on Global Surgery estimates five billion people have no access to safe surgery at all. Access, not raw signal fidelity, is where the addressable market actually opens up. ## But access to what, exactly? Here is where the strategic framing quietly skips a step. Access only means something once you have defined what the patient is getting access to, and that is the part the field has not settled. STAT reported in February that Synchron, Neuralink, and Paradromics are all finding it easier to show their devices are safe than to define and measure an improvement in patient outcomes. A brain interface does not cure a disease, so the therapeutic value is genuinely hard to specify and harder to prove to a regulator or a payer. That is not a surgery problem. A non-surgical device with an undefined endpoint is exactly as unfundable as a surgical one. I ran into how much the definition matters at Galen. The same physical device could be framed as a remote manipulator, which lined it up against the da Vinci and drew a clinical-evidence bar we could not clear, or as a tool that improved a surgeon’s access at the end of a long instrument, which fit a de novo and cleared. The hardware never changed. What changed was the definition of what the device was, and that decided the entire regulatory and commercial path. ## Modality is a decision you earn, not the one you lead with So the order of operations is the reverse of what the headlines imply. First define the win you can measure: the specific capability restored, the outcome a clinician will document, the number a payer will reimburse against. Then choose the modality that delivers that win to the widest population at a risk the market will accept, which today often points to non-surgical. Lead with the modality and you have optimized the delivery of a benefit you never defined. This is product and market definition, the first thing I work through with any hard-tech founder, and in a medical device it belongs to the person binding the product to a market, not to the regulatory consultant who implements the plan later. Dave’s take The non-surgical camp has the direction right, and I think it has the market right too. Just do not let the modality argument stand in for the harder work. The company that wins here will be the one that can say, in a sentence a payer understands, what its interface gives back and how you measure it, and only then argue about how it gets into the body. Define the win first. Surgery, or no surgery, is a delivery decision after that. From Dave’s video library Dave on how to find the single assumption worth testing before you build, the same discipline a brain-interface team needs before it argues about surgery. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [The First PTSD Neurostimulation Device Cleared FDA. Why Veterans Get It First →](/field-notes/2026-07-07-ptsd-neurostim-de-novo-clearance-vs-market) [Launch on the Narrowest Indication You Can Defend →](/field-notes/2026-07-04-launch-narrow-fda-indication-expansion) [Aidoc’s Radiology AI Got a Breakthrough Tag, Not a Clearance →](/field-notes/2026-06-29-breakthrough-designation-is-not-clearance) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # New Robot Safety Standards Are a Market-Access Gate, Not Paperwork | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-07-12-robot-safety-standards-market-access-gate [← All Field Notes](/field-notes) July 12, 2026 Commercialization When I was helping get a surgical robot through the FDA, I made a call early that felt like overhead at the time. We treated the risk and quality discipline as a product requirement from the first prototype, not as a binder we would assemble the month before we filed. That decision is why a standards story out of the robotics world this week landed for me. The Robot Report, syndicating analysis from Interact Analysis, asked whether robot suppliers are ready for the updated ISO 10218 safety standard. Most of the coverage reads it as a compliance chore. I read it as a market-access gate, and the companies treating it as a chore are the ones about to lose their shelf space in Europe. ## What the standard actually changes The mechanics are worth getting right. In Europe, the Machinery Directive that has governed industrial equipment since 2006 is replaced by the new Machinery Regulation, EU 2023/1230, on January 20, 2027. ISO 10218:2025, the revised safety standard for industrial robots, is what companies will build to for a CE mark under that regime, once it is formally listed in the Official Journal of the European Union. That listing was still pending as of mid-2026, and when the standard was last revised in 2011 the harmonization took more than a year, so the full legal transition could slip past 2027. In the United States the parallel update, ANSI/A3 R15.06-2025, is voluntary, but integrators write it into procurement specs and OSHA reaches for it after an accident, so it is optional in name only. Then comes the part that decides winners. Interact Analysis interviewed manufacturers and found a wide readiness gap. The established global vendors are prepared. Some mid-sized suppliers, including European ones, have limited visibility into what the new standard actually requires, and emerging Asian entrants, especially Chinese firms pushing into Europe, are the least ready and most likely to underestimate what compliance will cost them in R&D. The analysts’ read was blunt: the well-prepared benefit, and the underprepared risk losing market access. ## Why the prepared win, and it is not size The incumbents are ready because they treated the standard as knowledge they own, not a service they buy at the end. That is a play I have run in every regulated thing I have built. Learn the regulation well enough to design to it yourself, do about ninety percent of the work in house, and pay the outside specialist only to tell you what you missed and to certify the result. What you never do is outsource the understanding. A supplier that hands the whole compliance question to a consultant six months before the deadline is buying a document, not a design, and the gap shows up as cost and delay at the worst possible moment. It is the same instinct as framing a product to fit the regulatory envelope it has to live in. Whoever owns the product owns how it meets the standard, because that decision shapes the roadmap, the bill of materials, and the test plan long before an auditor is in the room. Push it down to a compliance team as a checkbox and you have already given up the months the prepared vendors spent building it in. ## The move if you sell hardware into a gated market If you build industrial robots, automation, or any hardware headed for Europe, the takeaway is narrow and practical. Put ISO 10218:2025 into your product requirements and your cost model now, not in 2027. Plan as if the deadline holds even though it might slip, because being early costs you a little and being late costs you the market. And keep the understanding of the standard inside the company, in the people making design decisions, instead of renting it by the hour. None of this is unique to robotics. A founder selling safety-critical hardware into any jurisdiction with a certification gate, from medical devices to grid and energy hardware, faces the same test. Is the standard a requirement you designed to, or a surprise you have to survive? Dave’s take A safety standard changing is one of the most predictable events in hard tech. You can see it coming years out, which means the companies caught flat footed were not unlucky, they were unprepared. Build the standard into the product early and it stops being paperwork and starts being a moat, because on the day the rule takes effect your competitors are scrambling for certification while you are already on the shelf. A gate does not only keep unsafe products out. It keeps unprepared companies out. From Dave’s video library Dave on the value-of-your-time test for deciding which work to own yourself and which to hand off, the same judgment behind keeping a standard in house instead of renting it. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [What FDA’s QMSR Transition to ISO 13485 Means for Founders →](/field-notes/2026-06-24-qmsr-iso-13485-quality-system) [A Longer FDA Approval Clock Is a Runway Problem →](/field-notes/2026-07-09-fda-approval-clock-is-a-runway-problem) [Eighty Percent of Factories Have No Robots. That Is a Product Problem. →](/field-notes/2026-07-10-eighty-percent-of-factories-have-no-robots) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Europe’s Biggest Fusion Raise Buys Runway, Not Fusion | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-07-13-fusion-mega-raise-is-runway-not-a-moat [← All Field Notes](/field-notes) July 13, 2026 Climate Hardware When a hardware company raises a number this big, the coverage treats the raise itself as the milestone. I read it the other way around. Last week the Munich company Proxima Fusion announced a €411 million round, about $468 million, at a €2.4 billion valuation, the largest private fusion investment ever recorded in Europe, with Google and the German utility RWE among the strategic backers. That is real validation of a serious company. It is also a clock. The money does not buy fusion. It buys the one demonstrator that still has to prove fusion is possible, and it restarts the countdown on everything the company just promised the people who wrote the checks. ## What €411 million actually buys Proxima is a stellarator company, the first spin-out from the Max Planck Institute for Plasma Physics, built on the Wendelstein 7-X research program. The round, led by XTX Ventures and East X Ventures, funds a machine called Alpha, a net-energy demonstrator to be built near Munich and targeted for the early 2030s. Alpha is the bridge. Behind it sits Stellaris, which the company calls the world’s first commercial stellarator power plant, planned for later in the decade, with RWE lining up to host it on the site of a retired fission plant in Bavaria. In under three years Proxima has now raised more than €650 million, including €95 million in public grants, and this round alone cleared its goal of matching Bavaria’s €400 million public commitment. Read that timeline again as a founder, not a fan. The €411 million funds one demonstrator whose whole job is to show net energy, which is a physics result. The commercial plant that actually sells electricity is a separate machine, years later, and whether it wins turns on a different question altogether. ## Money is runway, not a moat A raise is not a moat. It is runway, a clock ticking against the milestones you promised investors, and the size of the headline tells you how loud the countdown is, not how strong the company is. The bigger the number, the higher the bar you have set for the round after this one. So when I see a mega-raise I do not cheer the banked figure. I ask what the money has to buy to matter, and then I watch what the company builds, not what it banked. For a fusion company the money has to buy answers to two questions, and the raise is neither. Can the physics that works in a demonstrator survive at plant scale. And when it does, will the delivered cost per kilowatt-hour beat what the grid can already buy. That second question is the one founders in every capital-heavy category underrate. Net energy in a demonstrator is a milestone. A competitive price on the grid is the product. The economics are not a slide you bolt on after the science works. They are the thing you are actually selling. ## The move if you build capital-heavy hardware If you build in a capital-heavy category, climate and energy hardware, advanced materials, anything where the first unit costs a fortune and payback is years out, the pull after a big round is to treat the capital as proof. It is closer to the opposite. A large round buys you the right to attempt the hard part on a deadline, and it raises the bar you must clear to raise again. Strategic investors sharpen that even more. Google and RWE are not only money, they are a possible first customer and a build partner, which is an asset and a dependency at the same time. Plan for both. Dave’s take I want Proxima to win and I hope the science lands. But I have watched enough big numbers to know the raise is the easy part to celebrate and the hard part to earn out. The founders who make rounds like this pay off are the ones who treat the capital as a countdown rather than a trophy, and who keep the cost question in front of them the whole way, so that the day the physics works they already know it can be sold at a price the grid will pay. Watch what they build with the €411 million. That is the number that will end up mattering. From Dave’s video library Dave on why hitting a giant valuation is not the same as building something real, and why fabricated value, not company size, is the thing worth watching. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Quantum Systems’ $1.2B Round: The Number Defense Founders Should Study →](/field-notes/2026-07-03-quantum-systems-1-2b-drone-round) [The 2026 Hardware Funding Records Hide a Higher Bar, Not Easier Money →](/field-notes/2026-06-08-hardware-funding-records-higher-bar) [Energy Dome’s CO2 Battery Wins Where Lithium Can’t →](/field-notes/2026-06-26-google-energy-dome-co2-battery-long-duration) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Console-Free Cryocatheter: Cutting the Console Is the Product | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-07-14-console-free-cryocatheter-is-the-product [← All Field Notes](/field-notes) July 14, 2026 Medical Devices Somewhere in Endocision’s development, someone decided to throw out the console. Not shrink it, not make it cheaper to build, remove it entirely. That is the decision I would want to understand first, because it is the one that decides whether this company has a business. Last week the Montreal company announced the first human cases of its Cryocision system, a flexible cryocatheter for lung procedures that cleared the FDA and Health Canada and then treated its first patients in Quebec hospitals. The clinical firsts are real. But the move that matters commercially is the one buried in the product description: there is no capital equipment to buy. Cut the console and you have not built a smaller device, you have changed who is allowed to buy it. ## What cutting the console actually does The Cryocision platform replaces the electromechanical console that legacy flexible cryo systems depend on with a miniaturized refrigerant dispenser. It is hand-held or clipped to the handle of a standard bronchoscope, and it pushes pressurized CO2 down the scope’s working channel to cool the tip of a single-use 5 French catheter. One operator, fast setup, nothing to sterilize and reload. The capital equipment that used to gate the whole category is simply gone. A capital console is not only an expensive line item, it is a different purchase entirely. It goes to a value-analysis committee, competes against every other capital request in the hospital, and can sit in a budget cycle for a year. A wholly single-use device is an operating cost, the kind a department can often approve without the capital gauntlet. Endocision’s own chief medical officer said the point out loud, that the system is meant to reach centers that could not afford the capital expenditure. Cut the console and you have not only lowered the price, you have changed the buyer and the path to yes. I learned how unforgiving that math is the hard way. On a surgical robot I worked on, we aimed at a niche no other system served, and the catch was that the underserved procedures were the low-reimbursement ones, so the economics had to close on day one or we would never see a second hospital. ## The disposable is the business model, and that cuts both ways Removing the console does not remove the economics, it relocates them. The razor-and-blade shape is proven, da Vinci runs on it, but it moves the burden onto every single procedure. Each case now has to carry its own margin, the per-use cost has to stay low enough that a cryobiopsy pencils out against a needle or forceps biopsy, and the reliability has to hold up thousands of times with no console to fall back on. The upside is a recurring revenue path that grows with adoption instead of a one-time box sale. The risk is that a thin per-procedure margin, multiplied across the volume you were counting on, is where a disposable business quietly bleeds. That is a spreadsheet you want built before the first launch, not after. ## The read if you build capital-heavy hardware If your product today ships with a console, a cart, a reader, a base station, anything a customer capitalizes, it is worth asking what falls away if you delete it. Sometimes the honest answer is nothing, the capital equipment is the product and the disposables are the accessory. But when you can push the intelligence into a single-use tip and let the customer bring the platform they already own, the standard bronchoscope here, you change the sale from a capital fight to an operating decision. This matters most to founders in interventional and diagnostic device categories, where a large installed base of general-purpose scopes and platforms is already sitting in the rooms you want to enter. The device that clips onto what is already there has a shorter path to the first yes than the device that asks the hospital to buy a new box. Dave’s take I like this launch because the hard decision is visible in the design. Somebody was willing to give up the control and margin of owning the console in exchange for a device a hospital can actually say yes to. That is a product-strategy call, made early, that most engineering teams never reach because they fall in love with the platform. The clinical firsts will get the headlines. The console they chose not to build is the part I would study. From Dave’s video library Dave on why so many founders underprice their own work, and how to set price from the value it delivers rather than what it cost to build. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [A Cleared Device Is Not a Business: Zimmer Buys Pacira’s iovera →](/field-notes/2026-07-02-cleared-device-is-not-a-business) [Why a 3-Hour-Faster Blood Culture Test Is a Real Product →](/field-notes/2026-06-25-faster-blood-culture-is-a-real-product) [Non-Surgical Brain Interfaces Win on Access. That Is Not the Hard Part. →](/field-notes/2026-07-11-non-surgical-brain-interfaces-access) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # In Orthopedic Robots, the Moat Is Never the Robot | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-07-15-orthopedic-robots-moat-is-not-the-robot [← All Field Notes](/field-notes) July 15, 2026 Surgical Robotics Years ago I worked on a university tech-transfer project that built a navigated sagittal saw, the tool that shapes bone to seat a knee implant. To make it, we bolted a navigation rig onto a Stryker saw someone had bought off eBay. The clever part was the patent, written around cutting planes when nearly every other bone-shaping robot used a burr. We convinced the patent office it was novel and won a grant. Then the company never pushed it forward, the clock on the patent ran down, and years later a much larger orthopedic company shipped a navigated saw that did the same thing better. I think about that project whenever an orthopedic-robot company raises money. THINK Surgical raised a lot of it this week. On July 14 the Fremont, California company said it had secured a debt facility of up to $65 million from Symbiotic Capital, with $25 million funded now, another $15 million tied to milestones, and up to $25 million more at the lender’s discretion. The size of the number will get the attention. The bet it funds is the part worth studying. ## What the money is actually funding THINK sells the TMINI, a hand-held miniature robot for total knee replacement. The differentiator is not the robot’s size. It is that the platform is open. TMINI runs with implant brands from multiple manufacturers, and the company says it is already compatible with roughly 70 percent of the knee-implant market by share. The surgeon chooses the implant and the robot adapts to it. CEO Stuart Simpson framed the facility as the capital that will, in his words, fully finance the company to profitability. That is a more specific claim than most device companies make. Most raise to reach the next milestone. THINK is saying this is the money that gets it to break-even. ## Why open is a real weapon against a locked incumbent The robots that dominate orthopedics today are generally sold as closed systems, matched to the maker’s own implants. Stryker’s Mako is the obvious example. That is lock-in of the coercive kind. Buy the robot and your hospital is steered toward one company’s implant catalog for years, whether or not a given surgeon prefers something else. It holds right up until a surgeon wants an implant the system will not run, and then it is a grievance sitting there waiting for a competitor to relieve it. THINK is aiming straight at that seam. An open platform lets a hospital keep the implants its surgeons already trust and add the robot on top. That is a genuinely smart place to attack, because it turns the incumbent’s lock-in into the reason to switch. ## Open is a smart opening move, not the moat Here is where my saw comes back. A differentiator convinces surgeons. It does not, by itself, keep a competitor out. Open is copyable. Any incumbent that decides the captive-implant model is costing it deals can open its own platform and erase the advantage in a product cycle. What actually holds is the surrounding system, the part a rival cannot clone by copying the box: surgeons trained and comfortable, the workflow built around the robot, service that shows up when something breaks, and the per-case economics that make each procedure pencil out. That is the kind of lock-in nobody resents, the kind you earn by solving the problem so completely that leaving makes no sense. The da Vinci has it, which is why a technically equal rival still cannot pry those operating rooms loose. My saw project had a real differentiator and nothing durable built behind it, so when the patent clock ran out the window simply closed on someone else’s better version. A debt facility, even one meant to reach profitability, buys the time to build that second advantage. It is not the advantage. Dave’s take In orthopedic robots the box is the copyable part, and THINK picked a smart place to fight, because open attacks exactly the lock-in that makes hospitals resent the incumbents. But the company’s own words point at the real work. Financing to profitability means the next few years have to go into the things a competitor cannot clone: surgeon habit, workflow, service, and case economics. I learned the price of skipping that step on a saw a bigger company eventually shipped better. The money buys the window. What THINK builds inside it decides whether open was a strategy or just a feature. From Dave’s video library Dave on separating real demand from polite encouragement, and the one kind of proof that tells you a market actually wants what you are building. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Medtronic’s Hugo Filing Is an Access Play, Not a Specs Fight →](/field-notes/2026-06-30-medtronic-hugo-general-surgery-access-play) [Karl Storz Is Retiring a Surgical Robot That Worked →](/field-notes/2026-06-23-karl-storz-retires-senhance-surgical-robot) [A Cleared Device Is Not a Business: Zimmer Buys Pacira’s iovera →](/field-notes/2026-07-02-cleared-device-is-not-a-business) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # The FDA's Review Clock Runs on FDA Days, Not Calendar Days | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-07-16-fda-days-are-not-calendar-days [← All Field Notes](/field-notes) July 16, 2026 Regulatory A founder pulls up the FDA’s newly published review-time commitments and does the obvious thing. She pencils a 510(k) clearance ninety days after submission into her Series A model. Clean number, official source, right there in the agreement. Then someone who has filed before points at three small words: ninety FDA days. Not ninety calendar days. The distance between those two phrases is where a funding plan quietly cracks, and the draft agreement the FDA put out this month is worth reading closely, because most founders will read the wrong number in it. ## What the FDA actually committed to This month the FDA released the draft commitments letter for MDUFA VI, the medical-device user-fee agreement that sets review performance goals for fiscal years 2028 through 2032. The House Energy and Commerce health subcommittee held a hearing on the reauthorization on July 15, and public comment is open until August 7. As reported by attorneys at Holland & Knight, the draft keeps the headline timelines about where they have sat: 90 FDA days to decide 95 percent of 510(k) submissions, 150 FDA days for 90 percent of De Novo requests, and 180 FDA days for 90 percent of premarket approvals that skip an advisory panel. That is stability, not acceleration. For an early-stage founder, a predictable clock is easier to plan a raise around than a fast one, so hold the disappointment. ## Read which clock it is The runway math lives in the definition. An FDA day counts only the time your submission sits with the reviewer. The moment the agency issues an additional-information request, your file goes on hold and the clock stops. Every day you then spend running the extra test or rewriting the software documentation is a day that never touches the ninety. The FDA’s own performance goals spell it out that way. So the commitment is a promise about the agency’s working time, not about your calendar. Your calendar is set by something the agreement never names: how many of those hold-the-clock loops you trigger. That count is not luck. It tracks submission quality, and submission quality gets decided long before you file. It is why I push founders to treat the pre-submission meeting as the real event. A pre-sub is not a question-and-answer session where you collect FDA opinions. You present a strategy, the predicates you plan to cite, the indication you plan to claim, the testing you plan to run, and the agency signals back whether it holds. The signal is rarely blunt. “You may be better suited to a De Novo” is not a suggestion, it is an answer, and a founder who files a 510(k) anyway after hearing it has just bought a stack of deficiency letters and all the paused-clock time attached to them. You put close to eighty percent of your eventual submission into that meeting for a reason. ## The line worth watching One item in the draft points somewhere useful. By the end of fiscal 2028, the FDA plans to run a pilot in which a device with the same intended use is submitted to the FDA and at least two other national regulators at the same time, with results published by September 2030. It is small, it is slow, and it will do nothing for anyone filing this year. But the direction matters to any hardware founder planning more than one market. Paying for the same certification twice in two jurisdictions is one of the quiet costs of regulated hardware, and it is not only a medtech tax. A grid-hardware startup clearing UL and its European counterpart pays it. So does a drone company answering to both the FAA and EASA. A regulator testing one submission read by three authorities at once is a small signal about where that cost drifts over the next decade. Dave’s take The founders who get hurt by a review clock are almost never surprised by the FDA. They are surprised by their own spreadsheet. They banked the official ninety-day figure, burned three review cycles fixing things they could have gotten right before filing, and watched a comfortable runway turn into a countdown. Read this commitments letter for what it is, a fixed cost you can budget for, as long as you budget against the right clock. The number you actually control is not the agency’s calendar. It is how clean the thing you hand them is on the first pass. From Dave’s video library Dave on making a confident call when the data is thin, the same discipline a founder needs when the one variable that sets the timeline sits on someone else’s desk. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [A Longer FDA Approval Clock Is a Runway Problem →](/field-notes/2026-07-09-fda-approval-clock-is-a-runway-problem) [Why the FDA’s New Hires Will Not Speed Up Your Review →](/field-notes/2026-06-28-fda-rehiring-wont-speed-your-review) [Launch on the Narrowest Indication You Can Defend →](/field-notes/2026-07-04-launch-narrow-fda-indication-expansion) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # For a New Implant, a Medicare Payment Code Is the Market Gate | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-07-17-medicare-payment-code-is-the-market-gate [← All Field Notes](/field-notes) July 17, 2026 Medical Devices The part of a medical device that gets written up is almost never the part that decides whether it makes money. A neuromodulation implant that runs without a battery inside the patient is a real engineering achievement, and it will get the press. Whether it has a market got settled somewhere much quieter, inside a Medicare payment table. This month Neuspera Medical pointed straight at that table, and the move is a cleaner lesson in commercialization than most funding announcements. ## What CMS actually did Neuspera makes an integrated sacral neuromodulation system, an implant used to treat urinary urge incontinence. The word integrated is doing work here: the system is powered from outside the body, so there is no implanted battery to replace in a second surgery later. On July 8 the company said the Centers for Medicare and Medicaid Services had assigned its billing code to payment level APC 5464 in the proposed 2027 Outpatient Prospective Payment System rule. In practice, its externally powered device would sit at the same Medicare payment as every other implantable neuromodulation therapy for that condition. The proposed rule also raises the payment for that group to $22,150 for 2027, an 11.8 percent increase over 2026. A year earlier the device had already been added to Medicare’s national coverage determination for sacral nerve stimulation. Coverage plus payment parity is the combination that opens the door. The market it is walking into runs about a billion dollars a year and has belonged to Boston Scientific and Medtronic. ## Why parity beats a premium A founder’s instinct is to want a bigger number. A novel device, the reasoning goes, deserves a new-technology code and a premium rate that reflects how different it is. That instinct is usually wrong. A brand-new code is one a hospital’s billing staff has never run, a payment level that can be revised later, and a reason for a value-analysis committee to wait and watch. Parity with the incumbents does the opposite. It drops the new device into a payment slot the buyer already trusts, already bills against, and already budgets for. The engineering can be genuinely new while the reimbursement stays boringly familiar, and boring is what gets a purchase order signed. What Neuspera actually bought is that a hospital no longer has to take a payment risk to try something unproven. ## The gate founders skip Founders map the FDA pathway obsessively and file reimbursement under later, something to sort out after clearance and the first few sales. That order is backwards. A clearance lets you legally sell a device. A payment code decides whether anyone can afford to buy it. Those are two separate gates, run by two separate agencies, on two separate clocks, and the second one sets the size of your market. Payment rules also move on their own annual cycle, which means the reimbursement answer can shift the year after you launch. I tell founders to bring the CFO and a coding expert into the room before the surgeons fall in love with the device, because they, not the surgeon, model whether the hospital gets paid. This is not only a medtech tax. A diagnostics startup waiting on a CPT code, or an energy-hardware company whose payback depends on a utility tariff, is standing at the same gate. Map the payment path before you lock the bill of materials. If the economics do not close for the buyer, a clean clearance is still a device nobody can order. Dave’s take Reimbursement is where a lot of good hardware quietly dies, and it dies late, after the money is spent and the device already works. Neuspera did the unglamorous thing and lined up a payment path that matches the incumbents before it had to lean on it. If you are building anything a hospital, an insurer, or a government buyer pays for on a schedule you do not set, treat that payment mechanism as part of the product, not as paperwork for after launch. From Dave’s video library Dave on selling as clarifying the buyer’s math instead of convincing them, the same move behind a hospital sale that turns on economics, not on how impressive the technology looks. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [A Cleared Device Is Not a Business →](/field-notes/2026-07-02-cleared-device-is-not-a-business) [Cutting the Console Is the Product →](/field-notes/2026-07-14-console-free-cryocatheter-is-the-product) [Brain Interfaces Win on Access, Not the Hard Part →](/field-notes/2026-07-11-non-surgical-brain-interfaces-access) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # In Defense Hardware, the Constraint Is Production, Not Design | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-07-18-defense-hardware-constraint-is-production-not-design [← All Field Notes](/field-notes) July 18, 2026 Defense Hardware Lockheed Martin does not hand part of its missile supply chain to a young robotics startup because the robots are clever. It does it because it cannot make enough parts. This month Machina Labs said Lockheed had awarded it a qualification contract to produce components for the JASSM missile using robotic metal forming, and the company’s CEO compressed the whole thesis into one line: missile programs are not constrained by design, they are constrained by production. For hard-tech founders, that sentence is worth more than the contract. ## What Lockheed actually bought Machina won a qualification contract in support of the Joint Air-to-Surface Standoff Missile program. It is the first component built with the company’s RoboForming process, two industrial robots shaping sheet metal from opposite sides, to reach qualification for a U.S. defense missile system. The parts are metal structures, specifically fuel tanks, delivered to Lockheed for testing and integration. Behind the contract sits Machina Factory 3, a 200,000-square-foot plant designed to hold up to 50 robotic cells for high-rate defense production, pitched as pulling production timelines from months down to days. No dollar figure was disclosed, and the figure is not the point. The word qualification is. In a defense supply chain, getting a new process cleared to touch a mission-critical part is the slow, expensive gate. Once you are through it, you are designed in. ## Why “constrained by production” is the whole game The design of a standoff missile is not the scarce thing. The blueprints exist and the physics is solved. What is scarce is the ability to turn titanium into thousands of identical, flight-qualified parts fast enough to matter. That is execution, and execution is where the value actually sits. I have said for years that an idea carries no intrinsic value on its own. You cannot patent an idea, only an embodiment you have built and demonstrated. Machina is the industrial-scale version of the same rule. Its edge is not a concept for forming metal, it is a qualified, repeatable process running inside a building. A competitor who reads the same press release still has to stand up the cells, pass the qualification, and hit the rate. That is the part nobody clones from a summary. ## The gate hard-tech founders design around too late Founders in defense and dual-use hardware pour their planning energy into the demonstration, the prototype that proves the physics and the video of the thing working once. The prime does not buy the demo. It buys rate, yield, and a supply chain it can qualify. If you are selling into a prime, your manufacturing readiness is your product, and it carries its own long qualification clock that runs alongside the technical one. This is not only a missiles problem. A battery startup with a working cell but no gigafactory-grade yield curve is standing at the same gate. So is a diagnostics company with an elegant assay and no way to run a million of them at spec. Map the production and qualification path before you freeze the design, because for anything mission-critical the buyer is grading the copies, not the original. A demo answers whether the thing works. A purchase order answers whether you can make ten thousand of them the same way, and those are different questions. Dave’s take The romance in hard tech is always the breakthrough, the moment the thing finally works on the bench. The business is almost never there. It lives in the boring, capital-heavy grind of making the same part a thousand times to spec, which is exactly why so few startups clear it and why a prime will pay a qualified supplier real money to have done it. If you are building something mission-critical, decide early whether you are in the invention business or the production business, because the second one is where the real company gets built, and it takes years longer than the demo makes it look. From Dave’s video library Dave reads Anthropic’s own founder playbook and shows the hard part of building did not vanish, it moved, the same reason production, not the idea, is where a company gets built. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Most Factories Still Run Without Robots →](/field-notes/2026-07-10-eighty-percent-of-factories-have-no-robots) [How to Read the Pentagon’s Drone Buyer Map →](/field-notes/2026-07-06-pentagon-drone-integrator-dod-buyer-map) [A Drone Mega-Round Is Runway, Not a Moat →](/field-notes/2026-07-03-quantum-systems-1-2b-drone-round) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # A Disposable Surgical Robot Is a Manufacturing Bet | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-07-19-disposable-surgical-robot-is-a-manufacturing-bet [← All Field Notes](/field-notes) July 19, 2026 Surgical Robotics Harel Gadot is not trying to build a better robot than Intuitive Surgical, and that is the smartest thing about Microbot Medical. The company’s Liberty system, cleared by the FDA in September 2025 to steer guidewires and catheters through the peripheral vasculature, is a fully disposable, single-use surgical robot. You use it once and throw the whole thing away. Gadot told Medical Design and Outsourcing that he wants to strip out the barriers to telesurgery so a future version can treat stroke patients remotely. Most of the coverage frames this as a technology race, a question of who gets the best robot to the brain first. I read it as a manufacturing bet, which is a much harder thing to win. ## What Microbot actually built Liberty is not a shrunken da Vinci. It is a robotic drive unit, a wireless controller, and an arm that mounts near the patient’s bedside, and because every unit is single-use, it shows up sterile and ready to run. Gadot’s pitch is that the physician walks in and starts, without the setup a fixed capital robot demands. In stroke, where his own line is that time is brain, minutes are the product. The system is cleared today for peripheral work, not the neurovascular procedures a stroke needs, but the direction is not subtle. Microbot is calling Liberty the first fully single-use, remotely operated endovascular robot, which is another way of saying it is trying to create a category rather than win an argument against an incumbent. That instinct is right. The moment a young surgical robot lines up against the installed base on the incumbent’s terms, on price and features, against a capital budget the big player already owns, it loses. Claiming a slot the incumbent does not occupy is the only move that tends to work. ## The bet is the cost curve, not the robot The disposable design is where it gets both clever and risky. A conventional surgical robot earns the way a printer does. You place the console, then sell the consumables and the service contract for a decade. Microbot has walked away from that model. There is no console to install and no capital line for a hospital to defend. The entire promise, no capital purchase, near-zero setup, and reach into a rural hospital that could never house a fixed system, rests on a single fact being true: that a robot you discard after every case can be built cheaply enough to price per procedure and still leave a margin. That is not a clinical question, and it is not a regulatory one. It is a question of bill of materials, suppliers, and volume, and it is the whole company. If the disposable cost does not fall as case volume rises, the barrier Microbot removed at the front door comes back as a per-procedure price the hospital will not sign off on. ## Which founders should read this closely This reaches past robotics. Any hard-tech founder whose product is consumed or replaced with each use is making the same wager. The diagnostics company whose margin lives inside a single-use cartridge. The drug-delivery startup whose wearable is thrown away after one dose. For all of them the manufacturing plan is the business plan. I have been on the wrong side of this. Years ago I worked on a surgical robot built for a field no other robot served, on purpose, so we would not go head to head with da Vinci. The catch was that those procedures did not carry the big reimbursement, so our economics had to come in low just to function. The opening we found in the market was open partly because the money there was thin. Picking an underserved category does not spare you the math. It usually makes the math worse, because the space is empty precisely where the economics have not closed. Dave’s take Microbot made the right positioning call and then signed up for the hardest possible way to back it. Defining a new category is the easy sentence in a press release. Building a robot cheap enough to throw away, at a price a stroke center will actually pay, at national volume, is a years-long grind through design for manufacture, suppliers, and yield, and no FDA clearance shortcuts any of it. For a disposable device, that cost curve is the moat, and it is the number I would be watching long before the first neurovascular case. From Dave’s video library Dave walks through a four-question filter for cutting a pile of startup ideas down to the one worth building, the same focus that lets a small robot own a single category instead of chasing every procedure. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [In Orthopedic Robots, the Moat Is Never the Robot →](/field-notes/2026-07-15-orthopedic-robots-moat-is-not-the-robot) [Cutting the Console Is the Product →](/field-notes/2026-07-14-console-free-cryocatheter-is-the-product) [Non-Surgical Brain Interfaces Win on Access →](/field-notes/2026-07-11-non-surgical-brain-interfaces-access) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Teleoperation Is a Trap Only If the Human Is There by Accident | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-07-20-teleoperation-trap-human-by-accident [← All Field Notes](/field-notes) July 20, 2026 Industrial Robotics Nikita Rudin spent his doctorate at ETH Zurich building simulation tools that much of the robotics industry now trains on, and he thinks the industry is quietly buying its way around the hard problem. Writing in The Robot Report, the Flexion co-founder argues that humanoid robotics has grown dependent on teleoperation, and that a machine which cannot handle anything new without fresh human input is closer to a labor system than an autonomous one. He is right about the economics. The general rule he draws from them does not hold everywhere, and hard-tech founders who miss the difference will pay for it. ## What Rudin is claiming Over the past eighteen months, he writes, humanoid robotics companies have raised billions of dollars, and most of that money is quietly paying humans to operate robots. Teleoperation datasets, by his figure, are more than 100,000 times smaller than the corpora behind language and vision models. A commercial ecosystem has grown to fill the gap, with startups in China, India, Europe and the United States selling teleoperation data and recruiting operators mostly in lower-wage economies. His alternative is reinforcement learning in simulation, which scales with compute rather than headcount. Flexion raised $50 million from DST and NVentures to build it. Founders should take that economic argument seriously. If your training data cost scales with human hours, then every improvement in the product costs you more labor rather than less, and that curve does not bend on its own. ## Where the general rule breaks Rudin's framing treats a need for fresh human input as evidence of failure, and that does not travel across hard-tech. I spent years in surgical robotics, where the human in the loop is the product rather than a stage of development. Laparoscopic surgery puts four or five rods into a patient, and da Vinci won because it was built to control more instruments than a surgeon has hands. Remove the surgeon and you have removed what hospitals are buying. The Galen design goal ran the same direction: sit smaller than the assistant it replaced, keep the surgeon in their normal working position, make the hard part of the procedure disappear under their own hands. Nobody in that operating room is asking for autonomy. The useful question is not whether a human sits in the loop, but whether that human is there by design or by accident. ## How to tell which one you have Capital is arriving faster than reliability. Manufacturing Dive reports Barclays figures putting humanoid installations at roughly 2,000 units in 2024, 15,000 in 2025, and a projected 60,000 this year, while Crunchbase counts $18.8 billion raised by robotics startups in the first half of 2026 against $15 billion for all of 2025. In the same piece, GM's Mikell Taylor sets the operating bar: 80% success is failure in production. Teleoperators live in the space between those numbers. The test I would apply is uncomfortable on purpose. Point to the human on your org chart and your bill of materials, priced and staffed for the next five years, and defend that as the product you meant to build. If the operator turns up in the demo but never in the unit economics, you are describing an unfunded liability to investors and calling it a roadmap. That risk sits heaviest on warehouse and industrial robotics founders raising on autonomy claims, and on physical-AI teams whose pilot deployments quietly run hybrid. For an interventional or surgical device the answer inverts, and the human in the loop belongs in the pitch as a feature you charge for. Dave's take Humanoids give really good demo, and that has always been the tell. I do not much care whether a founder keeps a person in the loop. I care whether they can say out loud which version they are selling, because the ones who cannot tend to sell autonomy to investors and staff it with contractors. Pick the honest version and price it, and at least the argument becomes about engineering instead of story. From Dave’s video library In this one I look at where AI agents already make decisions that affect people, and why the real risk is handing a limited system the keys with no human override. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Why Eighty Percent of Factories Still Have No Robots →](/field-notes/2026-07-10-eighty-percent-of-factories-have-no-robots) [Physical AI's Hard Problem Isn't Vision, It's State Recovery →](/field-notes/2026-07-01-physical-ai-state-recovery-not-vision) [New Robot Safety Standards Are a Market-Access Gate, Not Paperwork →](/field-notes/2026-07-12-robot-safety-standards-market-access-gate) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # En Carta Signed Its Buyer Before It Built the Line | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-07-21-en-carta-signed-the-buyer-before-the-line [← All Field Notes](/field-notes) July 21, 2026 Diagnostics En Carta Diagnostics announced a €5 million first closing on July 10. By then the Paris company had already signed a €13 million commercial agreement that commits it to delivering several million test units over five years. Buyer first, capacity second. Most hardware founders run that order backwards and then wonder why the scale-up hurts. The €5 million splits into €3 million of equity led by Blue Forest Ventures, with Ring Capital, CentraleSupélec Venture, 50 Partners Health and a group of angels alongside, plus €2 million in non-dilutive financing from Bpifrance and other partners. It follows a €1.5 million pre-seed in 2024. On its own that is an ordinary European seed story. Guillaume Horreard and Margot Karlikow founded En Carta in 2022 to put PCR-grade molecular testing into a single-use cassette a patient can run without a lab, with a result in about thirty minutes. The platform uses aptamers, which bind to a programmed target sequence, so the same cassette architecture can be pointed at different pathogens. The first indication is early Lyme detection. The second is chlamydia and gonorrhoea screening. ## The commercial commitment is larger than the round That second indication is where the AAZ agreement sits. AAZ is a French company that sells self-tests and rapid diagnostics. EU-Startups reported the co-development and commercialisation deal as a €13 million commercial partnership with committed volumes of several million units across five years in Europe. En Carta also went fabless on the hardware, partnering with Circum Medical to manufacture at volume rather than standing up its own line. An equity round tells you a handful of investors believe the thesis. A volume commitment from a company that already owns shelf space and distributor relationships tells you someone has put their own forecast behind your product. Those are different grades of evidence, and only one of them survives a bad quarter. I have watched too many hard-tech teams fall in love with the engineering instead of the customer. A working demo feels like progress, and it is not progress if nobody has proven they will pay at a margin that lets the company live. Build the customer proof before you build the fleet. En Carta did that in the right order, and the fabless decision means this capital goes into regulatory work and launch rather than into a factory nobody has ordered from yet. ## A Breakthrough Designation buys access, not permission In January 2026 the FDA granted Breakthrough Device Designation to EC Pocket Lyme, En Carta’s test for detecting Borrelia directly from skin interstitial fluid at a tick bite. The designation gives the company priority review and interactive communication with FDA reviewers and senior leadership. It does not authorize a single sale. Founders and their investors routinely read it as though it does. At Galen we learned how thin an FDA yes can be. Our Q-sub said we would run a sixteen-subject cadaver study using residents, and the agency wrote back and told us to go for it. Months later, in substantive review, the first question was whether we intended to market to residents. The study design they had approved in writing was suddenly the problem. I went to Johns Hopkins with my hat in my hand, got a free OR and cadaver heads, and re-ran the entire study with sixteen attending surgeons in two weeks. What that taught me is that a written agency response is one reviewer’s read on one day, and that your test population is itself a claim about who you intend to sell to. En Carta is running clinical trials in Boston and in Białystok, Poland, which is the right instinct. Build the evidence inside the markets you actually intend to enter, and build it as though the conversation is going to get harder rather than easier. Dave’s take The number I would want on the wall at En Carta is the AAZ volume commitment, not the FDA letter. The letter opens a door at the agency. The volume commitment is a company with a sales force saying it will move millions of units, and that is the only evidence telling you the product has somewhere to go once the door opens. If you are building a diagnostic, or any regulated hardware, get the committed order before you get the press release. From Dave’s video library Dave walks through what a fractional CPO engagement actually costs, what the work looks like week to week, and why six months is the floor before it produces anything real. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Merck's $11.3B Bio-Techne Deal and the Premium on Being Indispensable →](/field-notes/2026-07-05-merck-bio-techne-11b-tools-acquisition) [A Cleared Device Is Not a Business: Zimmer Buys Pacira's iovera →](/field-notes/2026-07-02-cleared-device-is-not-a-business) [Europe's Biggest Fusion Raise Buys Runway, Not Fusion →](/field-notes/2026-07-13-fusion-mega-raise-is-runway-not-a-moat) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # The Pulse Oximeter Worked. The Validation Didn't. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-07-22-pulse-oximeter-validation-not-the-device [← All Field Notes](/field-notes) July 22, 2026 Medical Devices For about five years, a pulse oximeter clipped to a patient's finger was treated as settled technology. Cheap, everywhere, accurate enough. What the bench tests kept missing was that the device read low blood oxygen less reliably on patients with darker skin, and it did so without setting off any alarm. This week Philips won FDA clearance for a version built to close that gap. The engineering fix is real. The more useful thing for a founder is what the last five years say about validation. Philips received a 510(k) clearance for a new reusable SpO₂ clip sensor, with optical refinements meant to hold signal quality across diverse patient populations, according to MedTech Dive. In its desaturation studies the company reported an accuracy root mean square of 1.6 percent, against the 3 percent the International Organization for Standardization recommends and the FDA recognizes. It also reported accuracy varying by less than half a percent between people with light and dark skin pigmentation. Medtronic cleared an updated Nellcor system on the same premise last month. Two of the largest device makers in the world are now competing on how well their oximeters work on everyone. ## The device always worked. The validation didn't. The hardware was never the problem. In 2021 the FDA warned about accuracy after a New England Journal of Medicine study found Black patients were three times more likely than white patients to have low blood oxygen the device missed. A 2022 JAMA study tied the same gap to delayed COVID-19 treatment eligibility for Black and Hispanic patients. The oximeters passed their tests and shipped anyway. They failed in the field, quietly, on a population the validation studies underweighted. That is the part hard-tech and medical-device founders should sit with. A verification and validation protocol is not a lab formality. It is a statement about who you think your patient is. If your desaturation study runs on mostly light-skinned volunteers, you have not proven the device works. You have proven it works on the people you happened to test. The test population is a product decision, and it belongs to whoever owns what the product is, not to a contractor filling in a template. When I ran product for a surgical robot, the definition of the device and the safety story we had to prove were mine to hold. Hand that to a regulatory consultant and you have paid senior rates for a call that should have been made in the product seat, earlier and cheaper. ## The market moved before the rule did. Here is the part that makes this more than a fairness headline. The FDA's own updated guidance on pulse oximeter testing, the one calling for at least 150 participants and a real spread of skin tones, is still only a draft. It went out in early 2025 and has sat on the agency's under-construction list ever since. Yet Medtronic and Philips both cleared devices that beat that unfinished bar, and both said so plainly. They built to a standard that is not yet law. I have been on the other side of that decision. When Galen went through its de novo, there was no regulation on the books for our device, so the FDA had to write one, and they asked us what it should say. We sent a paragraph. A day later our clearance came back with that paragraph as the regulation, word for word. We set the standard, and every later entrant has to match us. The lesson runs the same in both directions. A requirement that clearly reflects how your product gets used in the real world is not optional slack just because the agency has not finalized it. Build to it now, and you might get to shape what it becomes. Dave's take The fairness angle is real and it matters. The colder operating lesson is that a device can pass every test you wrote and still fail the patient you never put in the study. Decide who your product is for, and prove it on those people, before the FDA or a competitor decides it for you. That is product work, it sits in the product seat, and it happens long before the submission. From Dave’s video library Dave on why a working demo is not proof, and why the only clock that counts is a costly yes from a real buyer. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [FDA Days Are Not Calendar Days →](/field-notes/2026-07-16-fda-days-are-not-calendar-days) [A De Novo Clearance Is Not a Market →](/field-notes/2026-07-07-ptsd-neurostim-de-novo-clearance-vs-market) [A Cleared Device Is Not a Business →](/field-notes/2026-07-02-cleared-device-is-not-a-business) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # J&J’s Ottava Robot Cleared FDA. Beating da Vinci Is a Different Problem | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-07-23-ottava-cleared-fda-beating-da-vinci-is-different [← All Field Notes](/field-notes) July 23, 2026 Surgical Robotics Johnson & Johnson just put arguably the most differentiated surgical robot in years onto the US market. Its Ottava system cleared the FDA on Wednesday, it is smaller and more automated than the machine it is chasing, and the analysts who cover Intuitive still shrugged. One of them told clients it is unlikely to take meaningful share from da Vinci any time soon. That gap, between a genuinely better-built robot and a market that may not care yet, is the part worth understanding whether you sell robots, factory automation, or clinical software. ## What J&J actually cleared Johnson & Johnson received FDA De Novo authorization for Ottava in general surgery, covering procedures from Roux-en-Y gastric bypass and gastrectomy to cholecystectomy and appendectomy. It is the first soft-tissue system to build its robotic arms into the operating table rather than parking them on booms and carts, which is how it takes up 30 to 50 percent less room in the OR. J&J’s medtech chairman Tim Schmid called it a new category of surgical robotics, and the file cleared about six months after submission, fast for a De Novo, according to MedTech Dive. This is the second heavyweight to reach Intuitive’s turf in under a year, after Medtronic’s Hugo cleared for urology in late 2025. After more than two decades alone, da Vinci finally has company. None of this is a weak product. ## Why the better robot might still lose the room So why did J.P. Morgan’s analyst tell clients Ottava is unlikely to materially dent Intuitive’s share, and why did Stifel flag that the launch will be slow and training-heavy? Because Intuitive never defended its position on specifications. It defended on the surgeon’s hands. Twenty-five years of da Vinci means tens of thousands of surgeons whose muscle memory is wired to one console, hospitals with the systems bolted into their rooms and their instrument contracts, and residency programs that train the next generation on the same machine. Switching robots is not a purchasing decision. It is asking a skilled professional to relearn a task they perform under real consequence, and surgeons do not relearn their hands casually. There is an irony sitting inside J&J’s boldest move. The table-integrated design is the most interesting thing about Ottava, and it is also its steepest adoption cost, because a genuinely new workflow is exactly what an entrenched surgeon resists. The moment a product asks people to retrain habits they built over years, it is fighting muscle memory, and muscle memory is the strongest lock the incumbent has. ## The read for anyone selling into a trained workflow This is not only a surgical robotics problem. If your product lives inside the workflow of a trained operator, whether that is a surgeon, a machinist, a field technician, or a clinician clicking through software forty times a day, the incumbent’s real defense is almost never its feature list. It is the cost of making a skilled user unlearn what their hands already know. You have two honest options. Fit the existing workflow closely enough that switching is nearly free, or accept that displacement will be slow and deliberate and fund it accordingly. What does not work is assuming a better spec sheet earns the switch. I watched this up close building a surgical robot. If the thing did not fit what surgeons already did with their hands, it did not solve their problem, it added one, no matter how good the engineering looked on paper. Dave’s take Ottava is real, the engineering is genuinely new, and I would not bet against a company J&J’s size over ten years. But the number that tells the story here is not the footprint or the automation, it is the analyst’s shrug about market share. A cleared, better-built robot still has to beat the one thing that keeps da Vinci in every operating room, which is that thousands of surgeons already trust their hands to it. Before you count on a switch, find out exactly what your buyer would have to unlearn to choose you, and price that in. From Dave’s video library In this one Dave digs into why customers who sign up still vanish in the first month, and lands on a rule that fits this story. People do not buy better looking, they buy better working. A slicker spec sheet is not what makes someone switch. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [In Orthopedic Robots, the Moat Is Never the Robot →](/field-notes/2026-07-15-orthopedic-robots-moat-is-not-the-robot) [A Cleared Device Is Not a Business →](/field-notes/2026-07-02-cleared-device-is-not-a-business) [The First PTSD Neurostimulation Device Cleared FDA. Why Veterans Get It First →](/field-notes/2026-07-07-ptsd-neurostim-de-novo-clearance-vs-market) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # The Bearings, Not the Algorithm: Where Hardware Scale-Up Stalls | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-07-24-bearings-not-the-algorithm-hardware-scale-up [← All Field Notes](/field-notes) July 24, 2026 Commercialization At Automate 2026, the floor was full of humanoids pacing warehouse aisles and robotic arms torquing fasteners for the cameras. The conversation worth listening to happened off to the side, and it was about bearings. Analysts from ARC Advisory Group sat down with two component suppliers and left with a flat conclusion: what stops a robotics company from scaling is rarely the software. It is the bill of materials under the demo. That gap is not a robotics problem. It shows up in every hard-tech category I work in, and it is where I watch good products stall out. ## The bottleneck moved from the algorithm to the loading dock ARC's reporting from Automate 2026 found that many robotics companies now put mechanical component availability and supplier qualification near the top of their operational problems. Bearings, linear guides, actuators, structural extrusions, custom castings: each one has to be sourced, qualified, and delivered on a schedule before a single unit ships. The headline version of hardware is about autonomy and better models. The version on the production calendar is about lead times and parts with exactly one supplier. You can have a flawless controller and still slip two quarters because one actuator runs twenty weeks out and there is no qualified alternate. Procurement and supplier qualification, not algorithm development, is increasingly where the schedule actually lives. ## Reshoring makes the squeeze worse, not better The Automate discussion surfaced a tension procurement teams already live with. Building more domestic manufacturing capacity, partly to answer a skilled-labor shortage, drives more demand for automation. That automation needs its own components, and those still come from North America, Europe, Japan, and China. So the push to reshore production speeds up the same global supply dependencies it was meant to reduce. New factory orders flow back to component suppliers that are already running near capacity, and the queue gets longer for everyone behind them. For a founder, the lesson is concrete: your supplier's supply chain is now part of your risk. Its lead times, its single points of failure, and its geographic concentration belong on your diligence list next to your own runway. This is the same trap founders fall into with a working demo. A demo proves the thing can work. It says nothing about whether you can build a hundred of them on a schedule a buyer will accept, or sell them at a margin that keeps the company alive. Teams fall in love with the product and skip the unglamorous half of the job: supplier strategy, design for manufacturing, and a real second source for the parts that gate the line. The two suppliers ARC pointed to, MISUMI on the catalog side and Fictiv on the custom side, exist because the handoff from a one-off prototype geometry to a repeatable, sourceable part is exactly where teams lose months they did not budget for. ## What buyers grade now The other shift in the room was on the buyer's side. ARC found industrial buyers applying stricter tests: reliability, serviceability, deployment readiness, and parts availability came up ahead of headline performance in a lot of the conversations. A robot you cannot service because the replacement part is twenty weeks out is not production-ready, however good the autonomy looks in a clip. The demo that used to open the door does not close the volume order anymore. And none of this is unique to robotics. ARC named aerospace mechanisms, medical devices, and clean-energy hardware as the same story, all of them fields where product cycles are compressing and both prototype and production sourcing have to be handled with care. Medical devices are the vertical I know deepest, and the pattern holds there without any translation. The teams that convert a pilot into real volume are the ones that can show a qualified, resilient supply chain, not just a good highlight reel. Dave's take Productization is the least glamorous phase of a hardware company and the one that quietly kills the most of them. The roadmap that matters is not only what the product does. It is which parts you can actually get, from whom, at what lead time, and what happens to the line when your first source slips. Sort that out while you are still building the demo, not after a buyer has finally said yes and you cannot ship. From Dave’s video library Dave breaks down how to separate real operating value from AI hype, and where the money actually goes. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [In Defense Hardware, the Constraint Is Production, Not Design →](/field-notes/2026-07-18-defense-hardware-constraint-is-production-not-design) [Why Eighty Percent of Factories Still Have No Robots →](/field-notes/2026-07-10-eighty-percent-of-factories-have-no-robots) [New Robot Safety Standards Are a Market-Access Gate, Not Paperwork →](/field-notes/2026-07-12-robot-safety-standards-market-access-gate) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Cepheid’s New GeneXpert Kept the Old Menu. That’s the Product. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-07-25-cepheid-genexpert-kept-the-menu [← All Field Notes](/field-notes) July 25, 2026 Diagnostics Cepheid will show the third generation of its GeneXpert system at the ADLM meeting on July 28, and the most disciplined thing about it is what the company left alone. The new instrument is designed to run Cepheid’s existing menu of FDA-cleared tests, in the workflows labs already use, from the day it ships. The touchscreen and rebuilt software are real, but the harder engineering call was refusing to break the installed base. When you already own the machines on the bench, the product you are defending is the habit around them. Cepheid, part of Danaher, is not a startup making a bet. It is the incumbent defending one. James Green, the company’s chief technology officer, framed the redesign around keeping the simplicity and ease of use customers expect while supporting more sophisticated assays. Connie Savor, the chief medical officer, described the goal as pulling molecular tests that are scattered across separate instruments onto one system. Both statements point the same way: add capability without asking the customer to change how they work. ## The design brief was mostly subtraction Read the launch materials and the restraint shows up everywhere. The third-generation system keeps true random-access testing, so a tech can run a sample the moment it arrives instead of waiting to batch it. It keeps the existing Xpert test menu, so labs are not re-validating a new menu to go live. The first configuration out the door is a compact four-module unit that Cepheid expects to start shipping to US customers in the coming months, with higher-capacity and Infinity versions still in development. The forward-looking parts, high-plex panels, faster time to result, and sample-to-answer workflows that may eventually reach next-generation sequencing, are described as designed-for, not shipping. There is an operating reason to build this way. In a clinical lab, the workflow is muscle memory. The techs have run these cartridges thousands of times, and the lab has a validated, documented process wrapped around every assay. The moment a new instrument makes them relearn that or re-validate a menu, you have handed them a problem instead of a solution. Friction is not a rough edge you sand down later. It is the thing that makes a good product get quietly set aside for the one already on the bench. ## The menu is the moat, not the machine This is also the part that should worry a diagnostics startup. Cepheid says GeneXpert is the largest installed base of any single sample-to-answer PCR system, and whether or not you accept the exact ranking, the shape of the advantage is clear. A challenger is not competing with the box. It is competing with a cleared test menu that took years to assemble, a fleet of instruments already paid for, and a lab staff that would rather not change anything. You do not out-feature that. So do not try. The losing move is to line up against GeneXpert on menu breadth or a spec sheet, which is a budget fight the incumbent already won. The better move is to find the assay or setting the big platform serves poorly, own it, and refuse the direct comparison. That might be a near-patient workflow it cannot reach, or a decentralized site it was never built for. Pick the ground where being different is the point, and build there instead of climbing the incumbent’s hill. Dave's take The interesting product decisions in an installed-base business are almost always subtractive, and I would rather see a team guard the existing menu and workflow than chase a longer spec sheet, because what the customer bought in the first place was continuity. When a founder is going up against a platform like this, my first move is to take the head-to-head comparison off the table. The box was never the point. The opening is the work the installed base cannot do, and that is the only conversation worth having. From Dave’s video library Dave walks through five questions that surface why customers actually leave, a useful lens when the thing keeping them is habit rather than features. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [En Carta Signed Its Buyer Before It Built the Line →](/field-notes/2026-07-21-en-carta-signed-the-buyer-before-the-line) [In Orthopedic Robots, the Moat Is Never the Robot →](/field-notes/2026-07-15-orthopedic-robots-moat-is-not-the-robot) [Console-Free Cryocatheter: Cutting the Console Is the Product →](/field-notes/2026-07-14-console-free-cryocatheter-is-the-product) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Vicarious Surgical Raised $425 Million and Never Reached a Submission | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-07-26-vicarious-surgical-never-reached-a-submission [← All Field Notes](/field-notes) July 26, 2026 Surgical Robotics On Tuesday, Vicarious Surgical's shareholders voted to sign the company's assets over to a trust for its creditors. The company had twelve years behind it, more than $425 million raised from investors including Bill Gates, a $1.1 billion valuation at its 2021 SPAC merger, and the first breakthrough device designation the FDA ever granted a surgical robot. It planned to file its dissolution paperwork in Delaware as early as the next day. MedTech Dive reports the company never achieved regulatory authorization for the system, and was still hoping to file for ventral hernia repair as its first indication. I have watched a surgical robotics company from the inside, and I keep looking past the $425 million to the twelve years, and to what those twelve years were aimed at. ## What the designation actually bought Vicarious was the first surgical robot developer to win breakthrough device designation, and that was a real accomplishment. Keep in mind what it does, though. It buys earlier and more frequent access to the reviewers, plus a signal that the agency considers the clinical need worth prioritizing. It doesn't lower the evidence bar or shorten the engineering. You still have to write the submission. Founders let that designation stand in for progress, and boards let them. I've sat in those meetings. The designation goes on slide three, everyone nods, and nobody asks the follow-up, which is what date the filing goes in. ## Twelve years, and the design still wasn't frozen A year earlier, Vicarious canceled its clinical trial plans so it could finish the design of the commercial version of the system. Look at the order of those two things. About eleven years in, a trial was on the calendar before the design was done, and then the trial had to give way so the design could catch up. Design freeze is the gate that makes everything downstream mean anything. Founders outside regulated hardware tend to file it under paperwork. Verification and validation run against a specific design, so every time the design moves, your test evidence goes stale, your risk file goes stale, and you pay for that work a second time. A company that can't freeze is buying the same year over and over. I've seen the small version of this, where a founder tells the board they need two more weeks, and two weeks becomes two months, and two months becomes a year. Vicarious looks like the same pattern with a much bigger balance sheet underneath it, which mostly meant it could run longer before anything forced the issue. ## The money was never the constraint I led the De Novo effort at Galen Robotics, twelve FDA examiners on the call and me answering the questions. We were a rounding error next to Vicarious in capital, and we came out with an authorization and a regulation the agency adopted essentially as we had drafted it. I'm not claiming we were smarter, because we weren't. We were poor enough that the submission had to be the design target from early on. There was never a version of our plan where we could afford to build the whole robot and work out the filing afterward. That constraint turned out to be worth more than the money would have been. When the submission is the target, every design decision gets tested against a question with an actual answer: does this make the file easier or harder to defend? When the target is a finished product and the filing comes later, nothing arbitrates the scope, and a large balance sheet just lets the scope keep growing before anyone stops it. The timing this week is almost too neat. The same week Vicarious moved to dissolve, Johnson & Johnson picked up De Novo authorization for its Ottava system, which I [wrote about on Thursday](/field-notes/2026-07-23-ottava-cleared-fda-beating-da-vinci-is-different). That's not a fair fight and I won't pretend it is, since J&J can fund a decade of robotics without noticing. But one of those two calendars was organized around a submission date and one of them wasn't. (The 2021 SPAC that valued Vicarious at $1.1 billion is its own separate autopsy, and other people have written it better than I would.) Dave's take A breakthrough designation is a place in line, not a device, and Vicarious never converted its place in line into a filing the agency could act on. That gap has nothing to do with capital. I've stopped asking regulated-hardware founders how much runway is left, and started asking what date the submission goes in, and whether anyone on the team can say that date out loud without hedging. From Dave’s video library Dave walks through how to pick the smallest test of your riskiest assumption instead of building the whole product first. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [J&J's Ottava Robot Cleared FDA. Beating da Vinci Is a Different Problem →](/field-notes/2026-07-23-ottava-cleared-fda-beating-da-vinci-is-different) [Europe's Biggest Fusion Raise Buys Runway, Not Fusion →](/field-notes/2026-07-13-fusion-mega-raise-is-runway-not-a-moat) [A Disposable Surgical Robot Is a Manufacturing Bet →](/field-notes/2026-07-19-disposable-surgical-robot-is-a-manufacturing-bet) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Sartorius Got Its Tariff Money Back. It Handed the Refund to Customers. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-07-27-sartorius-gave-the-tariff-refund-to-customers [← All Field Notes](/field-notes) July 27, 2026 Commercialization Michael Grosse spent part of Sartorius's first-half earnings call explaining a decision most chief executives would have left in a footnote. The company had paid US tariffs, applied for refunds after those duties were declared not in line with existing law in February, and had started getting the money back. Instead of keeping it, Sartorius is compensating its customers for the surcharges they were charged. Analysts treated the resulting dent in reported revenue as noise and moved on. For Sartorius, it is noise. For a hard-tech startup hit by the same cost shock, it would not be, and the reason has almost nothing to do with tariffs. ## What the gesture actually cost The company applied for 40 million euros in refunds tied to US tariffs and had received 26 million of it by the end of June, with roughly 14 million still outstanding. Group sales grew 7.7 percent on an operational basis in constant currencies, while reported growth came in at 2.5 percent. Investing.com's coverage of the call put the tariff refunds and customer compensation among the things separating those two numbers, and the Americas were the softest region on a reported basis partly for the same reason. Grosse's position was that the underlying economics of the business had not changed. He is right. Underneath that gesture sits an underlying EBITDA margin above 30 percent, free cash flow up 70 percent to 208 million euros, and a consumables business growing more than 9 percent operationally, which means money keeps arriving whether or not anybody buys a new instrument this quarter. A company built like that can absorb a policy cost, hand the refund back to its customers, take the hit to reported growth, and correctly call the difference presentation. The generosity is real. It is also cheap, and it was made cheap by structural decisions taken years before this duty existed. ## The startup version of the same event A hardware company two years from first revenue does not experience a tariff as a margin question. It experiences it as a pricing question, and the price is usually already gone. Quotes go to a hospital system, a channel partner, or a defense prime long before the line runs. Numbers get committed in a letter of intent, a distributor agreement, a grant budget, a term sheet model. When the landed cost of the bill of materials moves after that, nothing stands behind it. There is no consumables annuity to soak up the difference and no 30 percent margin to bury it in. There is a number you promised and a cost you do not control. And the customer, having watched a supplier the size of Sartorius eat the difference, is not going to be receptive when you show up asking for a surcharge. I have been on the wrong side of this math. On a multi-field surgical robot I worked on, we deliberately stayed out of a head-to-head fight with da Vinci and went after an underserved area instead. The catch showed up in the reimbursement. Those procedures did not carry the big payments, so our economics had to come in lower just to function. There was no slack in the model for a cost surprise. Hospitals post-COVID do not fund science experiments, and the math has to close on day one or you never get into the first hospital at all. A cost shock arriving after you have already quoted is one of the faster ways to reopen that gap. ## Where the work goes If your bill of materials crosses a border, and for medical device, defense, and energy hardware founders it almost certainly does, then price is a policy variable and not only an engineering one. That means knowing the landed cost of every major component and its country of origin, and knowing what a duty change does to gross margin per unit before somebody asks you in a diligence session. It means quoting with the variable named: a validity window, escalation language, a defined pass-through for duties. Buyers in these categories negotiate that language routinely and will not be startled by it. Second-sourcing belongs across geographies and not only across vendors, because two suppliers inside the same country are effectively one supplier. Mostly it means being honest about your cushion. If the answer to "what happens if landed cost moves eight points" is a shrug, the problem is not procurement. It is that the product economics were never built to take a hit, which is a much earlier and much more expensive thing to go fix. Dave's take The tariff is the least interesting part of this. What made Sartorius's decision possible was a margin structure somebody built years before this duty existed, and that structure is what turned an expensive gesture into a cheap one. Founders tend to file unit economics under problems that volume will eventually solve. They are actually what decides which shocks you get to survive. From Dave’s video library In this one I go through the five decisions that quietly lock in the moment engineering starts, and why pricing is one of them. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [The Bearings, Not the Algorithm: Where Hardware Scale-Up Stalls →](/field-notes/2026-07-24-bearings-not-the-algorithm-hardware-scale-up) [For a New Implant, a Medicare Payment Code Is the Market Gate →](/field-notes/2026-07-17-medicare-payment-code-is-the-market-gate) [Europe’s Biggest Fusion Raise Buys Runway, Not Fusion →](/field-notes/2026-07-13-fusion-mega-raise-is-runway-not-a-moat) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # The Pentagon Rewrote the Drone Supply Chain. That’s a Product Requirement. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-07-28-pentagon-rewrote-the-drone-supply-chain [← All Field Notes](/field-notes) July 28, 2026 Defense Hardware Travis Metz runs the Pentagon's Drone Dominance Program, and on Monday he put numbers on a problem most people already assumed. Ukraine will build six to seven million small FPV attack drones this year, roughly 500,000 a month. The $1.1 billion American program Metz oversees will have ordered just under 200,000 drones, cumulatively, by February. Reuters reported the comparison, and most of the coverage stopped right there, at the gap. What Metz said around it matters more. He was straight about the phase he is in, saying it is much harder to get from zero to 200,000 than it will be to scale from there once domestic capacity exists. He is right about that. It is also worth noticing that four days before he said it, his own department made getting to 200,000 harder. ## The bill of materials changed four days before the interview A revised supply chain framework released July 23 added language stating that the government's goal is a wholly domestic small unmanned aircraft system supply chain. In practice that means none of the drones from the 19 companies competing in an August test event may contain banned parts like Chinese-made brushless motors or battery packs, a stricter standard than the program's first phase. Metz estimated that the vast majority of drones bought in the earlier phase probably had Chinese motors in them. That reads differently from an engineering desk than it does from a policy desk. Nineteen companies were told, weeks out from a test, that a component sitting at the center of their airframe is now disqualifying. A brushless motor is not a slot on a parts list. It is a torque curve, a mounting pattern, a thermal budget, an ESC firmware relationship, and a supplier somebody already spent months qualifying. Swap it and you reopen flight test, EMI, and vibration. That is not a purchasing task. ## Your buyer's sourcing rules are design inputs The failure mode I keep seeing in hard-tech companies is treating procurement rules as paperwork that happens after the product works. It is the wrong order. Sourcing restrictions and banned-vendor lists constrain a design about as hard as mass and power do, and they move on the buyer's schedule rather than yours. I have watched the identical shape play out in hospitals. A team designs a connected device around always-on cloud connectivity, and then the hospital's cybersecurity department, a gatekeeper most founders never put on their map of the buyer, says the thing has to work with no active internet connection. That department is not reviewing your paperwork. It is handing you an architecture requirement, and it can stall a deployment no matter how good the clinical data looks. Written into the requirements document at the start, that requirement costs almost nothing. Discovered after design freeze, it costs a year. ## What the joint ventures are actually buying The second half of the Reuters piece is the part I would put in front of any founder selling to a government buyer. All six Ukrainian companies invited to next month's Gauntlet II test at Fort Carson, Colorado have entered or are moving toward joint ventures with US manufacturers, as a condition of winning future orders. F-Drones partnered with Ohio-based Ukrainian Defense Drones, which opened a facility near Toledo. General Cherry agreed to a joint venture with New Hampshire's Wilcox Industries. Those six companies have been building and losing small drones under fire for years, and that experience still is not sufficient on its own. The buyer is not purchasing a design. It is purchasing a design, plus a domestic line, plus an ownership structure it finds acceptable. Access to the order depends on where the line sits and who owns it, and flight performance does not answer that question. Founders in energy hardware will recognize the shape. Domestic-content conditions attached to incentives do the same work there, deciding who is eligible before anyone gets around to judging whether the product is any good. Dave's take I have never seen a hardware company killed by a requirement it knew about. It gets killed by the one that shows up late and resets a clock it had already spent. So before you lock a motor, a cell chemistry, or a chipset, go find every rule your buyer can put on your bill of materials, write them down next to the performance specs, and treat them like physics. The ones you cannot find are the ones worth a phone call. From Dave’s video library Dave works through a government move against a technology and sorts the part that actually constrains a business from the part that is just noise. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [In Defense Hardware, the Constraint Is Production, Not Design →](/field-notes/2026-07-18-defense-hardware-constraint-is-production-not-design) [New Robot Safety Standards Are a Market-Access Gate, Not Paperwork →](/field-notes/2026-07-12-robot-safety-standards-market-access-gate) [The Bearings, Not the Algorithm: Where Hardware Scale-Up Stalls →](/field-notes/2026-07-24-bearings-not-the-algorithm-hardware-scale-up) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # How the Indication Sentence Decides Your 510(k) | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-07-29-indication-sentence-decides-your-510k [← All Field Notes](/field-notes) July 29, 2026 Regulatory NeuraNova cleared a 510(k) this month for a device that reads fluid asymmetry inside the skull. The hardware is bioimpedance spectroscopy and it's genuinely clever. The part of that submission worth studying, though, is one sentence of plain English. ## What the indication sentence actually does A 510(k) runs two equivalence tests and they run in order. First the indication: is what you say your device is for substantially equivalent to what the predicate is for? Then technological characteristics, which is looser and depends on how different your engineering really is. Most founders pour their energy into the second test. The first one is where submissions die. NeuraNova's cleared wording says the device aids "in the assessment of" fluid volume asymmetry between the cerebral hemispheres, in adult patients who are already undergoing a neurological assessment. The company describes it as "an adjunct to standard clinical evaluation" methodologies. Now count what that sentence declines to do. It doesn't diagnose, and it doesn't use the word stroke anywhere. It aids an assessment a clinician is already performing, and it sits beside the existing standard instead of replacing it. Those are boundaries the company drew on purpose, and every boundary it drew shrank the pile of evidence it had to produce. The submission behind the clearance was engineering validation, repeatability testing, software verification and validation, and risk management review. No clinical outcomes trial. That isn't a loophole. That's what an evidence base looks like when the claim is scoped to what a bench can actually prove. ## Why "better" is the expensive word The trap here is the marketing instinct. A founder who has spent four years on a sensor wants the indication to say the device finds the bleed faster than what's in the room today. The moment you write faster, or more accurate, or better, you've made a superiority claim. Superiority is definitionally not equivalence, and asserting it can push you out of the 510(k) lane and into a review nobody budgeted for. The disciplined version argues that your differences are safely evaluated inside the same bounds as the predicate. Not better. Equivalent, done a little differently, with an engineering reason the difference raises no new question of safety or effectiveness. One thing I noticed: NeuraNova used to be called StrokeDx. The cleared sentence doesn't mention stroke. I have no idea whether the rename and the indication language were the same decision and I'm not going to pretend I do. But the gap between that old company name and that cleared claim is roughly the gap most hard-tech founders have to close before a submission is ready. ## Whose job that sentence is This is where I'll argue with how most companies staff the work. The indication gets filed mentally under regulatory paperwork, so it goes to the regulatory consultant, who writes it late, after the engineering is frozen. That's backwards. Deciding what the product is, in one sentence a regulator will accept, is product strategy, and it belongs to whoever owns binding the product to a market. I learned that the expensive way. At Galen we first framed the robot as substantially equivalent to a remote-control telemanipulator, with da Vinci as the predicate. The FDA didn't buy it. We reframed the same hardware, physically unchanged, as a device that improves a surgeon's access at the end of a long instrument, and that framing went through as a De Novo. We ended up drafting the regulation ourselves and the agency took our language almost word for word. The machine never changed. The sentence did. Change that sentence after the design history file is written and you get to reconcile every document sitting behind it. Nobody budgets for that either. Dave's take The founders I watch get blindsided by the FDA are almost never blindsided by the engineering. They're blindsided because a claim somebody made casually in a pitch deck two years ago is now a claim they have to defend with data. Write the indication sentence early, say it out loud with your engineering lead and your regulatory advisor in the same room, and make everything downstream agree with it. It's the cheapest work in the program and it's the piece most teams leave for last. From Dave’s video library In this one I walk through the three jobs a founder keeps on purpose, and Direction is the one that covers calls like this. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [FDA Days Are Not Calendar Days →](/field-notes/2026-07-16-fda-days-are-not-calendar-days) [J&J's Ottava Robot Cleared FDA. Beating da Vinci Is a Different Problem →](/field-notes/2026-07-23-ottava-cleared-fda-beating-da-vinci-is-different) [The Pulse Oximeter Worked. The Validation Didn't. →](/field-notes/2026-07-22-pulse-oximeter-validation-not-the-device) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # ENvue’s Feeding Tube Robot Came Second. That’s the Product Decision. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-07-30-envue-feeding-tube-robot-came-second [← All Field Notes](/field-notes) July 30, 2026 Medical Devices A feeding tube that ends up in a lung instead of a stomach is one of the quieter ways a hospital hurts someone. By ENvue Medical's own accounting it happens in three to five percent of the more than one million placements done in the United States every year. ENvue has been selling into that problem for years. This month it showed a robot, and the robot is the least interesting part of the announcement. What the company unveiled, per The Robot Report on July 28, is ENvue Drive: electromagnetic navigation, AI recognition to flag anomalies, and a robotic drive that advances the tube, reading its sensor forty times a second across five degrees of freedom. CEO Doron Besser described the goal as a procedure that's safer for patients, easier for clinicians, and less expensive for hospitals. It's a prototype. The company says it plans to pursue clearances for it over the coming years. The part worth a founder's attention is the order ENvue built in. ## They cleared the unglamorous version first The ENvue Navigation Platform, the electromagnetic guidance system the robot rides on, is already 510(k) cleared for adult use and already running in US hospitals. The company also has numbers to talk about. In a June release it cited a peer-reviewed series of 531 consecutive placements with no lung placements, a 67 percent reduction in ventilator-associated pneumonia, roughly $1.5 million in annual cost avoidance inside a single health system, and more than 350 nursing hours freed per year. Those are the vendor's own figures, so read them as a best case. They're still the useful kind of number, because each one converts a clinical outcome into money a hospital already spends. So by the time ENvue bolts a robot onto that platform, the expensive questions have mostly been answered before the engineering starts. The company owns its own predicate, which means the regulatory path is a known quantity rather than a guess. There's a clinical series behind the claim. And the hospitals that would buy the robot have already trained their staff on the workflow it plugs into, which is usually the slowest thing to change in a hospital and the thing founders budget the least time for. Compare that with the more common sequence, where the robot is the fundraising story and the evidence stays in the future tense. I wrote about [Vicarious Surgical](/field-notes/2026-07-26-vicarious-surgical-never-reached-a-submission) on Sunday: $425 million raised, no submission ever filed. Different category, same lesson about what you're supposed to prove first. ## Run it through the four gates I keep a short triage for medical robot ideas, and I don't get interested unless all four gates clear. Does it enhance a tool or workflow clinicians already use, instead of asking them to relearn the procedure? Is there filed IP on the actual embodiment rather than on the concept? Can you explain it to the person who'll use it in one sentence? And is a credible key opinion leader already enthusiastic? ENvue clears the first gate cleanly. Nothing about the procedure changes. It's the same tube going to the same place, with the hard part of the navigation handled. Gate three is fine too, because "the robot drives the tube where the navigation says it belongs" is something a bedside nurse can picture without a whitepaper. The company describes developing its sensor technology in house, which gestures at gate two without settling it publicly. Gate four is the one the announcement leaves open, and it's a different problem here than it would be in an operating room, because this procedure belongs to nursing and critical care rather than to a surgeon with a following. ## The word doing the work is "automatic" Most of the coverage calls this automatic placement. That's the word I'd want tightened long before a submission. My position on interventional robots is that you enhance the clinician and you do not automate them: make the hard part disappear under their own hands, and leave them in the driver's seat. At Galen we built the robot smaller than the assistant it replaced so the surgeon never had to leave their normal working position. Enhancement was a physical decision, not a slide. A tube advancing itself toward a patient's stomach at a bedside, in a room that may not have a physician standing in it, is precisely where "what is the clinician doing during this, and how do they stop it" turns from an engineering detail into the indication. That question has an answer. Whatever the answer is, it decides how much clinical evidence the agency asks for, and that decides how long the runway has to be. Dave's take ENvue is running the sequence I argue for and almost nobody actually runs. Clear the simpler device, publish the outcomes, get the workflow installed, then build the ambitious thing on a predicate you own. The robot isn't the bet here. The bet was placed years ago on a navigation box, and the robot is what you get to build once the unglamorous work is already earning. From Dave’s video library Dave walks through why a stated yes is not evidence, and which kinds of proof actually survive contact with a real buyer. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Vicarious Surgical Raised $425 Million and Never Reached a Submission →](/field-notes/2026-07-26-vicarious-surgical-never-reached-a-submission) [How the Indication Sentence Decides Your 510(k) →](/field-notes/2026-07-29-indication-sentence-decides-your-510k) [The Pulse Oximeter Worked. The Validation Didn't. →](/field-notes/2026-07-22-pulse-oximeter-validation-not-the-device) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Momentis Built Its Own Predicate, Then Spent It Three Times | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-07-31-momentis-built-its-own-predicate [← All Field Notes](/field-notes) July 31, 2026 Surgical Robotics Momentis Surgical said on July 23 that the FDA cleared a multiport configuration of its Anovo system, and the company can now claim something no other surgical robot can: natural orifice, single-port, and multiport access on one platform. The capability is the headline. What interests me is the regulatory decision that made it possible, which Momentis made years earlier, back when it looked like the slower option. Anovo's original authorization was a De Novo for a robotic transvaginal surgical system. No predicate existed, because the category didn't. Momentis had to define what the device was, name what could go wrong with it, and decide what testing would prove it safe enough. That is more work than pointing at somebody else's cleared robot and arguing you're close enough to ride along. What that bought them shows up in every filing since. Each one has been a 510(k) against a predicate they own. Single-site transabdominal ventral hernia repair. Robotic camera control. Now the multiport configuration, which puts gynecologic procedures and ventral hernia repair on the same core platform across different access approaches. Three expansions, each one a routine submission, because the hard categorical argument got settled once. ## I lost this argument from the other side At Galen we were building a cooperative-control robot for skull-base and laryngeal surgery. Cooperative control means the surgeon is still driving. It's a joystick, not a remote. We thought we had a clean 510(k), and we shared some software lineage with the da Vinci, since Russ Taylor wrote the original code. The predicate logic looked airtight from where we sat. The FDA didn't buy it. Our indication didn't match, they said, and a Delta robot isn't a remote-control telemanipulator. We pushed back and they said no again. Then came the sentence you have to learn to hear: it would be advisable that we consider the De Novo path. That's not a suggestion. That's the answer. So we took it, and the work turned into framing. We stopped arguing that we were a da Vinci and anchored the indication in what was physically happening, which was stabilizing an instrument the surgeon was already holding at the end of a 40-centimeter shaft. We drafted the proposed regulation ourselves. The FDA took our language essentially word for word, and Galen became the predicate. The sequencing is the part that matters. We got there by losing first. ## Your first clearance is inventory Founders treat the first clearance as a finish line, and the whole company organizes around clearing anything at all as fast as it can. That's the wrong thing to optimize. The first clearance is inventory you'll spend for the next decade, and its value sits almost entirely in what it lets you file next. A De Novo looks expensive on the first pass and cheap on the fourth. You write the safety story instead of guessing at somebody else's, because you don't have their submission and never will. The regulation that comes out the other side describes your device, which means later entrants have to be substantially equivalent to you, and your own next feature becomes a clean filing against yourself. A borrowed predicate gives you none of that. It gives you a faster first clearance and a ceiling. The other thing is where Momentis started clinically. It didn't open by trying to be a cheaper da Vinci. It opened in transvaginal procedures, a place the incumbent console-and-cart geometry serves poorly, and worked outward from there. Those two decisions aren't separate ones. Picking a procedure the existing category handled badly is what put a new device classification on the table at all, and walking into the agency asking permission to be somebody else would have closed that door. One caution before anyone reads this as a template. A De Novo is only on the table when the agency already considers your risk moderate and manageable. If they think you're high risk you're in a PMA, and no amount of clever framing rescues that. The pathway isn't something you pick in a conference room. It's a read you take off the signals the agency gives you, and reading those signals correctly is most of the job. Dave's take I've watched founders deflate when a pre-sub routes them toward De Novo, like they've been handed a punishment. It's closer to a compliment. The agency is telling you it thinks your risk is manageable and it's willing to let you define the controls, which is about the most leverage a regulator will ever hand you. Momentis is three clearances into cashing that in. Take the hint when it comes. From Dave’s video library In this one I make the case that a company built to serve a defined market well is not a consolation prize. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [How the Indication Sentence Decides Your 510(k) →](/field-notes/2026-07-29-indication-sentence-decides-your-510k) [J&J's Ottava Robot Cleared FDA. Beating da Vinci Is a Different Problem →](/field-notes/2026-07-23-ottava-cleared-fda-beating-da-vinci-is-different) [Vicarious Surgical Raised $425 Million and Never Reached a Submission →](/field-notes/2026-07-26-vicarious-surgical-never-reached-a-submission) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Peak Energy Picked the Weaker Chemistry. Then It Deleted the Cooling System. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-08-01-peak-energy-deleted-the-cooling-system [← All Field Notes](/field-notes) August 1, 2026 Climate Hardware Peak Energy co-founder Cameron Dales told Latitude Media that the company started with the market, not with the technology. In practice that meant choosing a battery chemistry that loses to lithium on the number the whole industry tracks, and it is why Peak is now building a 4 GWh factory outside Sacramento with General Motors behind the cells. The easy way to read sodium-ion right now is as the cheap substitute: abundant materials, no lithium exposure, a hedge against Chinese supply. All of that is true and none of it is the interesting part. Peak’s cost advantage does not come from a cheaper cell. It comes from what the cell lets the company leave out. ## The number they agreed to lose on Sodium-ion stores less energy per kilogram than lithium iron phosphate. In a car that is close to disqualifying. In a steel box sitting in a field for twenty years, mass and volume are nearly free, and the buyer is optimizing something else entirely. IEEE Spectrum quoted Dales making the trade plain: sodium-ion goes lower on energy density and pays for that with better stability and safety at lower cost. Peak’s cells use a sodium iron pyrophosphate cathode and, according to IEEE Spectrum, run safely at roughly double the operating temperature of lithium iron phosphate. That one property is what lets the rest of the design go away. The system cools passively, so there are no fluid loops to maintain and no fans or pumps to fail. Dales described the conventional alternative as having to “power a refrigerator in the desert for 20 years.” Peak’s own announcement claims that removing battery refrigeration could save California ratepayers around $100 million a year. ## What the trade actually buys The numbers Peak leads with are lifetime numbers rather than spec-sheet numbers. IEEE Spectrum reports its GS1.1 system holding 80 percent of capacity after about 20,000 cycles across 20 years, against a lithium iron phosphate baseline of 70 percent after 8,000. Round-trip efficiency comes in at 96 percent, two to three points better. The company claims 20 percent lower lifetime cost, 99 percent guaranteed uptime, and no scheduled maintenance. Buyers are signing. Peak reports more than 6 GWh of customer commitments, including a Jupiter Power agreement worth up to $500 million for as much as 4.75 GWh through 2030, with a first 720 MWh block in Texas, a 1.5 GWh data center deal with Energy Vault signed in February, and an RWE Americas pilot near Milwaukee. The Sacramento plant runs to $71 million and 183,000 square feet, with 239 jobs and first shipments in the first quarter of 2027. I watched this same move pay off once, on a much smaller stage. We sold a terminal emulator to government contractors, and the spec for one terminal type ran 1,800 pages. Building it properly would have taken years. Then we found out what the customers were actually doing with the thing: displaying CAD files to verify parts. That was the whole job. The project collapsed into six months of coding and still took millions of dollars across tens of thousands of seats. Peak made the same call one level down. When you find the job the customer is actually paying for, a great deal of engineering stops being necessary. ## Where the bet is still exposed Two things are unfinished. Peak is currently sourcing cells from Chinese suppliers while it builds a domestic factory, which means the supply-chain argument for sodium-ion is a promise the company has not yet kept for itself. The GM cells are also not the cells shipping now. GM plans to start prototyping at its Wallace Battery Cell Innovation Center at the end of this year, in 170 and 190 amp-hour formats, with commercialization targeted for 2028. Policy is moving the same way regardless. Beijing said in July that lithium-ion batteries lose an exemption they have held since 2015 and take a 2 percent consumption tax from September 1, rising to 4 percent in September 2027. Sodium-ion, solid-state, and fuel cells stay exempt through the end of 2028. No founder can plan around Chinese tax policy, but it does say the chemistry Peak picked is not a fringe position. Dave’s take The failure mode I see most often is a team that fell in love with its product instead of its customer, and the tell is a roadmap built to win the metric the trade press ranks. Peak went the other way and let itself be beaten on energy density, because the people writing the checks do not buy energy density. If you are building hardware right now, name the spec you are allowed to lose on. If you cannot name one, you have not defined the product yet. You have only agreed to compete on everything. From Dave’s video library Dave walks through how to make a real call when the data you want does not exist, using rough estimates that are good enough to decide on. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Europe’s Biggest Fusion Raise Buys Runway, Not Fusion →](/field-notes/2026-07-13-fusion-mega-raise-is-runway-not-a-moat) [The Bearings, Not the Algorithm: Where Hardware Scale-Up Stalls →](/field-notes/2026-07-24-bearings-not-the-algorithm-hardware-scale-up) [Cepheid’s New GeneXpert Kept the Old Menu. That’s the Product. →](/field-notes/2026-07-25-cepheid-genexpert-kept-the-menu) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # MMI Didn't Build a Robot for Alzheimer's. Two Surgeons Decided It Was One Anyway | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-08-02-mmi-symani-alzheimers-who-defines-your-market [← All Field Notes](/field-notes) August 2, 2026 Surgical Robotics On July 30 the FDA signed off on a supplement to an investigational device exemption, and the practical effect is that a study in Jacksonville can now enroll ten more patients than it could the week before. That is a small regulatory event. It is also a good look at something founders building platform tools tend to assume, which is that they get to decide what their product is for. They mostly don't. ## What the agency actually approved MMI, short for Medical Microinstruments, makes the Symani Surgical System, a robot built for microsurgery on structures under a millimeter across. It scales the surgeon's hand motion down and filters out tremor. The REMIND study, short for Robotic-Enabled Microsurgical Intervention for Neurodegenerative Disease, uses it to reopen lymphatic drainage in the deep cervical lymph nodes of people with mild to moderate Alzheimer's and confirmed lymphatic abnormalities. The theory, and it is still a theory, is that better drainage clears the neurotoxins researchers believe drive the disease. The FDA reviewed 30-day safety data from the first patients and approved a supplement to the IDE, per MMI's July 30 release. Enrollment goes from an initial cohort of up to five to the full early feasibility population of fifteen, each followed for twelve months. The first patient was treated at Baptist Health in Jacksonville back in May. Look at what actually cleared the gate here. Thirty days of safety data on a handful of people bought the expansion to fifteen. MMI's own release says the safety and effectiveness of Symani for lymphatic reconstruction in brain disease has not been established by any regulatory agency, and that sentence is doing real work rather than sitting there as legal decoration. Nobody has shown this helps anyone yet. ## The robot didn't change. The question did. Symani got a De Novo in 2024 covering microsurgery, then a 510(k) in 2025 that added soft tissue dissection, per MedTech Dive. Neither of those says anything about Alzheimer's. Neither anticipated a neurointerventionalist and a plastic surgeon standing over a patient's neck. That's the whole thing. The machine that showed up in Jacksonville is the same machine MMI has been selling to reconstructive surgeons for free flaps and lymphovenous anastomosis. What changed is that somebody with a hypothesis about lymphatic drainage needed to sew a sub-millimeter vessel to a vein, reliably, in a space with no room to work, and there was exactly one commercially available instrument that could do it. Which gets at something I believe about robots generally. There is no one robot to rule them all, and the machines that win bounded physical tasks are engineered around the task rather than around a general-purpose ambition. Symani is about as narrow as a surgical robot gets, and that narrowness is why it got picked up. A general-purpose arm would never have been in the conversation. ## Who actually defines your market The commercial math is not subtle. Reconstructive and lymphatic microsurgery is a real business and a narrow one. Alzheimer's is not narrow. If REMIND goes anywhere, and that is a real if at fifteen patients, MMI's addressable market changes by more than any sales hire could deliver. MMI did not invent that possibility. Two clinicians did. That is the pattern I look for in a platform tool. The way you find real demand in surgery is to stand in the OR, watch the procedure, and ask what the hardest part is. When the surgeon says this part is awkward and I just work around it, that is your opening, because every surgeon trained the same way hits the same wall. What MMI got here is that same discovery at a distance. Somebody found a problem only their tool could reach, and picked up the phone. So the question for a founder is not which market do I want. It is whether you are cleared, available and reliable enough to be picked when somebody else finds the problem. A company sitting on a research prototype with no clearance would not have been reachable at all. The trap on the other side is real, and worth a sentence before I leave it alone. Building a deliberately general tool and waiting for somebody to invent a use for it is not what happened here. Symani was built to do one thing extremely well, and the pull came because of the narrowness, not in spite of it. Dave's take I would not read this as MMI buying a lottery ticket on Alzheimer's, and I would not read fifteen patients as proof of anything. What it shows is that the tool was narrow enough and cleared enough to be handed a question its builders never asked. What I would actually watch from here is whether MMI can keep saying no to the story with the bigger market while the reconstructive surgeons paying the bills today are still waiting on their roadmap. From Dave’s video library In this one I go through why founders stop doing the unscalable customer work right after their tenth customer, and what it costs them. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [ENvue's Feeding Tube Robot Came Second. That's the Product Decision. →](/field-notes/2026-07-30-envue-feeding-tube-robot-came-second) [How the Indication Sentence Decides Your 510(k) →](/field-notes/2026-07-29-indication-sentence-decides-your-510k) [Orthopedic Robots: The Moat Is Not the Robot →](/field-notes/2026-07-15-orthopedic-robots-moat-is-not-the-robot) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Walter Reed Cleared the Implant. Now the Hospital Is the Manufacturer. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-08-03-walter-reed-hospital-is-now-the-manufacturer [← All Field Notes](/field-notes) August 3, 2026 Medical Devices Someone at Walter Reed decided the hospital would hold the clearance itself. Not the vendor, not a contract manufacturer. On July 28, 3D Systems announced that the 3D Medical Applications Center at Walter Reed National Military Medical Center had won FDA clearance for the 3D MAC Titanium Cranial Plate System under K253116, the first cleared implant ever held by a point-of-care institution. The plate is a good piece of engineering. The bigger news is that a hospital is now the manufacturer of record on an implant. The device is a patient-specific titanium plate that reconstructs cranial and facial defects after trauma, disease, or cancer, and it will be used on active duty personnel with traumatic head injuries. Dr. Peter Liacouras, who directs services at 3D MAC, designed it. They cleared it on their first 510(k) submission. ## What the clearance holder actually signs up for Holding a 510(k) means being the manufacturer of record, and that title carries a long list of standing obligations. Design controls. Complaint handling. Corrective and preventive action. Adverse event reporting. Supplier controls. Process and sterilization validation. A quality system that has to survive an FDA inspection, operated by an institution whose actual job is treating patients. Then layer on the fact that these implants are patient-specific, so the design record is not written once and filed away. It regenerates with every skull. 3D Systems did real work here. The company supplied process control expertise, design optimization, additive manufacturing workflow guidance, regulatory strategy, and help standing up the quality system, and its CEO Jeff Graves framed point-of-care manufacturing as bringing personalized devices closer to the patient. But 3D Systems did not take the clearance. Walter Reed did. That matters, because the entity named on the clearance is the one that has to live inside the quality system for as long as the device exists. ## I once paid about $480,000 for blue ink At Galen, our bill with Hogan Lovells on the De Novo came to something like $480,000. My board insisted on the brand-name firm, end to end. Meanwhile our own internal people were writing the documentation. The regulatory work product was being produced inside the company, by the engineers and clinicians who understood the device. What the board was really buying was the signature on it, because a recognizable firm changes how the next investor reads the file. That has value and I want to be fair about it. The submission cleared eleven days after we filed, so the firm probably did buy us some certainty. But I was paying premium hourly rates for regulatory thinking that belonged with a specialist consultant at a fraction of the price, on top of the engineering and clinical work that produced the document in the first place. I got disciplined about it eventually, logging every billable call and reading the invoices line by line. I told them I was not paying for ten minutes of pleasantries at the top of every call. The pattern I have run ever since, across IP portfolios, stock option plans, cadaver studies, and FDA work, is to learn enough to do about ninety percent of it myself, pay the expensive experts to tell me what to go figure out, then hand it back for the final refinement. That is roughly the shape of what Walter Reed did. They kept the design, the manufacturing, and the ownership. They bought the expertise they did not have. ## What the clearance actually covers Walter Reed intends to distribute the system to other point-of-care facilities across the Department of Defense, and possibly into the VA. The plan includes design hubs, where other military institutions submit device files and Walter Reed manufactures them. Scans, design, and production all happen in house. That turns the clearance into a distribution plan. A clearance written around a process rather than one finished good lets an institution extend the capability to its peers, and it puts Walter Reed in the position every other DoD facility now has to route through. The version of this that applies to a founder is simpler. If your product only works when it is built or fitted close to the patient, you have already decided that somebody out near the patient runs a quality system. That is not a manufacturing detail you get to settle after the design freeze. It sets who your customer is, what you can charge, and who carries the regulatory burden when something goes wrong. Dave's take Most founders I talk to treat the quality system as overhead, something a consultant produces and hands over in a binder a month before the audit. Walter Reed spent years doing the opposite and came out holding a clearance its peers will now route through. If your product has to be built close to the customer, decide on purpose who runs that quality system, because whoever runs it ends up owning the product. From Dave’s video library Dave on how to tell which work a founder should still be doing personally, and what it costs to hold on past that point. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [How the Indication Sentence Decides Your 510(k) →](/field-notes/2026-07-29-indication-sentence-decides-your-510k) [Momentis Built Its Own Predicate, Then Spent It Three Times →](/field-notes/2026-07-31-momentis-built-its-own-predicate) [The Pulse Oximeter Worked. The Validation Didn't. →](/field-notes/2026-07-22-pulse-oximeter-validation-not-the-device) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Reimagine Wants Your Line Worker to Teach the Robot. Walden Wants to Do It for You. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-08-04-who-teaches-the-robot-reimagine-vs-walden [← All Field Notes](/field-notes) August 4, 2026 Industrial Robotics Jonathan Scholz spent seven years running the applied robotics team he founded at Google DeepMind in London. The company he started after that came out of stealth on Monday, and its pitch has almost nothing to do with model size. Reimagine Robotics wants the person who already does the job to teach the robot by showing it. Scholz calls the approach monkey-see, monkey-do. That sounds like an AI claim. It's a product decision, and it's the one robotics founders are most likely to make by accident. According to The Robot Report, in one deployment at an electronics disassembly facility, where a three-robot cell pulls valuable material out of used hard drives, Reimagine says prototyping and testing a new robot behavior went from about a day to about ten minutes. A worker demonstrates the task, watches the robot try it, and corrects it while it runs. No specialist programmer in the loop. Another customer, a made-to-order plastics shop, has the robots tending 3D printers overnight: removing beds, working latches, pressing controls. Scholz founded the company in April 2025 with three former DeepMind colleagues, and raised a pre-seed round from Fly Ventures and firstminute capital. The amount has not been disclosed. ## Two answers to the same question Three weeks earlier, Walden Robotics launched at a $1.1 billion valuation with $300 million behind it. Russ Tedrake, the MIT professor who ran large behavior models at Toyota Research Institute, is co-founder and CEO. Toyota led the round. NVIDIA and Boeing are in it. Walden's semi-humanoid robots, a torso with two arms on a wheeled base, have been working inside a Toyota plant in North America since February, moving from pilot to production manufacturing and logistics tasks in two months. Tedrake says the company is building the "full stack": hardware, software, AI, and the application layer that sits on top. Both companies name machine tending as a target task. Both are selling robots that learn on the job. What separates them is who does the teaching once the machine is standing on somebody else's floor. Walden is building that layer itself. Reimagine is handing it to the customer's own operator. I don't know which one wins. I do know most founders never write the decision down. ## The spec is not the job Years ago I sold a terminal emulator to government contractors. One of the terminal types we had to support was the 3179G, and the spec on it ran 1,800 pages. Implementing the whole thing would have taken years, and we were planning the work like it would. Then we found out what the customers were doing with those terminals. They wanted to display CAD files so they could verify parts. That was the job. It was that simple. The work collapsed into a six-month coding project, and we still sold tens of thousands of seats for millions of dollars. The 1,800 pages described everything the terminal could do. The customer needed one thing it did, and nobody on my team could have found that gap by reading the spec, because it only existed in the heads of the people doing the work. That gap sits inside every general-purpose robot on the market. Machine tending is not one task. It is a particular latch on a particular machine in a particular cell, plus a knack the operator picked up in their second week and has never once said out loud. ## What it costs you to decide late If your robot needs your engineers to teach it every new behavior, you have built a services business with a hardware invoice attached. Every site is a deployment. Every new part number is a ticket. Your margin becomes a function of how many of your people are standing in someone else's factory. That business can work. Walden raised $300 million, which buys a lot of engineers in factories, and having Toyota as both lead investor and first plant makes the early deployments unusually friendly. If the operator can teach it, you're selling a tool, and your cost to serve drops. You inherit a harder design problem in exchange. The teaching interface has to work for someone who will never read the manual, and a robot that learns from a human learns the bad habits with the same fidelity as the good ones. Neither answer is free. Both of these teams picked one on purpose, which already puts them ahead of most of the hardware companies I see. Dave's take I've watched teams spend a year building what the spec described and about a week finding out what the customer needed. In robotics that mistake hides inside the training interface, because founders file it under AI and hand it to the research team when it belongs to product definition. Decide who owns the behavior after the sale and you've decided what business you're in. From Dave’s video library Dave walks through the NBER study of an AI copilot dropped into a 5,179-agent support org, where the newest reps gained the most and the veterans gained almost nothing. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Teleoperation Is a Trap Only If the Human Is There by Accident →](/field-notes/2026-07-20-teleoperation-trap-human-by-accident) [The Bearings, Not the Algorithm: Where Hardware Scale-Up Stalls →](/field-notes/2026-07-24-bearings-not-the-algorithm-hardware-scale-up) [ENvue's Feeding Tube Robot Came Second. That's the Product Decision. →](/field-notes/2026-07-30-envue-feeding-tube-robot-came-second) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # The Pentagon Wants Four Times the THAADs. The Shopping List Says Muzzle Covers. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-08-05-pentagon-interceptor-surge-buys-muzzle-covers [← All Field Notes](/field-notes) August 5, 2026 Defense Hardware On August 3 the Pentagon said it wants to triple production of the PAC-3 interceptor and quadruple production of THAAD. The framework agreements sitting behind that announcement buy solid rocket motors, ignition safety devices, THAAD mid-body shells, muzzle covers, and rail car assemblies. No new missile anywhere in it. If you build hardware and you read this as the defense market opening up, I think you have it backwards. ## What the Pentagon actually bought The money is real. Defense News reports a long-term procurement objective of nearly 14,000 Patriot and THAAD interceptors, and Stars and Stripes puts an Army contract with Lockheed Martin at $58 billion over seven years for Patriot production. The urgency is real too. CSIS estimates the United States burned through roughly two-thirds of its pre-war Patriot inventory during the Iran conflict. But look at where the new agreements land. Northrop Grumman becomes a second source for Patriot solid rocket motors. Production expands for ignition safety devices on the PAC-3. Capacity goes up for THAAD structural parts. Michael Duffey, the under secretary of defense for acquisition and sustainment, described the goal as "robust, dynamic supply chains at every level of the industrial base." Read that as a sentence about redundancy. The Pentagon is not shopping for a better interceptor. It is shopping for a second company that can build the one it already has. ## The thing that gets qualified is your process So here is what a second-source award asks a supplier to prove. That unit four hundred is the same as unit one. That the process making it is documented, validated, and frozen, so a change three tiers down the supply chain does not quietly move a dimension. That the records still hold up in an audit years after the engineer who set up the line has moved on. None of that is a performance question. A founder who walks into that conversation with a cheaper seeker or a smarter guidance stack is answering something nobody asked on August 3. I spent years on the medtech side building this discipline into design controls, and it travels. Making the first one work is the part founders celebrate. Proving the four hundredth is identical, on paper, to somebody who was never in the building is the part that wins the contract. ## The ramp sits on your balance sheet Now the part that worries me. This surge is priced off a five-month war. Stars and Stripes reports estimates that the United States consumed about two-thirds of its Patriot stockpile and half its THAAD missiles in that window. Build a line against that consumption rate and you are betting the rate holds. I watched that bet go the other way once. I was in Australia as a guest of Telstra for the opening of Sydney Olympic Park, standing on an ordinary tourist walk, when the guide announced over the microphone that Telstra was going to skip OC-96 fiber and go straight to OC-192. Lucent had been betting on OC-96. I took it back to the group president and was told not to involve myself in technology outside my department. About a year later Lucent began to come apart, in part because of all the OC-96 hardware sitting in the warehouses. The signal was free. It was announced to tourists over a PA system, and it still could not reach anyone able to act on it. If you take a second-source award and stand up a line to serve it, that line belongs to you when the consumption rate returns to peacetime. Negotiate the ramp so the tooling, the floor space, and the headcount are not stranded on your books the quarter the shooting stops. Dave's take The Pentagon is buying reliability this month, and reliability gets bought from people who can document it. So if you are a hard-tech founder eyeing defense revenue, ask whether you can be qualified at volume by an auditor who has never met you and never will. That capability is boring to build, it wins nothing at a demo day, and it is the one getting funded right now. From Dave’s video library Dave walks through how to make a call like a capacity bet when the data you want does not exist yet. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [The Pentagon Rewrote the Drone Supply Chain. That's a Product Requirement. →](/field-notes/2026-07-28-pentagon-rewrote-the-drone-supply-chain) [In Defense Hardware, the Constraint Is Production, Not Design →](/field-notes/2026-07-18-defense-hardware-constraint-is-production-not-design) [The Bearings, Not the Algorithm: Where Hardware Scale-Up Stalls →](/field-notes/2026-07-24-bearings-not-the-algorithm-hardware-scale-up) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # BD's Needle Was Out of Tolerance for Twenty Months. Part of the Fix Is a Label. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-08-06-bd-needle-out-of-tolerance-fix-is-a-label [← All Field Notes](/field-notes) August 6, 2026 Medical Devices Intraosseous access is what a paramedic reaches for when the veins have collapsed and there are seconds left. You drill a needle into the bone, pull out the stylet that kept the tip from clogging, and run fluid straight into the marrow. For a run of BD needle sets, the stylet wouldn't come out. On July 31 BD recalled five catalog numbers of its Intraosseous Vascular Access System Needle Sets. The affected lots shipped between September 30, 2024 and June 12, 2026. Through June, BD had logged 75 complaints and 45 serious injury reports, along with four deaths, all four during out-of-hospital cardiac arrest resuscitations where the obturator would not come out. FDA posted an early alert. MedTech Dive wrote it up on August 5. BD gives the cause in one line. The affected lots were manufactured with out of tolerance dimensions. That reads like a factory problem, and treating it as one is how a company earns the right to repeat it. ## The tolerance held everywhere except a real resuscitation FDA is specific about when the parts lock together. The obturator can seize if the drill isn't pulled straight back during insertion, or if the obturator gets rotated on the way out. Neither of those is abuse. Both are what hands do when the patient is on the floor and the clock is the only thing in the room anybody is thinking about. So the dimension was in spec for a clean axial withdrawal by a steady operator. That describes a bench, not the device's job. I ran a shipping test on a surgical robot once. The company doing the test dropped the container off a raised forklift, and I'm not naming names. The robot didn't bang around inside at all. But there was enough force that one of the actuators tore itself off its mounts. I never would have thought about internal kinetic energy issues. We rebuilt the container, the roll-in and roll-out process, and the entire internal securing approach off that one drop. There's video. It was terrifying to watch. Nothing in our analysis predicted that failure, because the analysis assumed the payload was the thing at risk. The payload was held perfectly still. That was the problem. A tolerance stack behaves the same way. Every part passes on its own, and the failure lives in what happens between two of them under conditions nobody wrote down. ## Seventy-five complaints arrived before the recall did BD's own count is 75 complaints and 45 serious injury reports through June. Those did not land in a single week. They accumulated across a shipping window of roughly twenty months, and each one arrived at a company with a functioning complaint system, because BD has one and it is audited. The mechanism is never the interesting part. What differs between device companies is how many reports of a stylet being hard to remove it takes before somebody stops evaluating them one at a time and goes back to check whether the drawing is wrong. Every one of those complaints is individually explainable. The pattern is what carries the information, and patterns only surface if the person who owns the dimension ever sees the pile. ## Part of BD's fix is a label BD says it implemented corrective and preventive action, including product and labeling improvements, to enhance device robustness and reduce the likelihood of recurrence. The product improvement is the real work. The labeling half is worth staring at. ISO 14971 ranks risk controls in a fixed order. Inherently safe design first. Protective measures second. Information for safety, which means labels, warnings and the instructions for use, comes last. It ranks last because a label only works if somebody reads it, understands it, and recalls it at the moment of use. Intraosseous access exists for the moments when nobody is reading anything. A control that asks a paramedic to remember a withdrawal angle during CPR has a known failure rate, and the rate is not low. To be fair to BD, they found the root cause and they changed the product, and a labeling update running alongside a design fix is legitimate practice. The useful question for anyone building a device in this category is about what their own file looked like before the complaints started, not about what BD did once it knew. Dave's take I've never seen a risk file that handles this shape well. The hazard that gets you is rarely one part failing. It's a part sitting at the edge of its tolerance meeting a user doing something reasonable under pressure, and describing it takes two entries instead of one, which is exactly why it doesn't get written. Go read yours. Every control that ends in "per the instructions for use" is a bet that somebody reads the label during the worst ten minutes of their week. From Dave’s video library Dave walks through what a fractional product lead does week to week, what it costs against a full-time hire, and why six months is the floor before the work produces anything real. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [The Pulse Oximeter Worked. The Validation Didn't. →](/field-notes/2026-07-22-pulse-oximeter-validation-not-the-device) [Walter Reed Cleared the Implant. Now the Hospital Is the Manufacturer. →](/field-notes/2026-08-03-walter-reed-hospital-is-now-the-manufacturer) [A Disposable Surgical Robot Is a Manufacturing Bet →](/field-notes/2026-07-19-disposable-surgical-robot-is-a-manufacturing-bet) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Solventum Is Separating $1.4 Billion of Software. It Ran on a Different Clock. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-08-07-solventum-software-runs-on-a-different-clock [← All Field Notes](/field-notes) August 7, 2026 Medical Devices Bryan Hanson runs a medical device company that owns a $1.4 billion hospital software business, and this week he explained why he is getting rid of it. The software needs a different level of investment, he said, and a different pace of innovation. That is not a complaint about the software. That is a CEO saying his company cannot run at two speeds. Solventum plans to separate its health information systems unit, which sells hospitals software for things like computer-assisted physician documentation and coding automation, MedTech Dive reported Thursday. The unit brought in $1.4 billion last year, about 16 percent of company sales. Hanson is open to either a sale or a spinoff and expects the process to run twelve to eighteen months. What is left afterward is medsurg and dental, which is the point: he wants Solventum read as a true medtech company. The mechanics are unusually clean, and Solventum knows it. The company came out of 3M in April 2024, and the two-year anniversary of that separation is what gave it room to start selling pieces. It has already sold purification and filtration to Thermo Fisher for $4.1 billion. Hanson's read on this one is that the software business barely touches manufacturing, so the split is about as easy as these get. ## Two clocks under one roof Device work runs on a regulatory and manufacturing clock. You freeze the design on purpose, because unfreezing it means retooling, revalidating, and answering for the change. Hospital software runs on a release clock. Its competitors ship every month, and the IT director evaluating documentation tools is comparing feature velocity, not design history files. Put both under one P&L and one of them gets starved. It is usually the fast one, because the slow one carries the bigger revenue line and all of the audit exposure. Run that arrangement long enough and you arrive where Solventum is, holding a healthy asset that its own CEO thinks would move faster somewhere else. ## I have been on the naming side of this At Ascend Communications I was the global line manager for the access concentrator product line. Ascend was a hardware company down to its bones, and software was treated as the stuff that made the iron work. Inside those boxes sat about 30 million lines of source code I did not write. At peak, Ascend gear was roughly 90 percent of what ISPs used to connect the world to dial-up, so that code was already running most of the internet. It had no name. Cisco had already solved this on the other end of every connection we shipped. Their routers ran IOS, a named, branded operating system. So I walked into Mory's office in Alameda, opened with "I think you're doing everything wrong," and argued that we should brand our software the same way. He asked what it would be called. I said TAOS, on the spot. The code did not change. The architecture did not change. What changed is that customers, competitors, analysts and the press had a noun to point at when they explained why an Ascend box was different. I did not build TAOS. I recognized it, and I pushed hard enough inside a hardware-first company to get it treated as a product. That is the same recognition Hanson has made, run to the opposite conclusion. Once you name the software for what it is, there are two honest answers. This is our product and we should sell it as one. Or this deserves funding we are never going to give it. Ascend got the first answer. Solventum got the second. ## What this looks like before you have $1.4 billion Nobody reading this is separating a business unit next quarter. But the decision underneath it arrives early, and it arrives as a roadmap question. Almost every device I work on now has software that decides whether the customer keeps buying: the planning step, the calibration routine, the data pipeline, the dashboard the hospital's IT department will audit before it lets your box on the network. So ask which half sets your release cadence. If the hardware sets it, your software team ships on hardware time. Better to decide that in year one, while it is still a staffing and roadmap question. Two things follow from it. A change that touches cleared functionality is a regulatory question before it is a sprint question, which makes your software plan and your submission plan one plan. And engineers who arrived from a monthly release cycle will not sit through an eighteen-month design freeze unless somebody told them the truth about the cadence when they were hired. Dave's take The expensive version of this decision is a divestiture. The cheap version is a paragraph in your product definition naming which half of the product the company is organized around. Write the cheap one now. Hanson is spending eighteen months and an investment bank on the other. From Dave’s video library Dave on why software gets expensive at the point you have to change it, not the point you build it. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [ENvue's Feeding Tube Robot Came Second. That's the Product Decision. →](/field-notes/2026-07-30-envue-feeding-tube-robot-came-second) [MMI Didn't Build a Robot for Alzheimer's. Two Surgeons Decided It Was One Anyway →](/field-notes/2026-08-02-mmi-symani-alzheimers-who-defines-your-market) [The Bearings, Not the Algorithm: Where Hardware Scale-Up Stalls →](/field-notes/2026-07-24-bearings-not-the-algorithm-hardware-scale-up) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # An FDA Advisory Panel Reads Your File Cold. Grail Finds Out September 23. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-08-08-fda-advisory-panel-reads-your-file-cold [← All Field Notes](/field-notes) August 8, 2026 Regulatory Grail filed its premarket approval application for the Galleri blood test on January 29. On August 7 the company announced when it finds out what a room of outsiders makes of that file. September 23, in front of the FDA's Molecular and Clinical Genetics Panel of the Medical Devices Advisory Committee. Nearly eight months from filing to that meeting, and the room that shapes the decision is staffed by experts who never took the pre-submission call and have no history with the company. Founders read that meeting as a formality on the way to a decision the agency already made. It is closer to the opposite. A panel is a cold second read of your evidence by people who were not in the room when you designed the study. ## What a panel is asked to do CDRH convenes outside experts to assess a device's safety and effectiveness, and they vote. The vote does not bind the agency. Maisel-Campbell and colleagues, writing in PLOS One in 2022, reviewed 49 device panel meetings across 11 panels held between January 2011 and June 2016, and found the majority vote went in favor of approval in 43 of them. Call it 88 percent. The base rate is friendly, which is why founders relax into it. What the base rate hides is the subject of the argument. Panels spend their time on the fit between the evidence you brought and the claim you wrote. Those are two different objects, and the space between them is where a finished submission gets opened back up. ## Your study population is a claim you already made I learned this the expensive way at Galen. We ran a Q-sub before the study. We told the FDA we would do a 16-subject cadaver study, we would show that the difficult maneuvers got easier with the robot, and we would run it with residents. The FDA wrote back: sounds great, go for it. Then we filed, and we got to substantive review. First question out of their mouths, on a Zoom call because it was still the COVID era: we noticed you did your study with residents. Are you planning on marketing to residents? The thing they had approved in writing two months earlier was the thing they were now questioning. I went to Johns Hopkins with my hat in my hand, got free OR time and cadaver heads out of the neurosurgery department, recruited 16 attending surgeons, and re-ran the entire study in two weeks. One of them was a last-minute replacement, an otolaryngologist who does balloons and tonsils and will never do microsurgery in her career. She nailed the washout tests. That one data point turned out to be the strongest evidence we had that the device did what we said it did. We got there. Nobody should plan on getting there. A pre-sub answer is one reviewer's read on one day, and it does not bind the team that reviews your submission later. The population you tested in is a statement about who you intend to sell to, and a reviewer who never met you will read it that way. ## What Grail brought to the room Look at what Grail assembled and you can see a company that settled the population question early. The application rests on 25,490 consented participants from PATHFINDER 2 with a year of follow-up, plus more than 70,000 people from the intervention arm of the prevalent screening round of the NHS-Galleri trial. Asymptomatic adults, tested in the setting where the product would be used. Galleri has been on the market as a lab-developed test for years, past 475,000 commercial tests as of January, so the company had every reason to lean on its existing customer base. It ran the prospective screening studies anyway. Then read the intended use. The test is meant to be used "in addition to, and not as a replacement for, guideline-recommended screenings." That sentence is a fence. It tells a room of geneticists and pathologists what Grail is not claiming, before anyone in the room has to ask. Dave's take Founders with a submission in motion treat each FDA yes as a deposit that accumulates. It doesn't. Every round is staffed by different people with a different mandate, and the only thing that carries forward is the evidence itself. Build the file for a reviewer who reads it cold, in a room you are not in, with no memory of anything anyone told you. If the file only makes sense with your explanation attached to it, you have not finished it. From Dave’s video library I take apart a widely cited study where the system under test helped grade its own results, and what to check before you trust a number like that. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [How the Indication Sentence Decides Your 510(k) →](/field-notes/2026-07-29-indication-sentence-decides-your-510k) [The Pulse Oximeter Worked. The Validation Didn't. →](/field-notes/2026-07-22-pulse-oximeter-validation-not-the-device) [Momentis Built Its Own Predicate, Then Spent It Three Times →](/field-notes/2026-07-31-momentis-built-its-own-predicate) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # HII Put $900 Million on the Table for Shipyard Robots. The Qualification Comes First. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-08-09-hii-shipyard-robots-qualification-comes-first [← All Field Notes](/field-notes) August 9, 2026 Industrial Robotics HII will pay two robotics companies up to $900 million over seven years to help build aircraft carriers and submarines. It hasn't committed a dollar of it yet. The agreements HII announced on August 6 are performance-based production agreements. Path Robotics and GrayMatter Robotics get shipbuilding work only after they hit technology and manufacturing-readiness milestones that HII defines, and only if cost, schedule and quality hold up once real work starts. Every outlet that covered it led with the number. The words doing the work in that sentence are "up to." ## HII wrote a spec and put a payment schedule behind it The deals sit under HYPR, HII's High-Yield Production Robotics program, launched earlier this year. There are two stages. In the first, both companies work alongside HII to develop, validate and qualify high-precision production techniques. Only in the second does HII start routing actual shipbuilding work through the new line, and even then the release says that sourcing is contingent on favorable cost, schedule and quality performance. Target programs include carriers, submarines, destroyers, amphibious ships, future frigates and unmanned surface vessels. HII's reason is in a different number. The company plans to send more than 2.5 million hours of shipbuilding work outside its own yards in 2026, about 30 percent more than last year. That's a capacity problem it has already decided it can't hire its way out of. What it bought from Path and GrayMatter is a qualified process, with the acceptance criteria written by the customer and the invoices gated behind them. ## The customer specified welds and finished surfaces Path Robotics brings welding: its Obsidian physical-AI model and the Rove mobile system, which moves welding out of a fixed cell and onto assemblies too large to bring to the robot. GrayMatter brings surface preparation, coating and inspection. Grinding, blasting, painting. Between them the program covers welding, material handling, surface prep and quality checks on one coordinated line. Look at what HII named. Tasks, with tolerances behind them, and no opinion about what shape the machine should be. That's worth noticing in a week when the robotics headlines were humanoid unit counts and a new robotic hand for teaching dexterous skills. A weld on a submarine hull has an inspection standard sitting behind it. You can write a milestone against that. You can't write a milestone against general-purpose capability, which is why the nine-figure agreement went to the companies working on bounded jobs. ## What seven years of qualification is worth Here's the part founders underprice. Once a welding process is qualified into carrier and submarine production, HII has no appetite for requalifying it with somebody else. The switching cost is the qualification itself, sitting inside a production line the Navy audits. That's lock-in of the good kind. I've spent years drawing that distinction for founders who confuse it with the coercive kind, the pre-sold consumables and contract handcuffs that create a customer who wants out. Intuitive's da Vinci is the model I keep returning to. Surgeons are trained and credentialed on it, the OR is built around it, it's easy to buy and easy to service. Nobody is trapped. Leaving is just obviously the worse choice. The robot is copyable. The system around it isn't. That's what Path and GrayMatter are buying with the development stage, and it isn't the $900 million. It's a position inside a shipyard that a competitor would have to requalify from zero to take. Dave's take A big number attached to your company is a clock, not a crown, and it makes no difference whether it came from an investor or a customer. What these two companies owe now is a qualified process, on somebody else's schedule, in somebody else's yard, measured against milestones they didn't write. Watch what gets qualified, not what got announced. From Dave’s video library Dave makes the case that the clock on a build is external proof from a real buyer, not months or features. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Reimagine Wants Your Line Worker to Teach the Robot. Walden Wants to Do It for You. →](/field-notes/2026-08-04-who-teaches-the-robot-reimagine-vs-walden) [The Pentagon Rewrote the Drone Supply Chain. That's a Product Requirement. →](/field-notes/2026-07-28-pentagon-rewrote-the-drone-supply-chain) [The Bearings, Not the Algorithm: Where Hardware Scale-Up Stalls →](/field-notes/2026-07-24-bearings-not-the-algorithm-hardware-scale-up) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Caristo Built a Heart Test for the Patients Who Get Sent Home. The Billing Codes Came First. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-08-10-caristo-heart-test-billing-codes-came-first [← All Field Notes](/field-notes) August 10, 2026 Diagnostics Caristo Diagnostics built a heart test for the patients a heart scan sends home. That was the product decision. Everything the company has done since follows from it. On July 29 the FDA granted De Novo authorization to CaRi-Heart, software that reads coronary inflammation off a routine CT angiogram. The scan is one the cardiologist already ordered. CaRi-Heart measures the fat around the coronary arteries, scores the inflammation there, and returns an estimate of the patient's risk of dying from cardiovascular disease over the next ten years. Caristo, an Oxford spinout, plans to start selling it in the United States this quarter. ## The market was the negative result Coronary CT angiography is good at finding blockages. A patient turns up with chest pain, gets scanned, the radiologist finds no obstructive disease, and the patient goes home reassured. Caristo's validation data has something to say about that group. In a UK study of more than 40,000 patients scanned at eight hospitals, published in The Lancet and followed for a median of 2.7 years, the patients with no obstructive coronary artery disease accounted for two-thirds of the major adverse cardiac events and two-thirds of the cardiac deaths. There are two products sitting in that finding. One is a better blockage reader, and that shelf is crowded. HeartFlow cleared an updated plaque analysis platform last September, and Caristo's own CaRi-Plaque, cleared in 2025, handles plaque and stenosis. The other product is a test for the people the blockage reader clears. Nothing on the market did that, which is why there was no predicate to point at and why the file went in as a De Novo. ## The billing codes were live in January The American Medical Association gave CaRi-Heart two Category III CPT codes, 0992T and 0993T, effective January 1, 2026. The authorization came through on July 29. The billing path was in place close to seven months before the product could legally be sold here. Category III codes are tracking codes. No payer is obligated to pay against them, and plenty of them expire without ever converting to Category I. What they do is give the claim somewhere to go. An imaging center running the test next month has a number to put on the form, and the volume that number collects is the evidence any future Category I application will stand on. I push clients toward that sequence and it is reliably the least popular thing I say in the room. Engineering wants to finish the device. The regulatory consultant wants the submission in. Pulling the CFO and a coding specialist into a product meeting feels early to everyone except the person who has watched a cleared product sit on a shelf because nobody could work out how a hospital would bill for it. ## What the De Novo bought Founders hear De Novo and think slow lane. It is more work. It is also a signal, and the signal is decent news: the FDA routes you there when it considers your device moderate risk and inside the boundaries it is prepared to accept. A high-risk device is already in a PMA. At Galen we got that signal in a pre-submission response and took the hint. At the very end of the process the FDA asked us what the new regulation should say, because there was nothing in the CFR describing what we had built. We wrote a paragraph we liked and sent it over as a PDF. They asked for it in Word. A day or two later the clearance arrived, and the regulation describing our device was word for word what we had sent them. Caristo holds that position now for measuring coronary inflammation on CT. Anyone who wants to sell into that category has to be substantially equivalent to Caristo. That is worth having. It is worth a good deal more when there are codes to bill against and a launch already scheduled behind it. Dave's take The authorization is the last thing that happened at Caristo, not the first. Someone decided the market was the patients whose scan comes back clean, and the study population, the claim, and the billing codes all followed from that one sentence. That call belongs to whoever owns the product. If a regulatory consultant is making it at your company, the wrong person is holding the pen. From Dave’s video library I walk through how to tell whether you are building the right thing before you have spent months on the wrong one. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Momentis Built Its Own Predicate, Then Spent It Three Times →](/field-notes/2026-07-31-momentis-built-its-own-predicate) [Cepheid’s New GeneXpert Kept the Old Menu. That’s the Product. →](/field-notes/2026-07-25-cepheid-genexpert-kept-the-menu) [En Carta Signed Its Buyer Before It Built the Line →](/field-notes/2026-07-21-en-carta-signed-the-buyer-before-the-line) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # iRhythm Paid $287.5 Million to Stop Waiting on Its Own FDA Clearance | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-08-11-irhythm-vitalconnect-bought-out-of-the-clearance-clock [← All Field Notes](/field-notes) August 11, 2026 Medical Devices Quentin Blackford runs iRhythm, and he explained a $287.5 million purchase with a sentence about scheduling. "We're not going to wait around for an MCT clearance to then figure out whether to introduce one product versus the other," he told MedTech Dive. The clearance he's not waiting on is his own. iRhythm announced on August 7 that it will buy VitalConnect, the San Jose maker of the VitalPatch biosensor, for $237.5 million in cash and $50 million in stock. The deal should close by the end of the year. MedTech Dive's headline called it growing heart monitoring leadership, which is true, and which is the least interesting thing about the decision. Look at what iRhythm expects to be selling once it closes. Zio AT, the monitor it sells today, reaches roughly half the mobile cardiac telemetry market. Zio MCT, the next one, is sitting in front of the FDA and aims at another 20 to 30 percent. VitalPatch covers what those two leave behind, in hospitals and in remote monitoring, which is how Blackford gets to a market near $1 billion with about $500 million of it new to iRhythm. Three products, overlapping, sold at once, on purpose. ## The gap was in the calendar It's easy to treat a submission already in review as a launch date. The FDA hands you a range instead, and the range is widening. MedTech Dive ran the numbers in July: 510(k) clearances averaged about 156 days in the first half of 2026, up from about 148 a year earlier, and original PMA approvals stretched to roughly 599 days from about 402. The device center is hiring back after cutting more than 3,000 positions in 2025. None of that is a scandal. It's variance, and variance in a file you have already handed over is variance you can't manage. So Blackford bought his way out of the sequencing problem. With VitalPatch in the portfolio, nobody at iRhythm has to pick which product leads based on when a letter arrives, or hold a launch while the answer comes back. That's a different purchase than market share, and it's the one worth studying. ## Overlap only works if you can name the buyer Focus is the standard advice, and it's usually right. Carrying three products that do similar things goes wrong when the only thing separating them is a spec sheet. iRhythm's don't split that way. VitalPatch is cleared for hospital and remote monitoring, wears for seven days, and watches for 21 arrhythmias alongside vital signs. Zio is the long-wear ambulatory patch a cardiologist sends home with a patient. Different care setting, and a different person deciding to order it. I learned that difference from floppy disks. In high school I bought the premium tier every time, the box advertised as 100 percent surface tested, because I had allowance money and no way to check whether a cheaper disk would eat my work. A friend of mine was a product manager at Dysan and claimed credit for the tiering, graded by how much of each production run got scanned for defects. As he told it, the executives scoffed that nobody buys a floppy disk over testing percentages. The top tier outsold everything. Internally the cheap disks almost never failed, but I couldn't see the factory. I could only see the fear the expensive box removed. Every tier needs a plausible home. That's the test I'd run on iRhythm's three: not whether the products differ, but whether a hospital, a cardiologist and a payer would each land on a different one and feel well served. A customer who buys the smaller option and grows into the bigger one is healthy. A customer who buys the smaller one for exactly the right reasons and gets burned is damage you paid $287.5 million to build. Dave's take What I'd steal here is the honesty about the clock. Blackford has a file in review, he refused to let its timing set his product strategy, and he spent cash to take that decision back. Most of us can't write that check, so the move is to stop drawing roadmaps that assume a review lands on schedule, and to start naming which customers go unserved for every month it doesn't. From Dave’s video library Dave walks through the commercial questions that have to be answered before a clearance letter turns into a product anyone buys. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Cepheid’s New GeneXpert Kept the Old Menu. That’s the Product. →](/field-notes/2026-07-25-cepheid-genexpert-kept-the-menu) [ENvue’s Feeding Tube Robot Came Second. That’s the Product Decision. →](/field-notes/2026-07-30-envue-feeding-tube-robot-came-second) [Solventum Is Separating $1.4 Billion of Software. It Ran on a Different Clock. →](/field-notes/2026-08-07-solventum-software-runs-on-a-different-clock) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Hadrian Raised $1.37 Billion for Factories. The Drawings Are Still Your Problem. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-08-12-hadrian-factories-drawings-are-your-problem [← All Field Notes](/field-notes) August 12, 2026 Commercialization Hadrian raised $1.37 billion on August 6, and almost none of it goes toward inventing anything. The money buys floor space, machines, and the technicians to run them. Founder and CEO Chris Power framed it in national terms: "Production is now the frontline of deterrence." Underneath that framing sits a plain commercial bet, and the bet is about somebody else's problem. Hadrian is wagering that American hardware companies have parts they cannot get made. The round values the company at $7.87 billion. WCM Investment Management, Washington Harbour Partners, Valor Equity Partners, 137 Ventures and Baillie Gifford co-led it, with JPMorganChase's Strategic Investment Group anchoring. Hadrian runs four factories totaling just under three million square feet, two in Torrance, one in Mesa, one in Muscle Shoals, and it sells the output as what it calls Factories-as-a-Service. You don't buy the machine or hire the machinist. You send a part and take delivery. ## What you can buy, and what you still have to write down For a founder holding a working prototype, that promise is easy to like. Somebody else carries the capital equipment and the labor. You send the drawings. That last sentence is where programs stall. An automated factory runs what the drawings say and only what they say. It can't walk over to the machinist who has been cutting your part for two years and ask what they do differently on the third operation. They aren't there, and what they do isn't on the drawing. ## 2,800 ways to open a file I watched this play out in software, which is where I spent the first half of my career. When Apple changed the Mac file system, it deprecated the old primitives. Over the years there had been two or three sanctioned ways to open and save a file, and none of them had ever been properly retired. That's normal. Operating systems accumulate. When InterCon went to make its software compatible, we found more than 2,800 separate instances of file open and save calls in our own codebase. They weren't consistent enough for a search and replace. One engineer, Jim Bates, spent about six months cataloging every different way that code called the file routines, then went and updated every single one. A lot of applications got caught flat-footed when the cutoff came, because they weren't that thorough and missed instances of their own. The knowledge existed in the building. It had never been written down in one place, so no tool could reach it. A person had to go find it, one instance at a time, for six months. ## The part that actually gates you An automated factory is the same kind of machine pointed at a different kind of file. It executes a specification. Tolerances copied off a template and never examined, a finish callout nobody can measure, a revision that doesn't match the part sitting on your bench: none of that survives contact with a shop that has no relationship with your engineers and no memory of why the part looks the way it does. Inside your own building those gaps are invisible, because the person who fills them is standing right there. There's a cheap test for this and it takes about a week. Send your drawing package to a shop that has never met your team, and ask for a quote and a first-article plan. Then count the questions that come back. Every question is a piece of your product that lives in somebody's head instead of your documentation. That count is your real distance to production, and it has nothing to do with how much capacity exists in Arizona. The founders who will get the most out of a Hadrian, or any of the shops that follow it, are the ones who did that counting early, while a purchase order wasn't waiting on the answer. The capacity is the easy half now. It's the half with a price tag on it. Dave's take I've never seen a company reach production faster than its documentation allowed. Capital can buy the building and the machines, and Hadrian is proving it can buy them at a scale I wouldn't have predicted three years ago. What capital can't buy is the six months somebody has to spend writing down what your team already knows. Do that work while it's still cheap. From Dave’s video library Dave reads Anthropic's own founder guidance against its headline and shows where the hard work actually moved to. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [The Bearings, Not the Algorithm: Where Hardware Scale-Up Stalls →](/field-notes/2026-07-24-bearings-not-the-algorithm-hardware-scale-up) [HII Put $900 Million on the Table for Shipyard Robots. The Qualification Comes First. →](/field-notes/2026-08-09-hii-shipyard-robots-qualification-comes-first) [The Pentagon Rewrote the Drone Supply Chain. That's a Product Requirement. →](/field-notes/2026-07-28-pentagon-rewrote-the-drone-supply-chain) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Joby Is Paying $500 Million to Move Its Defense Work Out of the Building | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-08-13-joby-moved-defense-work-out-of-the-building [← All Field Notes](/field-notes) August 13, 2026 Defense Hardware Joby Aviation is paying about $500 million for a 250-person company in Dayton, Ohio. Then it is moving its own defense programs into the thing it just bought. Joby's announcement says why: housing that work inside Resonant Sciences lets the commercial aviation organization keep its primary focus on certifying, manufacturing, and commercializing the electric air taxi. That is a company spending $450 million in cash to stop two of its engineering programs from fighting over the same calendar. ## What Joby actually bought Resonant Sciences designs, manufactures and delivers RF and mission systems for U.S. national security customers. It runs out of Dayton with facilities in Virginia, West Virginia, Michigan, Colorado and North Carolina. Joby's release puts it at roughly 250 employees, more than $100 million in trailing revenue growing 40 percent year over year, and EBITDA margins in the high teens. The price is $450 million cash and $50 million in Joby stock, with certain Resonant employees keeping an equity stake. Co-founder and CEO J. Micah North stays in the chair. Close is expected in the first half of 2027, subject to regulatory approval. In a normal tuck-in the acquired team gets absorbed into the buyer's org chart. Here it runs the other way. Joby's dual-use turbine-electric and hydrogen-electric aircraft programs move into Resonant, along with its dual-use autonomy stack and, per TechCrunch, the L3Harris collaboration on a gas-turbine hybrid VTOL. The hydrogen-electric work flew 521 miles in 2024. That is not a program you shut down. It is a program you cannot run next to a type certification. ## A certification program takes the whole company, on no notice I ran into the sharp end of this at Galen. We did our Q-sub and told the FDA we planned a 16-subject cadaver study using residents. It was in the letter. The FDA wrote back that it sounded great and to go for it. Then we got into substantive review, and the first question out of their mouth was whether we intended to market to residents. The study design they had approved in writing two months earlier was now the thing they were questioning. So I went to Johns Hopkins with my hat in my hand. Neurosurgery gave me a free OR and cadaver heads, on the condition that I not disturb the brain. In two weeks I recruited 16 attending surgeons and re-ran the entire study. That is the notice a regulator gives you when it decides your file needs something it did not ask for before, and there is no version of that scramble where your best engineers and your quality lead are half-committed to a customer program with its own delivery date. FAA type certification runs on the same physics. So does a DoD qualification. ## Underneath the calendar fight is a configuration fight A certification program wants a frozen design baseline and a traceable record of every change made against it. Contract R&D wants a fresh configuration per program and per customer. Both are legitimate ways to run engineering. Put them in one building and the freeze leaks, because the people holding the baseline are the same people being asked to modify it for the customer who is paying this quarter. I have watched medical device teams grow a smaller version of the same fork. A surgeon champion asks for a variant, the team builds it because the relationship is worth more than the hours, and the design history file quietly picks up a branch nobody documented. The same thing happens in climate hardware when a utility pilot wants a one-off configuration. The defense version is the same fork with a classification on it, which makes it harder to see and much harder to unwind. None of that makes Joby's defense revenue a mistake. A business doing north of $100 million, growing 40 percent, at high-teens margins is a better business than most of what the eVTOL sector has produced in fifteen years, and Joby is keeping it. What Joby is buying is somewhere else to put it: its own CEO, its own P&L, its own address. Dave's take I have watched founders treat contract revenue as a hedge against a long certification path, and it works right up until the regulator moves. Then you find out you have two products and one team, and the program with a customer paying this quarter wins the argument every time. Joby's fix costs $450 million in cash; the cheap version is a separate team, a separate design baseline, and the willingness to tell a paying customer no. From Dave’s video library Dave walks through what happens when a small team carries several products at once, and the four questions that cut the list down. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [iRhythm Paid $287.5 Million to Stop Waiting on Its Own FDA Clearance →](/field-notes/2026-08-11-irhythm-vitalconnect-bought-out-of-the-clearance-clock) [The Pentagon Wants Four Times the THAADs. The Shopping List Says Muzzle Covers. →](/field-notes/2026-08-05-pentagon-interceptor-surge-buys-muzzle-covers) [How the Indication Sentence Decides Your 510(k) →](/field-notes/2026-07-29-indication-sentence-decides-your-510k) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Cook Medical's Breach Never Touched the Device. It Went Through a Person. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-08-14-cook-medical-breach-went-through-a-person [← All Field Notes](/field-notes) August 14, 2026 Medical Devices On July 2 somebody at Cook Medical said yes to the wrong person. No password got cracked. An employee was deceived by a social engineering attack and gave an outside party access to company systems, MedTech Dive reported on August 13. Cook spotted it and shut it the same day. Here is what went out the door. Contact information for customers in the United States and Canada. Records of communications with Cook employees held in Salesforce. Some internal business files. Employee names and company email addresses. Cook says it has no evidence that sensitive or protected data was accessed, and that products, manufacturing, and its ability to serve patients and customers were unaffected. ## What a CRM is worth to the person who took it Nothing on that list is the device. It's the commercial layer: who your customers are, what they said to your sales and service people, and which of your employees to write to next. For a company Cook's size, that CRM is a map of hospital relationships built over decades. A competitor would find it useful. Somebody planning the next round of scam emails would find it more useful, which is why Cook is telling customers and employees to watch for follow-on attempts. Founders building connected hardware spend real money on the other layer. If you're writing a premarket submission for a device with software in it, you have a cybersecurity section, and you've put months into threat modeling, the software bill of materials, the patching plan. That work is worth doing. It also protects a completely different thing than the one that broke here. ## Where I learned to put people at the top of the list I knew Kevin Mitnick a long time ago. He lived in Maryland, I lived in DC, and Interop drew everyone in or near the hacker community, so it was close to impossible for someone in my line of work not to have met him. He was a good friend before he died. What he was famous for was hacking without ever touching a computer. He had a personality people responded to, and allegedly he could charm his way into information that was supposed to require special access. Nobody handed him a keyboard. They handed him the answer. So social engineering sits at the top of my own threat model, above anything on the network diagram. On the surgical robots I've worked on, hospital IT never allowed a live IP connection during a procedure. Store and forward only. That's a real control and I'd argue for it every time, but store and forward still carries injection risk, and my larger worry was always someone with live access. A friendly person in a company vest who talks their way to a terminal. Not Kevin, who was a good guy, but somebody with his skills and worse intentions. Hospital cyber departments already understand this, which is part of why they can gate your deployment independent of whether the device works. Design for human access control, man in the middle, and packet injection from the architecture stage, along with the unglamorous stuff like an open USB port. Security planned in is cheap. Security added after the design freeze is expensive and leaky. ## The half of the model with no section number Designed-in security is about stopping the incident. Cook's disclosure teaches the other half, which is what you do once one starts. Same-day detection and containment on July 2, public disclosure in August, an honest inventory of what was reached and a plain statement of what was not, operations running the whole time. That's a company that had practiced its incident response rather than written it. The device file covers the device. In every submission I've worked on, the cybersecurity questions were about the product: its interfaces, its software components, its update path. None of them asked who inside the company could pull the entire customer list, or what happens when that person takes a convincing phone call. The regulation is doing its job. It asks about the product because the product is what it governs. The gap is in what founders count as done when they say they've handled security. Dave's take The strongest attack on any system runs through people, and people don't show up on the network diagram. Cook caught theirs in a day, which is the outcome you want and the one you only get by practicing. Go find out who at your company could hand over the customer list this afternoon, then go ask them how they'd know not to. From Dave’s video library Dave on handing decisions to systems that were never smart enough to hold them, and where a person still has to stay in the loop. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Solventum Is Separating $1.4 Billion of Software. It Ran on a Different Clock. →](/field-notes/2026-08-07-solventum-software-runs-on-a-different-clock) [BD's Needle Was Out of Tolerance for Twenty Months. Part of the Fix Is a Label. →](/field-notes/2026-08-06-bd-needle-out-of-tolerance-fix-is-a-label) [iRhythm Paid $287.5 Million to Stop Waiting on Its Own FDA Clearance →](/field-notes/2026-08-11-irhythm-vitalconnect-bought-out-of-the-clearance-clock) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # The Medicare Coverage Decision Now Happens Before Your Pivotal Trial | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-08-15-medicare-coverage-decision-before-pivotal-trial [← All Field Notes](/field-notes) August 15, 2026 Regulatory A breakthrough device that clears the FDA waits an average of 5.7 years before Medicare will pay for it. That number comes from Stanford research by Josh Makower, and it belongs on the wall of any company about to design a pivotal trial. Market authorization is not the day you get paid. It is the day a second review begins, run by an agency asking a different question from the one you spent four years answering. In April, CMS and the FDA announced a pathway built to close that gap. They call it RAPID, for Regulatory Alignment for Predictable and Immediate Device coverage. The mechanism fits in a sentence: CMS issues a proposed national coverage determination the same day the FDA grants market authorization, then targets a final decision within 60 to 90 days. A device that would have spent a year or more in what Andrew Ibrahim of Viz.ai calls “coverage purgatory” can be covered in about two months. ## The entry gate is a trial-design decision In August, CMS filled in the part that changes how you plan. The two agencies start working together before the pivotal trial is designed, and a device has to be in early testing to get in. Eligibility runs through Breakthrough Device designation: all Class III breakthrough devices meeting the criteria, and Class II breakthrough devices also enrolled in the FDA’s Total Product Life Cycle Advisory Program. The investigational study has to enroll Medicare beneficiaries and measure outcomes both agencies agreed to in advance. That last requirement is where the work is. The population you enroll to satisfy the FDA and the population that answers CMS are not automatically the same people. Medicare beneficiaries are 65 and older, and they show up with everything else that comes with being 65 and older. A trial built around a healthier cohort can produce a clean safety and effectiveness story and still leave CMS holding no evidence about the patients it covers. Under the old sequence you could find that out after authorization and go run a second study. Under RAPID you’ve already spent your one shot. ## Who owns the coverage decision The instinct is to hand this to a reimbursement consultant. That is the wrong seat. Deciding which patients your device is for, and what outcome you will claim it produces, is the same decision that sets your indication, your predicate logic, your study population, and your labeling. Framing the indication is the regulatory strategy, and it’s the argument I make first when a founder asks me where regulatory work goes wrong. The device you build and the product you define are two different things, and only one of them goes through review. RAPID stretches that same framing decision to cover who pays for it. The person who owns the product owns that call. A scale check is worth running before anyone rebuilds a roadmap around this. CMS estimates roughly 40 devices qualify today, with about 20 more expected. That’s a real number and a small one. It’s also bigger than what came before it: the Parallel Review program the two agencies launched in 2011 and made permanent in 2016 capped participation at five candidates a year, and very few devices ever completed it. Letting more companies in is the easy half. Getting them to use it is the half that killed the last attempt. For everyone outside the breakthrough lane, none of the underlying math moved. Hospitals cannot fund science experiments anymore. Margins are tight enough that a promising capital purchase with an unproven payback does not get budget, and the buyer runs that arithmetic long before your first system reaches a loading dock. Coverage is one input to that calculation. RAPID makes the input arrive sooner for a few dozen companies. It does not make the number close. Dave’s take The speed is what gets the headlines. What matters more is that RAPID drags the question of who pays into the room while the protocol is still editable, and that is the only moment when answering it is cheap. If your pivotal trial is already enrolling, you missed that window. Plan for the wait. From Dave’s video library A companion on evidence: how to design proof of demand that survives a person who has every reason to tell you what you want to hear. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Caristo Built a Heart Test for the Patients Who Get Sent Home. The Billing Codes Came First. →](/field-notes/2026-08-10-caristo-heart-test-billing-codes-came-first) [How the Indication Sentence Decides Your 510(k) →](/field-notes/2026-07-29-indication-sentence-decides-your-510k) [iRhythm Paid $287.5 Million to Stop Waiting on Its Own FDA Clearance →](/field-notes/2026-08-11-irhythm-vitalconnect-bought-out-of-the-clearance-clock) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # H.B. Fuller Already Made Surgical Glue. It's Paying £715 Million for the Evidence. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-08-16-hb-fuller-715-million-for-the-evidence [← All Field Notes](/field-notes) August 16, 2026 Medical Devices H.B. Fuller already makes surgical glue. About $68 million worth, assembled by buying four smaller medical adhesive companies. This week the shareholders of a fifth, the British tissue-healing company Advanced Medical Solutions, approved a takeover valued at £715 million. AMS makes surgical adhesives, sealants and dressings. So does the buyer. ## The chemistry was the cheap part On an earnings call in June, H.B. Fuller CEO Celeste Mastin described the market she was buying into as one whose barriers to entry include "regulatory approvals, clinical validation and rigorous customer qualification." The offer announcement made the same point in fewer words, calling medical a core growth market for its "high regulatory-based entry barriers." The barriers are the reason to buy. A manufacturer with its own formulation chemists, its own plants, and four prior acquisitions in this exact space looked at what it would take to grow a medical adhesives business out of its own labs, and decided £715 million was the faster road. Making the product was never what stood in the way. The money points the same direction. AMS brings roughly $300 million in annual revenue, and H.B. Fuller is paying 12.9 times EBITDA before synergies, expecting about $55 million a year in combined revenue and cost savings by 2031. Some of that is the cost of running a public company going away. What the buyer cannot grow on its own schedule is the part AMS spent years accumulating: products already through their regulatory gates, and the clinical evidence sitting behind them. ## A patent is a clock. A clearance isn't. A patent is a clock, not a moat. It buys a defined window, and the only thing that window is good for is building something behind it that doesn't expire on a date your competitor can read as easily as you can. More than a decade ago, maybe closer to fifteen years, I worked on a University of Nebraska Medical Center project that built a navigated sagittal saw for preparing bone to take a knee implant. Nearly every bone-shaping robot at the time used a burr, which grinds material away. Our claims were built around cutting planes instead. I went to Washington and walked a patent examiner through why cutting a plane with a saw and destroying material with a burr are different operations that produce different results. He agreed. The patents issued to the university, and I managed them. Then the company didn't push it forward. Tick tock. The patents drifted toward expiry, and Stryker now ships Mako RPS, a handheld robotic saw for total knee replacement that does the job better than our version ever did. Somebody read the expiration date and built the better product for the day after. The H.B. Fuller deal prices the difference between those two kinds of asset. A patent runs out on a schedule anyone can look up. A clearance doesn't lapse because a calendar says so, a clinical data set doesn't stop being evidence, and a hospital that has already qualified you as a supplier doesn't hand your competitor that status for free. ## Who this is for This applies to any founder whose product needs a regulator's sign-off or an institutional buyer's qualification before a single unit ships. Medical devices and diagnostics, plainly. Also defense hardware moving through test and evaluation, grid equipment facing utility interconnection review, and industrial robots that have to clear a plant's safety qualification before they go near a line. It's easy to staff that work as a compliance function and park it off to the side of the product organization, where the only question anyone asks is whether it's holding up the launch. Price it the way an acquirer prices it and it belongs on the roadmap, owned by whoever owns the product, with the attention the technical milestones get. Dave's take I've watched teams treat the regulatory and clinical work as the tax they pay to get a product out the door. H.B. Fuller just published a receipt for what that work is worth on its own: £715 million for a business doing about $300 million a year, paid by a buyer that could already make the product. Build the file like it's the asset, because to whoever eventually writes the check, it is. From Dave’s video library A related look at the distance between what a health technology promises and what the measured evidence supports. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [iRhythm Paid $287.5 Million to Stop Waiting on Its Own FDA Clearance →](/field-notes/2026-08-11-irhythm-vitalconnect-bought-out-of-the-clearance-clock) [Momentis Built Its Own Predicate, Then Spent It Three Times →](/field-notes/2026-07-31-momentis-built-its-own-predicate) [Caristo Built a Heart Test for the Patients Who Get Sent Home. The Billing Codes Came First. →](/field-notes/2026-08-10-caristo-heart-test-billing-codes-came-first) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Abbott Makes the Glucose Number. Google Charges for What to Do About It. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-08-17-abbott-lingo-google-owns-the-coaching-layer [← All Field Notes](/field-notes) August 17, 2026 Medical Devices A continuous glucose sensor produces a number every few minutes. What a person pays for is the sentence that comes after the number. On August 11, Abbott and Google announced a multi-year partnership that hands those two jobs to two different companies. Abbott took the first one. ## What each company took Abbott's Lingo is an over-the-counter sensor. It launched in the US in 2024, it's for adults 18 and over who don't take insulin, and it isn't intended to diagnose disease. Under the partnership, Lingo readings flow into the Google Health app, where a user sees glucose sitting next to sleep, activity and whatever else Google already collects. Google's AI health coach reads that pile and returns guidance on nutrition, activity, sleep and recovery. The coach is a premium subscription feature. Integrations roll out later this year. Neither company disclosed financial terms. So Abbott carries the sensor: the manufacturing, the regulatory file, the consumable economics, the box on the shelf. Google carries the app icon, the daily open, the subscription, and the screen where a habit either forms or doesn't. Google Health is built on the $2.1 billion Fitbit acquisition Google closed in 2021, so the other signals were already being collected before glucose showed up. ## I have built the innards before My group built the original Apple Airport at a request from Steve Jobs to Lucent. Lucent owned most of the early Wi-Fi patents and we built the guts. Apple designed the plastic and put its name on the outside. Crack open one of those first flying-saucer Airports and there's a Lucent Orinoco card sitting in it. No consumer went shopping for a Lucent product. They bought an Apple one. That was a reasonable trade for Lucent in 1999 and I'd take it again. It's also true that Apple later shipped an Airport card of its own, which is how component relationships tend to go. The name on the outside keeps the customer, and the part on the inside is a decision somebody revisits. ## The study is the term I'd negotiate over The two companies also plan what they describe as one of the largest real-world metabolic health studies to date, combining continuous glucose, wearable, lab and survey data. Abbott didn't disclose enrollment. The stated purpose is to inform future AI coaching in Google Health and new Lingo features. Neither announcement says who ends up owning that data set. That's the term I'd want written down before anything else in the agreement, because a coach trained on years of paired glucose and behavior data keeps its value long after the sensor generation that produced it has been replaced by something cheaper and thinner. ## Dexcom aimed the same idea at a different gate Dexcom, Abbott's main rival in glucose sensing, is taking part in an FDA digital health pilot testing AI-enabled glucose management, with the goal of improving glycemic control. Same underlying idea, opposite direction: put the intelligence inside the regulated product, where it moves slower and can carry a claim. Abbott put its coaching layer outside the device, in a wellness app that ships this year and can't say it treats anything. Both are defensible. The choice decides what your marketing is allowed to say, who signs off before you change the algorithm, and whether an insurer ever has grounds to pay for the thing. ## Where else this shows up The pattern isn't confined to medtech. It arrives any time a hardware company gets invited into somebody else's app: a battery maker into a fleet operator's dashboard, an industrial sensor company into an OEM's monitoring platform, a diagnostics instrument into a hospital's records system. The partner brings distribution you couldn't buy at any price, and takes the layer where the relationship lives. Platform dependency is a line item on your cap table that nobody writes down. You never raised it, it never hit your bank account, and you can't renegotiate its terms the way you'd renegotiate with an investor. You pay for it in control instead of equity. Dave's take I'd price a deal like this the way I price capital, and I'd ask one question before signing it: if the partner turned hostile tomorrow, could I switch my own business back on without them? Abbott can. It has the sensor, the shelf space and a brand people already trust, and it was selling Lingo before Google showed up. A smaller company signing the same structure usually can't, and the day you're the biggest thing on somebody else's platform is the day it costs the most to find out. From Dave’s video library Dave walks through where AI spending returns something you can measure, and where it quietly turns into a meter you keep feeding. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Solventum Is Separating $1.4 Billion of Software. It Ran on a Different Clock. →](/field-notes/2026-08-07-solventum-software-runs-on-a-different-clock) [iRhythm Paid $287.5 Million to Stop Waiting on Its Own FDA Clearance →](/field-notes/2026-08-11-irhythm-vitalconnect-bought-out-of-the-clearance-clock) [Caristo Built a Heart Test for the Patients Who Get Sent Home. The Billing Codes Came First. →](/field-notes/2026-08-10-caristo-heart-test-billing-codes-came-first) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # J&J Cleared AI on Its Lung Robot. The Fourth Update in Eighteen Months Matters More. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-08-18-jj-monarch-fourth-update-in-eighteen-months [← All Field Notes](/field-notes) August 18, 2026 Surgical Robotics Johnson & Johnson has now put four software releases onto the same lung biopsy robot in eighteen months. MedTech Dive reported the latest clearance, Monarch Quest 3, on Monday. Most of the coverage leads with the AI, because the update draws nodule boundaries automatically during planning. The AI is the least interesting part of it. ## What actually shipped Monarch is the robotic bronchoscopy system J&J uses to reach nodules deep in the lung for biopsy. Per MedTech Dive, Quest 3 adds AI-generated nodule boundaries during pre-procedure planning, updated navigation and visual guidance during the case, and a 3D view that helps the operator hold the relationship between the controls and the patient's anatomy. The cart in the room did not change. This is the fourth major software launch on that platform in eighteen months. Monarch competes with Intuitive's Ion and Noah Medical's Galaxy in robotic bronchoscopy. J&J also won FDA authorization for its Ottava soft tissue robot in July, which puts the company up against da Vinci and Medtronic's Hugo on a second front. Four cleared software releases in eighteen months works out to a release roughly every four and a half months, through a federal review, on a device already sitting in hospitals. That cadence is harder to build than the algorithm riding on it. ## The cadence is an architecture decision Whether you can ship a software-only improvement to a cleared device gets settled years before you file anything. It depends on how tightly your software is bound to your hardware, how your requirements and risk files are organized, and whether your verification strategy lets you re-test a changed function without re-testing the machine around it. Get those wrong and every improvement drags the whole device back into review. Get them right and you can hand the FDA a small, well-bounded change and keep your customers on the current version. I led the de novo effort at Galen Robotics, and the lesson that stuck was how much of a submission comes down to the boundary you drew around the device rather than the device itself. Founders tend to treat that boundary as paperwork to be sorted out once the engineering is done. It is engineering. Draw it in the wrong place and you have signed up to re-open your whole file every time somebody improves the navigation. That makes it a product decision, and it belongs to whoever owns the product. The roadmap question runs past what the next feature does, to whether you can put that feature in the hands of the customers who already bought the machine. ## The AI claim is narrow on purpose Look at how the AI is actually described. It generates nodule boundaries during planning. That is a specific job, on a specific input, producing an output somebody outside the company could examine. Nobody at J&J announced that the robot now understands the lung. That gap is the difference between a claim written for a reviewer and a claim written for a news cycle. My test on any AI announcement is whether the claim can be checked by someone with no financial stake in it, inside the window where checking still changes anything. A lot of what the big labs put out fails that test by design, and the pressure from it lands on hardware founders as a board question about what you are doing with AI. A regulated product forces the better answer, because you have to write the claim in a form a stranger can evaluate. If you cannot state your AI feature that way, what you have is positioning. Founders in diagnostics, industrial robotics, and defense autonomy are all being pushed toward the same announcement. The ones with a regulator or a qualification authority in the loop get handed the discipline for free. Everybody else has to impose it. Dave's take What I would protect on that roadmap is the release clock, ahead of any single algorithm riding on it. Intuitive did not win the operating room with a better arm. It won by solving so much of the surrounding problem that leaving stopped making sense to anyone, and a company that can improve a cleared product every few months is building that same hold one release at a time. The box gets copied. The clock is harder. From Dave’s video library Dave walks through how P.T. Barnum drew 38 million people, and why the same moves keep showing up in AI announcements. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [How the Indication Sentence Decides Your 510(k) →](/field-notes/2026-07-29-indication-sentence-decides-your-510k) [Momentis Built Its Own Predicate, Then Spent It Three Times →](/field-notes/2026-07-31-momentis-built-its-own-predicate) [MMI Didn't Build a Robot for Alzheimer's. Two Surgeons Decided It Was One Anyway →](/field-notes/2026-08-02-mmi-symani-alzheimers-who-defines-your-market) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # The Navy Wants Sixteen Times the Tomahawks. The Seven Years Is What Buys Them. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-08-19-navy-tomahawk-seven-years-buys-the-rate [← All Field Notes](/field-notes) August 19, 2026 Defense Hardware The Navy signed a contract on August 17 that asks Raytheon to build more than a thousand Tomahawk cruise missiles a year. Raytheon currently builds about sixty. Acting Navy Secretary Hung Cao called it a landmark contract, which is what you say about $22.9 billion, and the money is the least interesting number in the deal. The interesting number is seven. That's the length of the contract in years, and it's the term that makes a thousand missiles a year buildable at all. ## Sixteen times the rate, same missile Divide a thousand by sixty and you get roughly sixteen. Raytheon has to multiply its output by sixteen without changing the product, which means almost none of the hard work is engineering work on the missile. The design is done. What has to change is everything around it. Raytheon president Phil Jasper named the four things in one sentence when the deal was announced: workforce, technology, supply chain, facilities. Every one of those is a purchase with a payback measured in years. A second production building does not earn its cost back in a quarter. A qualified supplier for a rocket motor case does not appear because you sent a purchase order. Somebody has to pay for the tooling, sit through first article inspection, and carry that supplier through qualification before a single part counts. Hiring and training a machinist who can hold the tolerance takes as long as it takes. That is the shape of a rate problem in any hard-tech company. Rate is a design parameter, the same as mass or power draw, and it is the one nobody writes into the requirements document. You buy it with capital, and capital wants to know how long the demand lasts. ## A one-year order funds overtime A one-year order funds overtime and a second shift. A seven-year order funds a building. That difference has almost nothing to do with the size of the check and everything to do with how far out the customer is willing to commit. Raytheon's output was already moving before this contract. The company delivered three times as many Tomahawks in the first half of 2026 as it did in the first half of 2025, off the back of a framework agreement signed in February. February set the terms. August put seven years behind them, and that is when you can justify pouring concrete. The Pentagon has been running this play across programs, with similar multiyear deals for Patriot and THAAD interceptors signed earlier this month. The mechanism is the same every time. Give a manufacturer a horizon long enough to invest against and the capacity shows up. Take the horizon away and nobody sane builds the second building. ## Your version of this is a letter of intent No founder reading this is getting a seven-year Navy contract. What shows up instead is a letter of intent, a pilot agreement, a distributor's forecast, or a strategic investor who says they also want to be a customer. The question to put to any of them is the one Raytheon's finance team had to answer: how long is this good for, and does that horizon cover what I am about to spend? Getting the order wrong looks like success right up until it doesn't. The pizza robot graveyard is full of machines that worked. Real hardware, real demos, dead companies, because the founders fell in love with their product and not their customer. Build the customer proof before you build the fleet. The opposite mistake is real too. If a customer does commit and you cannot hold rate, you have handed them a reason to go qualify a second source, and you rarely get that position back. Both things are true, which is why there is no rule about building early or late. There is a number. Put the payback period on the capital next to the duration of the demand, and if the first one is longer than the second, you are not buying capacity. You are gambling. Dave's take Money is runway, not a moat, and a contract works the same way. Watch what a company builds with the number and how long it has to build it. Raytheon got seven years. Most founders are working against a customer who will not commit past next quarter, and that is a design constraint whether you write it down or not. From Dave’s video library Dave walks through what a large spending number does and does not buy, and where the judgment has to stay with you. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Joby Is Paying $500 Million to Move Its Defense Work Out of the Building →](/field-notes/2026-08-13-joby-moved-defense-work-out-of-the-building) [The Pentagon Wants Four Times the THAADs. The Shopping List Says Muzzle Covers. →](/field-notes/2026-08-05-pentagon-interceptor-surge-buys-muzzle-covers) [Hadrian Raised $1.37 Billion for Factories. The Drawings Are Still Your Problem. →](/field-notes/2026-08-12-hadrian-factories-drawings-are-your-problem) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # GM Walked Away From This Battery Plant Twenty Months Before It Opened | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-08-20-gm-walked-away-from-this-battery-plant [← All Field Notes](/field-notes) August 20, 2026 Climate Hardware In December of 2024, General Motors walked away from a battery plant it had nearly finished building. The Lansing, Michigan site runs to 2.8 million square feet. GM sold its stake to its joint venture partner, LG Energy Solution, kept its Ultium plants in Ohio and Tennessee, and its CFO Paul Jacobson said the company already had the cell and manufacturing capability it needed to grow with the EV market. On August 18 the plant started production without GM. Most of what comes off the line is not going into cars. ## What the building makes now LG has put more than two billion dollars into the 226-acre site since 2022 and is aiming at more than 35 gigawatt-hours a year. The plant runs lithium iron phosphate cells for grid and commercial storage, with DTE Energy as a named customer and LG's own Vertech arm integrating the systems. It also runs nickel manganese cobalt cells for cars, starting with Toyota's 2027 Highlander EV. About 900 people work there now. Full production is 1,700. None of those customers were the plan. The site was announced in 2022 as Ultium Cells, a joint venture that existed to feed GM's own vehicles. The owner changed, the chemistry changed, and the end market moved from cars to the grid. The building didn't. GM read its own demand wrong, and the plant was fine. A cell line can be tooled for LFP or NMC. The dry rooms, the electrical service, the trained operators, the suppliers qualified within a day's drive: all of it carries across a change of chemistry and a change of customer, because none of it was specific to a Chevrolet. ## I have been on the other side of this Before the internet went public I worked at a company that was the only game in town. Our customers were defense contractors and government labs, anyone with thousands of machines that needed to reach a mainframe. We hit a big monthly number one year and everybody got a custom leather jacket. Another time we had a party at the Smithsonian Air and Space Museum. Technically they held a private reception to thank us for a generous donation, which is a distinction I have never been able to make stick. Then the internet opened up and our software was not built for regular people. We tried. We built a browser. We sold a version to AOL for the Mac. Eventually a high-density modem company acquired us to bundle the software, and when I flew out to brief their sales team, their VP of Sales stopped me halfway through the pitch. "I'm really not sure why you're here." He wasn't rude. He was right. His customers were ISPs connecting somebody's grandmother, not Boeing and not Los Alamos. The asset in that story was code, which is supposed to be the flexible kind, and it would not bend at all. Who the user was ran down through the install process, the documentation, the support model, and the sales compensation plan. We were B2B in the DNA. The software faded out. It would have been cleaner if it had died. ## The question to ask while the line is still on paper Lansing and my old employer both got the market call wrong. One of them had somewhere to go afterward, and that was decided years earlier by people choosing tooling. Flexibility of that kind costs money on the day you buy it. Equipment that can run a second chemistry is more expensive than equipment that runs one. A supplier qualified for two form factors takes longer to qualify. Every one of those calls looks like waste at the time, because at the time the forecast still looks right. Your demand model will be wrong. What matters is what happens to the money you already spent when it is. Before the line gets built, that means asking: if this market comes in at half, what else can this equipment make, who buys that, and what would it cost us today to keep the door open. Sometimes the answer is that it costs too much and you take the concentrated bet deliberately. That's a fine answer. Arriving at it by default is not. Dave's take I have said before that if you want options, you need structure. That usually comes up about a founder's calendar, and it works the same on a factory floor. GM read the EV market wrong and it cost the company a stake in a building. It did not cost the building. Design for the market you expect, then spend a little on being wrong about it. From Dave’s video library Dave on the business that only runs while you are running it, and how to tell when you have built yourself a job. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Peak Energy Picked the Weaker Chemistry. Then It Deleted the Cooling System. →](/field-notes/2026-08-01-peak-energy-deleted-the-cooling-system) [Hadrian Raised $1.37 Billion for Factories. The Drawings Are Still Your Problem. →](/field-notes/2026-08-12-hadrian-factories-drawings-are-your-problem) [Joby Is Paying $500 Million to Move Its Defense Work Out of the Building →](/field-notes/2026-08-13-joby-moved-defense-work-out-of-the-building) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # The FDA Authorized a Blood Draw Robot. The Ratio in the Label Is the Product. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-08-21-blood-draw-robot-ratio-in-the-label [← All Field Notes](/field-notes) August 21, 2026 Medical Devices A machine in an outpatient lab can now find a vein, tell it apart from an artery, apply the tourniquet, prep the skin, insert the needle, change the tubes, and put on the bandage. Nobody touches the patient. The FDA authorized it on August 19 through the De Novo pathway, and gave Vitestro's Aletta a device category that did not exist the day before. The sentence that decides whether the company has a business is none of that. It's a ratio in the labeling: one phlebotomist, three machines. ## The ratio is the product The buyer is a hospital lab or an outpatient draw station with phlebotomy positions it has not been able to fill. What that buyer is purchasing is coverage. The FDA found the device drew blood at rates comparable to or better than trained human phlebotomists across diverse patient populations, which is the price of admission. A machine that draws as well as a person, one for one, gives a lab manager nothing. At one to three, the salary math opens up. I've been saying for years that in this category the economic story is the product, not a story you tell about the product. Labs and hospitals coming out of COVID don't fund science experiments. The CFO wants to know when the thing pays for itself, and the answer has to be a small number a reasonable person believes. Vitestro's small number is three. ## What the human still does The authorization keeps a supervisor trained in phlebotomy in the loop, and it spells out what that person does. They start the session. They confirm the collection tubes fill in the right order. They verify each tube is adequately full when the draw finishes. They answer the alerts. Someone trained cleans the device between patients. Read that list as a time study, because that's what it is. Whatever those duties take per draw sets the real ceiling on how many machines one person can watch. If verifying fill volumes takes ten seconds, three is conservative. If it takes ninety, three is optimistic and the second machine spends its afternoon waiting. The engineer who decided how an operator confirms a tube was writing the pricing model, whether or not anyone in the building described it that way. ## Adults, outpatient, supervised Aletta is authorized for adults in outpatient settings. Those few words draw the map of what Vitestro can sell tomorrow. Inpatient units are outside it. Pediatrics is outside it. So are the hard-stick populations where an experienced phlebotomist earns their reputation, the oncology patients and the elderly with fragile veins, which is where automation would be worth the most and where the evidence is hardest to produce. Each of those expansions is a separate argument with separate clinical data behind it. The FDA published special controls the same day, covering labeling, performance testing, and clinical testing, and a competitor can now meet them through a 510(k). Going first bought Vitestro the definition of the category and paid to pave the road behind it. I worked through that trade in [the Momentis note](/field-notes/2026-07-31-momentis-built-its-own-predicate) in July, and the arithmetic hasn't changed. The money already knew. Vitestro raised $70 million in March, from Labcorp's venture arm, Mayo Clinic, Sutter Health, and Fred Moll, who co-founded Intuitive Surgical. That list is mostly customers, plus the man who has done this before. Dave's take The headline result is that the robot sticks a vein about as well as a trained person, and founders in this category keep mistaking that for the win. I've made the same argument about surgical robots for a decade. The value of the da Vinci was never that it beat the best hands in the room, it was that it grew the population of surgeons who could perform a minimally invasive prostatectomy by orders of magnitude. More surgeons, more patients. Vitestro's version of that number is three, and they had to prove it to a regulator before a single lab could buy it. From Dave’s video library Dave works through which parts of a job belong to a machine and which stay with a person, and what it costs when you draw that line by guessing. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Momentis Built Its Own Predicate, Then Spent It Three Times →](/field-notes/2026-07-31-momentis-built-its-own-predicate) [Caristo Built a Heart Test for the Patients Who Get Sent Home. The Billing Codes Came First. →](/field-notes/2026-08-10-caristo-heart-test-billing-codes-came-first) [Reimagine Wants Your Line Worker to Teach the Robot. Walden Wants to Do It for You. →](/field-notes/2026-08-04-who-teaches-the-robot-reimagine-vs-walden) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # ARPA-H Bet $175 Million on Autonomous Stroke Robots. The Hard Part Is the Room They Land In. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-08-22-arpa-h-stroke-robots-the-room-they-land-in [← All Field Notes](/field-notes) August 22, 2026 Surgical Robotics More than half the people in this country live more than an hour from a hospital that can pull a clot out of their brain. Of the stroke patients who qualify for a mechanical thrombectomy, 12% get one. ARPA-H has put $175.3 million behind one reading of those two numbers, which is that the thing in short supply is the operator. The awards come out of a program the Advanced Research Projects Agency for Health calls Autonomous Interventions and Robotics, run by program director Ileana Hancu. The stated goal is an endovascular robot that can perform a stroke intervention with progressively less direct human input. Siemens Healthineers takes up to $31.1 million and adds $5.4 million of its own, which puts its project at $36.5 million, with Stryker along as a sub-awardee under a partnership the two signed in 2025. Philips North America takes up to $33.7 million. Magnendo, an MIT spinout building magnetic navigation through the vasculature, takes up to $32 million. UC San Diego is in as well. Everybody has five years. ## The shortage ARPA-H named is a building A thrombectomy isn't one machine and one physician. It needs a CT scanner to rule out a bleed before anyone touches the patient, a biplane angio suite, contrast, anesthesia coverage, and a transfer protocol that gets the patient through the door while the tissue is still salvageable. The hospital an hour from the nearest stroke center, the one this program exists to reach, is usually short several of those. Autonomy takes a person out of the room. It doesn't take away the room. That distinction decides what you're selling. A robot that removes the physician from a fully equipped neuro suite is a staffing product for large hospitals. A robot that brings thrombectomy to a community hospital is an infrastructure product, and its real bill of materials includes the imaging, the workflow, and the people who have to be trained to receive it. Those are two different companies. Pick one before you write the indication, because the indication is where the choice hardens. Getting the machine there is its own discipline. We ran a shipping test on a surgical robot once, and the test company dropped the container off a raised forklift. The robot never rattled around inside. One of the actuators tore off its mounts anyway, because the payload sat perfectly still while everything around it stopped very fast. Internal kinetic energy had never crossed my mind. It rewrote the container, the roll-in and roll-out process, and how we secured the machine internally. There's video, and it was terrifying to watch. If your access story depends on hospitals with no robotics service team, shipping is a product requirement and not a line item. ## Two of the four recipients already own the room Siemens Healthineers and Philips sell the imaging a cath lab is built around, and between them they hold the two largest named awards. Stryker, riding in under Siemens, sells the neurovascular devices that travel through the catheter. A lot of this money went to companies whose equipment is already bolted into the rooms where an autonomous thrombectomy would have to work. That's a read on where the difficulty sits. Navigation software that can't see the vessel through the imaging system a hospital already bought is software waiting on a building nobody plans to construct for it. Magnendo is the exception and the more interesting bet, because magnetic navigation changes the physics of getting a wire where it needs to go instead of automating the wire everyone already uses. If you don't own the installed base, changing the physics is the move available to you. For anyone building hardware that has to live in somebody else's building, that's the question worth sitting with. Not whether the thing works on your bench. Whether it works through equipment that someone else specified, bought, and has no budget to replace. ## The part that stays bespoke Threading a catheter from the groin to a cerebral artery is a geometry problem, and geometry is where machines have been winning for twenty years. Deciding whether a particular clot comes out with aspiration or a stent retriever is not. Neither is knowing after the second pass that a third one does more harm than good, or reading the anatomy that says stop and close. That's decision making under uncertainty, and it isn't on anybody's roadmap. The routine half of a procedure automates first because the routine half is the half you can specify. The bespoke half is the one keeping the patient alive. Dave's take Autonomy is the word on the grant, and it's the part I'd bet against. I've argued for years that the design goal is to make the hard parts of a procedure disappear under the physician's own hands, not to take their hands off the work. But ARPA-H picked the right problem to spend five years on, and if this money buys better imaging, better navigation, and a machine that survives the trip to a hospital in the middle of nowhere, that's the win whether or not anything ends up running unattended. From Dave’s video library Dave walks through the BCG numbers behind the jagged frontier, where the same AI tool lifted consultants 40% on one task and left them 19 points worse on another. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [The FDA Authorized a Blood Draw Robot. The Ratio in the Label Is the Product. →](/field-notes/2026-08-21-blood-draw-robot-ratio-in-the-label) [J&J Cleared AI on Its Lung Robot. The Fourth Update in Eighteen Months Matters More. →](/field-notes/2026-08-18-jj-monarch-fourth-update-in-eighteen-months) [Reimagine Wants Your Line Worker to Teach the Robot. Walden Wants to Do It for You. →](/field-notes/2026-08-04-who-teaches-the-robot-reimagine-vs-walden) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # The FDA's TEMPO Pilot Isn't Reviewing These Devices. The Companies Wrote Their Own Limits. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-08-23-fda-tempo-pilot-companies-wrote-their-own-limits [← All Field Notes](/field-notes) August 23, 2026 Regulatory Somebody at Limbic sat down and wrote out every kind of patient who should never be handed their product. The list runs long. Suicidal or homicidal ideation. Moderate to severe dementia. Any psychiatric disorder with psychotic features. Eating disorders of any severity. Pregnancy. Patients 81 or older with a frailty indication. People who don't have a telephone. That list now sits on an FDA webpage, and it's doing a job the FDA isn't doing. ## Enforcement discretion is not approval Limbic is one of four companies in the FDA's TEMPO pilot. The agency updated the participant table on August 21. TEMPO stands for Technology-Enabled Meaningful Patient Outcomes, and it runs on enforcement discretion. The FDA isn't approving anything here. It's saying it doesn't intend to enforce certain requirements, premarket authorization among them, when a device is offered inside a Medicare payment model called ACCESS that pays for chronic care. The agency prints the caveat on the same page: the effectiveness of these devices, for the uses they're participating under, has not been evaluated. The other three participants are SonderMind, with an app that runs alongside therapy for depression and anxiety; Cadence Solutions, with HypertensionOS, which supports clinician-supervised initiation and titration of blood pressure medication; and Dexcom, with a glucose health program. Each row of the table carries a manufacturer, a device name, a clinical use area, an intended use, and contraindications. Limbic's contraindications run most of a paragraph. SonderMind's name four conditions. Cadence and Dexcom both read N/A. ## The sentence you wrote is the whole boundary Nobody at the agency weighed those entries against a body of evidence and decided they were right. Each company wrote its own, and the FDA published it. Under a normal clearance, the indication statement is the settlement of an argument you can point at afterward. Here it's an assertion. So the sentence a company wrote about who its product is for is the only line anyone can hold it to. I've been on the writing end of that sentence. At Galen Robotics we first framed the device as a remote-control telemanipulator, which parked us next to da Vinci and failed. We reframed it as a stabilizer that improves a surgeon's access to anatomy at the end of a long instrument. Same machine, different sentence. The second one cleared as a de novo, and it cleared because the framing changed what evidence we owed. Claim your device makes surgery safer for the patient and you've invited a clinical trial. Claim it improves access and you've invited engineering and usability work, which is cheaper, faster, and yours to control. Read the TEMPO table as the FDA getting easier and you'll take the wrong lesson from it. The bar sits where it always sat. What changed is when you meet it. Under premarket review you learn what the agency thinks of your framing before a patient touches the product. Under enforcement discretion you learn later, from real-world data you collected inside boundaries you drew yourself. ## There is a later, and it arrives with your own data The FDA expects TEMPO participants to file a 510(k) or another marketing submission eventually, built on the data gathered during the pilot. That data can only ever answer the question the indication asked. Draw the indication too wide and you'll spend two years collecting evidence that supports no claim you can defend. Draw it too narrow and you end up with a cleared product nobody wants to pay for. The agency also asks participants for a proposed indications-for-use statement in the very first email they send, before any of the safety data. That ordering is the tell. Framing comes first because everything downstream is measured against it. I'd hold the enforcement-discretion posture loosely too. A regulatory attorney once told me that FDA agreement in a pre-submission doesn't obligate the review team that reads your file later, and I watched it happen. Galen had a study design the agency signed off on in the Q-sub, with resident surgeons performing the procedures. Substantive review came back wanting attendings. We re-ran a 16-subject cadaver study in two weeks because the neurosurgery department at Johns Hopkins handed us a free OR and cadaver heads. That kind of rescue isn't reproducible, and it taught me that a favorable posture from the agency today is exactly that. A posture. None of this is confined to digital health. Any time a regulator or a big customer hands you a lighter way in, whether that's an other-transaction agreement at DoD, a sandbox, or a pilot purchase order that skips the usual qualification, the definition work you skipped doesn't evaporate. It lands on you, and it stays there. Dave's take I'd take the pilot. Reaching patients while you build the evidence is worth a great deal, and the four companies in it clearly agree. What I'd fight about before anyone signs is the wording of that intended-use column, because it's the one decision in the whole arrangement nobody else will make for you and nobody else will catch. From Dave’s video library Dave walks through what the fractional CPO arrangement actually costs, what a week inside it looks like, and why six months is the floor before it produces anything real. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [The Medicare Coverage Decision Now Happens Before Your Pivotal Trial →](/field-notes/2026-08-15-medicare-coverage-decision-before-pivotal-trial) [An FDA Advisory Panel Reads Your File Cold. Grail Finds Out September 23. →](/field-notes/2026-08-08-fda-advisory-panel-reads-your-file-cold) [H.B. Fuller Already Made Surgical Glue. It's Paying £715 Million for the Evidence. →](/field-notes/2026-08-16-hb-fuller-715-million-for-the-evidence) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Every COVID-19 Test EUA Ends December 26. Authorized Was Never the Same as Cleared. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-08-24-covid-test-eua-ends-authorized-is-not-cleared [← All Field Notes](/field-notes) August 24, 2026 Diagnostics Nobody is revoking these tests. On December 26 the three declarations that let the FDA authorize COVID-19 devices at all are terminated, and every emergency use authorization resting on them stops being effective. No test fails a review. The ground under it is removed, and the product goes with it. The Secretary of Health and Human Services signed that determination on June 29, and the notice published in the Federal Register on July 2. It runs two pages. It opens a 180-day window that HHS called a sufficient period for disposition of the product, which is the formal phrase for getting inventory back off the shelves. Companies still selling a COVID-19 test under an EUA have 124 days of that window left. ## What the date actually terminates Three declarations end at once. The first, issued February 4, 2020, covered in vitro diagnostics for detecting SARS-CoV-2. The second, from March 2 of that year, covered personal respiratory protective devices. The third, March 24, covered medical devices broadly, including the improvised products hospitals pressed into service that spring. Ventilators, blood purification systems, remote patient monitors, and screening tools all rode in on that third one. An emergency use authorization does not outlive the declaration that created it. FDA states it plainly: on the effective date, the COVID-19 EUAs for devices listed on its site will no longer be in effect. FDA will publish separate notices for the individual authorizations, but those follow a decision already made. And the schedule has been readable for years. FDA floated the 180-day period in a draft guidance in December 2021, asked industry whether it was long enough to avoid shortages, took the comments, and kept it in the final guidance it issued in March 2023. ## The companies that already moved More than fifty COVID-19 tests now hold traditional marketing authorization, a count the termination notice cites as part of its own reasoning. Most came through 510(k). A few came through De Novo. BioFire's Respiratory Panel 2.1 was granted De Novo in March 2021, Ortho-Clinical's two antibody assays in May 2023, Quidel's Sofia 2 antigen test that December. Each of those created a classification where none existed, so the company that went first wrote the terms everyone behind it would be measured against. Writing those terms starts with how you define the product, and that is the decision I have lived. At Galen we took a 510(k) rejection and came back through De Novo, and the device on the bench never changed. What changed was the definition. A remote-control telemanipulator has to argue it is equivalent to something already cleared. A surgeon-access stabilizer has to prove what it claims about access. The second framing was defensible with engineering and usability evidence. The first one never was, and no amount of paperwork was going to rescue it. The conversions are still landing. ACON cleared a combined RSV, flu, and COVID rapid test on July 10 this year. Visby cleared a flu and COVID test in May, Assure Tech and CorDx in June. Companies are clearing the permanent bar four and five months out from a deadline that has been on the calendar since 2023. ## The two standards were never the same standard Here is the gap that made these transitions take years. To issue an EUA, FDA had to find it reasonable to believe a product may be effective, weighed against an emergency. To clear or approve that same product through the normal route, FDA has to find a reasonable assurance of safety and effectiveness. One is a judgment made under time pressure about a plausible benefit. The other is a finding about the device itself, and it holds after the emergency ends. That gap is about evidence, and no filing strategy closes it. A diagnostics company that started assembling its clinical and analytical package in 2023 is on FDA's list today. A company that treated the EUA as a finished product and kept shipping now has 124 days to produce data it never collected. Dave's take An emergency authorization was always a lease, not a deed, and the lease had a termination clause anyone could read. I say the same thing about patents: the grant is a clock, and the only thing a clock is good for is building something durable before it runs out. The founders on FDA's list this morning spent those years building. The ones who did not are about to find out that a market position you never earned on the permanent standard was inventory with an expiration date. From Dave’s video library Dave works through why the honest clock on a product is proof from a real buyer, and what it costs to go get it. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [iRhythm Paid $287.5 Million to Stop Waiting on Its Own FDA Clearance →](/field-notes/2026-08-11-irhythm-vitalconnect-bought-out-of-the-clearance-clock) [The Medicare Coverage Decision Now Happens Before Your Pivotal Trial →](/field-notes/2026-08-15-medicare-coverage-decision-before-pivotal-trial) [Caristo Built a Heart Test for the Patients Who Get Sent Home. The Billing Codes Came First. →](/field-notes/2026-08-10-caristo-heart-test-billing-codes-came-first) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # XPeng Raised $900 Million for a Humanoid. Its First Customer Is XPeng. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-08-25-xpeng-humanoid-first-customer-is-xpeng [← All Field Notes](/field-notes) August 25, 2026 Industrial Robotics XPeng's robotics unit raised more than $900 million this week. The number to look at is not the money. It's where the robot goes first. IRON, the company's humanoid, is scheduled into XPeng's own stores and campuses when mass production starts at the end of this year. Buyers in China and overseas wait until 2027. That's a sequencing decision, and it's the right one to make. It also tells you what the machine can handle so far. ## What the money buys IDG Capital led the round. Gaorong Ventures joined, with Tencent and Alibaba coming in as strategic investors, at a post-money valuation above $6.3 billion. XPeng keeps roughly 82 percent of the unit. The company's own list of uses runs: hardware and software R&D, training and iterating its physical AI models, high-quality data generation, end-to-end mass production facilities, and global commercial expansion. XPeng delivered 103,295 vehicles in the second quarter of this year. The stated target for IRON is more than a thousand units a month, and a million units by 2030. A car maker raising capital to stand up production lines is the safest bet in the announcement, because building at rate is the thing this company has already proved. There's a real engineering distinction hiding inside the word production. A car is tens of thousands of parts with a mature second-source market for nearly all of them. A dexterous hand is a novel actuator problem where the company designing the robot is often the only supplier that exists. XPeng develops its own chips, actuators and controllers for IRON. That reads as strength on a specification sheet and as concentrated risk on a factory floor, because every yield problem lands on one team with nobody to call. ## The spec sheet is not the job IRON carries 76 degrees of freedom across its body and 21 in each hand, running three of XPeng's Turing chips at up to 2,250 TOPS. Capability on that scale still has to be pointed at something. There's no one robot to rule them all, and I keep having to say it. Humanoids earn their keep in two places: co-bot environments where people and machines share the same floor, and human-shaped workspaces nobody is going to redesign. Everywhere else, the machines that win bounded tasks get engineered around the task. The da Vinci took laparoscopy because it drives more instruments than a surgeon has hands. Put a two-handed humanoid back in that room and you've handed back the limit the robot existed to remove. A store floor and a corporate campus are places XPeng controls. Lighting, flooring, foot traffic, who's allowed to walk up and touch the thing. That's a sensible place to learn, and it's a narrow envelope. The demo reels this whole category runs on were shot in envelopes just like it. ## The part the round does not cover I bought a Magic Link in 1994. Sony built it, engineers who had worked on the Mac designed it, and AT&T ran a data network behind it. Email anywhere, messaging on the go, thirteen years before the iPhone. It cost around $800 plus a monthly fee, and almost nobody bought one. The vision was there and it was still a fancy science project, because there weren't enough people yet who felt the problem it solved. XPeng can build a thousand IRONs a month. What I want to know is what the thousand and first one does, in a building XPeng doesn't own, for a buyer with a budget line and a payback expectation. That's a product question, and production capacity doesn't answer it. The 2027 date is the one to watch, because that is when someone outside the company has to write a check for a job the robot performs. Dave's take Capital is pouring into the manufacturing half of physical AI, and manufacturing was always the half these companies were going to solve. I would trade the entire spec sheet for one bounded task IRON does better than the alternative, in a room nobody staged for the camera. Everything before 2027 is a pilot with nine figures behind it. From Dave’s video library Dave walks through how to tell you're building the wrong thing first, before you spend months finding out. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Reimagine Wants Your Line Worker to Teach the Robot. Walden Wants to Do It for You. →](/field-notes/2026-08-04-who-teaches-the-robot-reimagine-vs-walden) [HII Put $900 Million on the Table for Shipyard Robots. The Qualification Comes First. →](/field-notes/2026-08-09-hii-shipyard-robots-qualification-comes-first) [Hadrian Raised $1.37 Billion for Factories. The Drawings Are Still Your Problem. →](/field-notes/2026-08-12-hadrian-factories-drawings-are-your-problem) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Boston Scientific's Plants Went Quiet. Its Filing Names Shipping, Not Devices. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-08-26-boston-scientific-filing-names-shipping-not-devices [← All Field Notes](/field-notes) August 26, 2026 Medical Devices At two in the afternoon on Tuesday, Boston Scientific sent the day shift home from its Cork plant with full pay. No fire, no failed machine. What had stopped was the network. The company filed a Form 8-K the next morning. I went and read it, because by then the headlines were already calling this a cyberattack and the filing is the only account Boston Scientific has signed its name to. It says the company identified a cybersecurity incident on August 25 affecting certain of its information technology systems, and that the incident "has resulted in a global disruption to the Company's operations." Then comes the line a founder should stop on. The disruptions include "the ability to process and ship customer orders." ## What the filing says, and what it leaves out Four things in that document are worth having straight. The incident was identified on August 25. Order processing and shipping are named. The timeline for full restoration is "not yet known." And the company has "not yet determined whether the incident is reasonably likely to have a material impact." Then there is what the filing leaves out, which is worth as much as what it says. Nothing about implanted or in-use devices. Nothing about patient data. When I am checking a story like this, a gap in the primary document is not something to route around, it is part of the finding. An open investigation is no clean bill of health, and Boston Scientific may say more next week. But on day two, with counsel reviewing every clause, what the company put its name to was a shipping problem. ## Why Item 8.01 is the ordinary route Boston Scientific filed under Item 8.01, "Other Events," and it is tempting to read that as a dodge. It isn't. The SEC's cybersecurity disclosure rules took effect in December 2023, and Item 1.05 is reserved for incidents a company has determined are material. In May 2024 the SEC's Division of Corporation Finance told registrants that where materiality has not been determined, they should disclose under a different item, and file an Item 1.05 within four business days if they later conclude the incident is material. Practice followed the guidance. Debevoise, which has tracked every one of these filings since the rules took effect, counted 29 issuers filing under Item 1.05 and 50 under Item 8.01 as of May 21 this year, with only five doing both. Item 8.01 is the ordinary front door. Read the item number for signal and you are reading the wrong line. ## Why a factory stops when the network does Boston Scientific employs roughly 59,000 people, more than 7,000 of them across three Irish plants in Galway, Clonmel and Cork. The Cork day shift went home at two with full pay. Evening contract staff were paid four hours. Galway and Clonmel staff were offered unpaid or annual leave, and by Wednesday the instruction was to work from home wherever the work allowed it. None of that happens because a robot stopped moving. In a regulated plant, the record is the permission. A device history record, a lot release, a serial number tied to a batch: those are what let finished goods leave the building. Product can be sitting on a shelf, complete and correct, and still be unshippable, because nobody can demonstrate it was released. The quality system is not a filing cabinet you keep for the auditors. It is a gate on revenue, and it runs on the same network as the email. So the question worth carrying back to your own company is narrower than "do we have a firewall." If the ERP is dark on Monday morning, can you release a lot? Is there a paper path, has anyone walked it, and does your own quality manual permit it? The third one is where it bites. A continuity plan written by operations does not override a quality procedure that requires an electronic signature on release. Dave's take Platform dependency is expensive money you never raised and can't negotiate, and I usually aim that at hyperscalers; this week it points inward. You don't own a business you can't switch back on yourself, and the test doesn't care whether the switch sits in somebody else's data center or your own server room. Boston Scientific will restore its systems and ship its backlog, and a company with eight months of runway and one ERP instance doesn't get that assumption. From Dave’s video library Dave reads the study behind a headline about AI outperforming lawyers, and finds the model helped grade its own work. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Cook Medical's Breach Never Touched the Device. It Went Through a Person. →](/field-notes/2026-08-14-cook-medical-breach-went-through-a-person) [BD's Needle Was Out of Tolerance for Twenty Months. Part of the Fix Is a Label. →](/field-notes/2026-08-06-bd-needle-out-of-tolerance-fix-is-a-label) [An FDA Advisory Panel Reads Your File Cold. Grail Finds Out September 23. →](/field-notes/2026-08-08-fda-advisory-panel-reads-your-file-cold) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Uber's Software Decided Who Kept Earning. The Missing Human Cost 825 Million Euros. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-08-27-uber-825-million-for-the-missing-human [← All Field Notes](/field-notes) August 27, 2026 Regulatory The Dutch privacy regulator has no authority over what Uber builds. On August 21 it fined the company 824,990,000 euros anyway, because software decided when a driver stopped earning and no person looked at that decision first. Read it as a product spec. ## What the regulator actually found Between 2018 and 2022, Uber ran software that watched driver behavior and rider ratings. A fraud flag suspended an account temporarily. Persistent low ratings removed it for good, and with it the driver's income from the platform. The Autoriteit Persoonsgegevens, the Netherlands' data protection authority, found two failures in that chain: no meaningful human involvement in a decision that carried consequences that large, and no adequate explanation to drivers of how the system reached it. Article 22 of the GDPR bans decisions made solely by automated processing when the effect on a person is that significant. Monique Verdier, deputy chair of the AP, gave the finding in one sentence: "A computer should not make decisions on its own that have major consequences" for a person. The penalty is 824,990,000 euros, roughly 1.85 percent of Uber's 44.5 billion euro 2025 turnover and about 46 percent of the statutory maximum. It is the second largest GDPR fine on record, behind Ireland's 1.2 billion euro penalty against Meta in 2023. Uber has appealed. ## I once asked my bank whether my manager was a computer Years ago my bank locked me out of my own credit card for making too many payments on it in one month. I was paying it down early, for my own convenience, and the security system read that as fraud. I got a person on the phone and asked them to release the lock. They told me they could not override the computer's decision. So I asked whether their manager was a computer. The rep took offense. No sir, I am the team lead here. Fine, I said, but you just told me you cannot override the computer, so the computer is calling the shots and you are not. Perhaps think about what that means for your career. That phone call is the Uber ruling in miniature. Somewhere inside that bank there was a documented human review step. On paper a person could reverse the lock. In practice the person on the line had no button, no authority, and no path to anyone who had either. Pedro Domingos framed the general case in _The Master Algorithm_: the danger was never that computers would get smart enough to take over, it is that they are not smart at all and already run things. If a human inside your product cannot practically reverse a decision the software made, you did not build an automation. You built an abdication. The override written into the design document does not count. The only one that counts is the one a tired person can actually reach. ## Who this bill is addressed to Article 22 is not a clearance you apply for. There is no premarket review, no pathway, no predicate, no reviewer to negotiate with. It applies to a product already shipping, the moment that product makes an automated call with significant consequences for a person in Europe. For hardware founders used to watching a safety regulator's clock, that is a different clock entirely, and it is already running. So walk your own system and list the places it decides something about a named human being with nobody in the loop. A diagnostic that scores a scan as negative and sends the patient home. A fleet controller that pulls a vehicle out of service and a driver off shift. An industrial cell that locks out an operator over a safety flag the operator disputes. A monitoring rule that downgrades an alarm at three in the morning. Every one of those is a consequence for a person, decided by software, on a schedule your safety regulator may never inspect. Automating a bounded task is fine, and I do it every day. The design work is where you draw the boundary and how fast a person can cross back over it. An override nobody can find is a deflection funnel wearing a compliance label, and the AP just showed what that costs when a regulator goes looking for it. Dave's take I keep noticing that the regulator here has no safety mandate at all. Uber's vehicle standards and background checks were never on trial: a data protection agency looked at one design decision, the missing human in the loop, and priced it near a billion euros. Go find every place in your own product where a person is nominally in charge and cannot override the machine, and fix it before somebody else puts a number on it. From Dave’s video library Dave walks through what an AI copilot changed for the newest support reps, and where the human override has to stay real. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [The FDA's TEMPO Pilot Isn't Reviewing These Devices. The Companies Wrote Their Own Limits. →](/field-notes/2026-08-23-fda-tempo-pilot-companies-wrote-their-own-limits) [Cook Medical's Breach Never Touched the Device. It Went Through a Person. →](/field-notes/2026-08-14-cook-medical-breach-went-through-a-person) [ARPA-H Bet $175 Million on Autonomous Stroke Robots. The Hard Part Is the Room They Land In. →](/field-notes/2026-08-22-arpa-h-stroke-robots-the-room-they-land-in) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Bedrock’s Excavators Dig With Nobody in the Cab. Bedrock Doesn’t Build Excavators. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-08-28-bedrock-retrofit-excavators-nobody-in-the-cab [← All Field Notes](/field-notes) August 28, 2026 Industrial Robotics Bedrock Robotics has excavators moving dirt on live job sites in Texas and Nevada with nobody in the cab. The company has raised $350 million, its founders came out of Waymo, and the line everyone is quoting is the empty seat. The decision worth studying is a different one. Bedrock does not build excavators. It sells a retrofit that goes onto a machine a contractor already owns, installed in a single day, with nothing on that machine permanently changed. ## The form factor was a money decision before it was an engineering one Start with the buyer. Excavators run about a quarter of a contractor's fleet, and an operator takes roughly five years to get good at one. A company selling autonomous earthmoving has two ways to reach that contractor. Build a new machine and ask them to buy a second excavator to do work the first one already does. Or leave their iron alone and sell them the operator hours they cannot hire. Bedrock took the second road, and the engineering follows from it. Same-day install. No permanent modification. The machine goes back into the fleet and a person can still climb in and run it. Sundt Construction has one working at a Nevada water treatment plant. Zachry Construction is running one on 1.2 million cubic yards of civil sitework, and Champion Site Prep on a multi-million cubic yard earthwork job in Texas. None of those contractors had to change what they own in order to say yes. I learned this on hospital capital equipment, and the lesson travels. Post-COVID, hospital margins have no room for science experiments. The economics have to make sense on day one or you never get into the first hospital, never mind the hundredth. My own multi-field surgical robot was built for procedures no other robot targeted, which sounded like an open lane until we priced it. Those procedures carried lower reimbursement, so our economics had to come in lower to work at all. The gap we found in the market was a gap because the money did not work there yet. ## The empty cab came a year after the capability did The machines could do this work before anyone was allowed to walk away from them. Bedrock ran them for a year on active job sites with a human supervising, on top of a model trained across tens of thousands of hours in the field, and only then took the person out. That gap between working and shipping is the discipline, and it is where hard-tech companies get impatient. Look at how narrow the claim is even now. A site manager sets the plan and the machine executes inside it. It runs in pre-approved site conditions. It carries 360 degrees of sensor coverage and stops on its own when a person or another object gets close. It pauses on a fault. When it cannot make progress, The Robot Report quotes CTO Kevin Peterson saying "it'll call home if it gets stuck," and there is a documented process waiting on the other end of that call. Bedrock tells a contractor which jobs the machine handles and what happens when it cannot. That is a smaller promise than the headline suggests, and it is one a superintendent can plan a week around. ## Fit the work people already do Years ago we bolted a navigation system onto a Stryker sagittal saw bought off eBay. That was a University of Nebraska Medical Center project, more than a decade back. The eBay saw sounds like a budget decision. The reason was the workflow. A total knee replacement runs on a cutting jig and a saw, five cuts and the surgeon is done. The robotic platforms of that era used a burr, which grinds bone away, takes far longer, and asked a surgeon trained on five cuts to work an unfamiliar way. Putting navigation on the saw already in their hand meant the new capability arrived inside a motion they knew cold. The moment a product adds friction or makes someone relearn their job, you have created a problem rather than solved one. The excavator operator is not the buyer here, but a superintendent and a fleet manager are, and both of them have a way of running a site that took years to settle. A kit that installs in a morning and leaves the seat usable asks neither of them to change it. Dave's take Founders want to open with the machine. The machine is downstream. Pick the buyer, run their payback math on the first unit they would ever purchase, and let that math tell you what form the product takes, because once engineering is underway you have quietly locked your price, your factory and your customer. Bedrock made that call early enough that its hardest engineering problem is the part the customer never has to finance. From Dave’s video library Dave on the one lesson from forty products: the conversation about who buys and what has to be true belongs before the engineering conversation, not after it. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [XPeng Raised $900 Million for a Humanoid. Its First Customer Is XPeng. →](/field-notes/2026-08-25-xpeng-humanoid-first-customer-is-xpeng) [HII Put $900 Million on the Table for Shipyard Robots. The Qualification Comes First. →](/field-notes/2026-08-09-hii-shipyard-robots-qualification-comes-first) [Hadrian Raised $1.37 Billion for Factories. The Drawings Are Still Your Problem. →](/field-notes/2026-08-12-hadrian-factories-drawings-are-your-problem) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Medtronic Pulled These Capsules Twice for the Same Failure. Only the Cause Changed. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-08-29-medtronic-bravo-recalled-twice-same-failure [← All Field Notes](/field-notes) August 29, 2026 Medical Devices A gastroenterologist threads a catheter down a patient’s esophagus, presses a small capsule against the wall, and lets go. Two things then have to happen. The capsule stays. The catheter comes out without it. On August 27 the FDA issued an early alert saying that with Medtronic’s Bravo CF delivery device, sometimes neither one happens. The capsule fails to attach, or it fails to release. As of August 10, Medtronic had reported 184 serious injuries and no deaths. The risks the FDA lists include perforation of the esophagus, obstruction of the airway, hemorrhage, and foreign bodies remaining in the patient. Medtronic recalled the same product numbers for the same failure in June 2025. The FDA’s notice this week says the underlying cause is a different one. ## Two recalls, one failure, some of the same lots The 2025 recall letter went out on June 3, with an update on June 16. Medtronic told customers the capsule might not attach or detach because adhesive had been misapplied during manufacturing. Thirty-three serious injuries, no deaths. The FDA classified it a Class I recall in January of this year, its most serious category, meaning continued use could cause serious injury or death. The 2026 letter went to customers on August 6. This time Medtronic traced the failure to the component that controls deployment of the capsule. Same product numbers, FGS-0635 and FGS-0636. Same failure, described to the same customers. And according to the FDA, a subset of the lots covered now had already been covered by the June 2025 notification, which means some of these boxes came off a shelf twice for two different reasons. The list of risks is the same in both notices, in the same order, down to the wording: aspiration, perforation of the esophagus, obstruction of the airway, hemorrhage, laceration, delay in diagnosis, foreign bodies remaining in the patient. The injury count went from 33 to 184. ## A closed corrective action is not a closed hazard From inside a quality system the sequence is orderly. Complaints arrive. You investigate, trace the failure to a cause, fix the cause, and verify the fix. Verification asks a narrow question: is the defect gone? Devices built after the change do not show it. The corrective action closes. The question verification does not ask is whether the harm stopped. Those two come apart whenever a single hazardous situation has more than one route into it. A capsule that will not release from its catheter is the hazardous situation. Adhesive was one road there. The deployment component was another. Close out the first road and the hazard is still standing, and the only place it shows up is the complaint file, where it reads as a new problem instead of the same one. That one sentence from the FDA, same failure and a different cause, is the most useful thing in the notice, and it is the sentence I would carry back to my own complaint data. ## An instruction is the weakest control you can write The notice adds a detail worth sitting with. Medtronic said the deployment problem can be made worse when the delivery device is not kept straight, as the User Guide instructs. Medtronic is probably right about that. It is also the weakest kind of risk control available. Design the hazard out, add a protective measure, or write it down for the user, and the order matters because it tracks how much of the work you are handing to somebody else. Information for safety sits last on that list for a reason. The person receiving the instruction is holding a catheter in a patient’s throat, watching a screen, and doing this procedure a handful of times a week. If your control depends on a clinician executing your written instructions correctly during a procedure, the risk moved onto them, and you wrote down where it went. That is a defensible thing to do when nothing else is available. It is a bad thing to mistake for a fix. Dave’s take The risk standards work. Compare device injuries over the past decade to what the seventies and eighties looked like and it is remarkable how safe these complicated machines have become, and that is not an accident. But a standard only earns that record when you ask your own data the right question, and the right question is not whether you found something to fix, it is whether the injuries stopped. From Dave’s video library Dave walks through pulling scattered records into a single number you can decide from. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [The FDA Authorized a Blood Draw Robot. The Ratio in the Label Is the Product. →](/field-notes/2026-08-21-blood-draw-robot-ratio-in-the-label) [The FDA's TEMPO Pilot Isn't Reviewing These Devices. The Companies Wrote Their Own Limits. →](/field-notes/2026-08-23-fda-tempo-pilot-companies-wrote-their-own-limits) [J&J Cleared AI on Its Lung Robot. The Fourth Update in Eighteen Months Matters More. →](/field-notes/2026-08-18-jj-monarch-fourth-update-in-eighteen-months) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # The FDA Wants to Know How You'd Test a Device That Answers Differently Every Time | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-08-30-fda-genai-devices-answer-differently-every-time [← All Field Notes](/field-notes) August 30, 2026 Regulatory Run the test twice, get the same answer twice. That assumption sits underneath every verification plan I have ever signed off on. Put a generative model inside the product and the assumption breaks. On August 18 the FDA published a discussion paper admitting it has no settled method to put in its place. The paper arrived with an open docket asking the industry for help. The paper is called Considerations for the Regulation of Generative AI-Enabled Medical Devices, out of the Digital Health Center of Excellence inside CDRH. Nothing in it is draft guidance and none of it is binding policy. What the FDA put on the table is 26 numbered questions, a comment deadline of October 19, and a docket number to file against: FDA-2026-N-7874. Read it even if you'll never file with the FDA. Nothing about the verification problem it describes is specific to medicine. If your product decides something, and somebody gets hurt when it decides wrong, these questions are coming for you too. The FDA is just the first regulator to write its confusion down in public. ## Two axes, not one The agency sorts these devices on two axes at once. The first is how independent the thing is, running from handing a clinician information that carries no recommendation all the way to acting on its own. The second is how bad the consequence is when the output is wrong. Teams argue about the first axis constantly. The second one they write down almost never, and that's backwards, because the consequence is where the evidence burden comes from. A warehouse robot that suggests a pick path and a warehouse robot that executes it can run identical software. The reviewer is weighing the consequence profile, and those two robots don't have the same one. "Assistive" is also a status you have to keep earning. Your roadmap is under constant pressure to break it. Each autonomy feature your sales team wants pushes the product further along the first axis, and somebody should have to approve that move in writing. ## Competency instead of a verdict For premarket work the paper sketches benchmarking a device across ten elements, covering safety behaviors, clinical proficiency, generalizability, and autonomous conduct. The benchmark then gets confirmed clinically, somewhere between a retrospective study and a prospective one. After launch the framework leans harder on ongoing monitoring than the agency does for a fixed-function device, on the grounds that the outputs vary. Benchmark-plus-monitoring is a different shape of work from what most hardware teams budget for. A test plan that produces a verdict finishes. A test plan that produces a score has to be re-run, and somebody has to own that score for as long as the product ships. I've watched founders model regulatory work as a submission-shaped expense that ends the day the letter arrives. If the thing you shipped keeps changing its own behavior, the expense doesn't end, and the burn model is wrong by whatever it costs to staff the difference. ## The part that decides is somebody else's code The paper also floats a voluntary Foundation Model Device Master File. The company that built the underlying model files its confidential information once, and your submission points at it. Strip the paperwork away and the FDA is acknowledging something founders already know and rarely price: the component making the call inside your device was built by a supplier you don't control and can't fully inspect. I've paid for that arrangement before. We sold internet software that ran on MacTCP, the Mac's TCP/IP stack, written originally by University of Michigan students and by then maintained by nobody at Apple. System 8 could page MacTCP out when virtual memory was on, and the stack threw up a debugger statement those students had left in the code. It wasn't even a crash. If you had MacsBug installed you told it to continue and you were fine. Our customers were defense contractors running a hundred thousand copies, and what they saw was a bomb icon, and it was all our fault. None of that was our code. Every support call still said our product crashed. A foundation model is the same bargain with more zeros. You're shipping a dependency you didn't write, can't read, and can't patch, and in front of the customer and the reviewer the debt is yours. The master file gives that problem a paperwork channel. It doesn't solve it. Dave's take I'd file a comment. There are 26 questions open until October 19, and the operating reality inside the eventual framework will be whatever the companies who bothered to file put there. Everyone else gets to comply with what the people who showed up decided. Filing costs you an afternoon. You'll be living inside the answer for the next decade. From Dave’s video library Dave walks through how to tell a checkable AI claim from one nobody can check, using the model the government supposedly banned. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [The FDA's TEMPO Pilot Isn't Reviewing These Devices. The Companies Wrote Their Own Limits. →](/field-notes/2026-08-23-fda-tempo-pilot-companies-wrote-their-own-limits) [The FDA Authorized a Blood Draw Robot. The Ratio in the Label Is the Product. →](/field-notes/2026-08-21-blood-draw-robot-ratio-in-the-label) [J&J Cleared AI on Its Lung Robot. The Fourth Update in Eighteen Months Matters More. →](/field-notes/2026-08-18-jj-monarch-fourth-update-in-eighteen-months) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Oura Says It Isn't a Medical Device. The Ad Compared It to a Sleep Lab. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-08-31-oura-not-a-medical-device-ad-cited-the-sleep-lab [← All Field Notes](/field-notes) August 31, 2026 Medical Devices Somebody at Oura approved a sentence putting a percentage next to the words clinical sleep lab. That sentence is what a federal court is now being asked to rule on. The algorithm behind it is barely in dispute. The case is Surber v. Oura, filed this month in the Northern District of California by the Clarkson Law Firm on behalf of a California buyer of an Oura Ring 4. It challenges three pieces of marketing: "Built for accuracy," "Unparalleled Accuracy," and "95% Sleep Staging Accuracy compared to clinical sleep lab." Oura told MobiHealthNews it stands behind its science and will defend the allegations. Call it a health wearable that oversold its AI and you have the easy reading. I don't buy it. The model is doing roughly what Oura says it does. What went wrong here was a product decision, and it went wrong long before any lawyer read the website. ## Both sides describe the same ring Set the complaint's technical argument beside Oura's response and they agree about the hardware. Sleep gets defined in the brain, the plaintiff argues, through electrical activity, eye movements and muscle tone, read by electrodes on the scalp and leads placed around the eyes. A ring on your finger reaches none of that. Oura doesn't say otherwise. Its own statement describes a device that estimates sleep stages from heart rate, movement, breathing and temperature, and calls itself neither a medical device nor a substitute for a clinical sleep study. What they are fighting about is one number and the phrase bolted to it. For its evidence the complaint leans on a 2025 study in Nature that put the ring's sleep-stage agreement near 53 percent across 45 nights. I'd want to read that paper before I treated the figure as settled. Forty-five nights is a small pile of data, and a number this convenient for the people suing deserves the same scrutiny as the claim it targets. Oura answers with more than 1,200 nights of collected data, which is a company number and gets read the same way. Neither figure has been tested by anyone except the party citing it. ## Not a medical device answers a question nobody asked Oura's own framing, that the ring was never a medical device, is also its defense, and that framing is the part worth sitting with. Plenty of founders are running the same play. Staying in general wellness is a real strategy. Keep off disease claims, keep the risk profile low, and the FDA has better things to do than review your app. Where founders get it wrong is treating that as the finish line. It's a boundary on one agency's attention, and that's all it is. It decides which regulator reads your claims before you ship. Nothing about it touches whether those claims are actionable afterward. California consumer protection law, express and implied warranty, and fraud by misrepresentation are all pleaded in this complaint, and not one of them turns on how the FDA classified the device. Then there's the comparator. "Compared to clinical sleep lab" reaches over and borrows the authority of polysomnography, the exact standard Oura says it should not be measured against. Declining the jurisdiction is allowed. Cite its gold standard in your advertising while you decline it, and you have volunteered for a different forum. ## The expensive word, in a different room That borrowed authority has a price. The same trade shows up in a room with no jury in it. In a 510(k), the most expensive word a founder can write is "better." Substantial equivalence means equivalent. Frame a technical difference as an improvement and you have made a superiority claim, which can push you out of the pathway the whole submission was built around. The disciplined version says the difference raises no new question of safety or effectiveness, so the device can be evaluated inside the same bounds. A comparison claim changes what you are asserting, and it summons a reviewer nobody budgeted for. For Oura, that reviewer is a federal judge. Oura's calendar makes that expensive. The company raised $900 million at a valuation around $11 billion, filed confidentially for an IPO in May, and says it is on pace to pass five million paid members. A company at that stage has an S-1 to write, and its accuracy claims are now part of what has to be disclosed. Dave's take Go read every accuracy number on your own site and find out who signed off on it. If no regulator is going to check those claims before they ship, that doesn't leave them unchecked. It means the first careful reader is opposing counsel. From Dave’s video library Dave walks through how a carefully worded company number got repeated as something bigger, and who ends up owning the gap. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Every COVID-19 Test EUA Ends December 26. Authorized Was Never the Same as Cleared. →](/field-notes/2026-08-24-covid-test-eua-ends-authorized-is-not-cleared) [The FDA Authorized a Blood Draw Robot. The Ratio in the Label Is the Product. →](/field-notes/2026-08-21-blood-draw-robot-ratio-in-the-label) [Abbott Makes the Glucose Number. Google Charges for What to Do About It. →](/field-notes/2026-08-17-abbott-lingo-google-owns-the-coaching-layer) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Boston Scientific Is Pulling Catheters Its Complaint Data Never Flagged | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-09-01-boston-scientific-catheters-complaint-data-never-flagged [← All Field Notes](/field-notes) September 1, 2026 Medical Devices Boston Scientific’s Imager II catheters are coming off the market, and the complaint data is not why. The FDA published an early alert on Friday saying no increase in tip-detachment complaints turned up for the affected lots. They are being removed because the company built them outside the approved design specification. The Imager II is a sterile, single-use catheter that doctors thread into the peripheral vasculature to image it. Boston Scientific determined that the tips on these lots were made with reduced levels of stabilizing agents, which raises the risk that a tip degrades and comes apart. The FDA describes two outcomes. Most of the time a detached tip costs a delay while somebody exchanges the device. In the remote case a fragment lodges in a vessel, cuts off blood flow, and takes an organ with it. As of July 16 the company had reported two serious injuries and no deaths. ## A gap against the spec, with no complaint behind it Somebody compared the units coming off the line against the specification on file and found a gap. Under the rules, that gap is grounds for removal by itself, with no complaint signal required. A design specification is the description of the device the agency reviewed, and what it reviewed included how the device gets made. A manufacturer building to a different specification is outside its own clearance whether or not a patient ever notices. The product can be performing acceptably in every hospital that uses it, and the removal still happens. ## Getting the catheters back out of the hospitals Boston Scientific’s letter went to customers on August 24, and most of the FDA notice is retrieval logistics: stop using the product, segregate it, fill in a tracking form, send the catheters back. The notice then warns that hospitals often discard the outer carton and shelve the catheters in the inner pouch, and that the code printed on the inner pouch is a different number from the one on the box. The tracking form carries only the outer codes. The catheters ship in five-packs, and returns have to be counted as individual units. So the person standing in a supply room with that form has to match a pouch in their hand against a number that was never printed on it. Nothing about that is a manufacturing failure. It is a labeling decision, made years earlier by people solving a different problem. The version of this I lived through was in software. When Apple changed the Mac file system and deprecated the old primitives, my company found more than 2,800 separate instances of file open and save calls in our own codebase. They were not consistent enough for a search and replace. One engineer, Jim Bates, spent about six months cataloging every variation and then updating every one of them. Companies that were less thorough shipped breakage the day the old calls stopped working. ## The line is being discontinued anyway Boston Scientific has also decided to stop selling the Imager II. The retrieval happens regardless, and every unit already sitting in a hospital supply room still has to be accounted for. The alert does not explain how the stabilizing agents ended up low, and I have no idea what happened inside that plant. I do know what a product on its way out looks like from the inside. The strongest engineers are on the next platform. A supplier change gets a lighter review than it would have gotten three years earlier. The process running the line has not had a hard look in a while, and the line is still shipping. If you build anything in lots (devices, diagnostic assays, defense hardware, battery packs), the specification question is cheap to answer now and expensive later. For every lot you have shipped, can you name the specification it was built to? Dave’s take Can somebody holding one of your units in a stockroom tell you which lot it came from? That is the question I would go ask today, because it is the one nobody can answer in the moment. The answer was settled long ago, probably in a packaging meeting nobody remembers. From Dave’s video library Dave works through what it costs to change one thing in software you shipped fast. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Medtronic Pulled These Capsules Twice for the Same Failure. Only the Cause Changed. →](/field-notes/2026-08-29-medtronic-bravo-recalled-twice-same-failure) [Boston Scientific's Plants Went Quiet. Its Filing Names Shipping, Not Devices. →](/field-notes/2026-08-26-boston-scientific-filing-names-shipping-not-devices) [The FDA Authorized a Blood Draw Robot. The Ratio in the Label Is the Product. →](/field-notes/2026-08-21-blood-draw-robot-ratio-in-the-label) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Two in Ten Storage Projects Now Treat the 30% Tax Credit as Optional | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-09-02-storage-projects-treat-tax-credit-as-optional [← All Field Notes](/field-notes) September 2, 2026 Climate Hardware Tony Song buys battery units for a living, and he carries three prices in his head. The Chinese ones start around $70 to $80 per kilowatt-hour, and by the time tariffs, shipping, and service costs are added they land near $120. Add ten or twenty dollars to that for units built elsewhere, most of which still fail the new foreign-entity rules because their supply chains run back through China anyway. American-made units are the outlier at $160 to $180. Song runs engineering, procurement, and construction at the storage developer GridStor, and he told Latitude Media that paying the American premium to qualify for the 30% investment tax credit is still his best option. For now. Song will change his mind at a number. If the Chinese units keep falling toward the high $50s or low $60s, they land under $100 in the United States even after tariffs, and the credit stops paying for the premium it was supposed to offset. He is not the only one running that arithmetic. Ravi Manghani, who runs strategic storage sourcing at the procurement platform Anza Renewables, told Latitude Media that roughly two projects in ten now treat proceeding without the credit as their baseline plan. He called it a math problem. ## What it costs to claim the credit Manghani's point is that a developer cannot subtract 30% from a project's cost and call it done. The credit requires prevailing wages, which in states like California means union labor at union rates. It requires documented foreign-entity compliance, and a developer who plans to sell the credit for cash on the transferability market pays legal fees to prove that compliance to the buyer. Add back the equipment premium that started the exercise. The 30% is a gross number, and all of those costs come out of it before the developer sees a dollar. ## The vagueness costs more than the premium A year after the One Big Beautiful Bill attached foreign-entity restrictions to the credit, the IRS still has not said what counts as Chinese control, or how much foreign-held debt disqualifies a developer. Manghani said the market is waiting. A developer signing equipment contracts this month is committing capital against a rule nobody has written, and that project will be audited later under whatever the rule turns out to say. Capital prices uncertainty the way it prices everything else. A credit that can vanish on a definition nobody has published yet is not a foundation to build on. It is a variable, and it belongs in the model where the other variables already live: the wind forecast, the commodity curve, the interconnection queue. I have watched a version of this play out in medical devices, where the payer is a hospital and the instrument is a reimbursement code. When I work with a founder building a surgical robot, the first thing I ask for is the payback model: what one procedure pays, and how many procedures it takes before the hospital has covered the cost of the box. I want the CFO and the coding people in the room before the surgeons fall in love with the device, because they are the ones who decide whether the money shows up. The economic story is the product, and that holds whether the money comes from a hospital's finance office or from the IRS. ## The compliant option is also the newer factory There is a cost the price per kilowatt-hour does not show. Samsung SDI and LG Energy will not have their U.S. capacity running before late 2026 or early 2027, so a developer who needs a project energized sooner buys from the mature Chinese supply regardless of what the credit pays. Song raised the problem facing developers who can wait: a new factory usually takes a year or two to work out its quality problems, so the first domestic cells off the line are early production. The premium that unlocks the credit also buys cells from a line that has not settled down yet. So a developer is weighing two different risks against a rule that does not exist, which is a harder problem than picking the cheaper unit. Two in ten is a small share. But GridStor used to pursue the credit as a matter of course and now weighs it project by project, and it is making that call while the guidance is still unwritten. Making it afterward would cost nothing and prove nothing. Dave's take Run the project twice. Once with the incentive and once without it, and let the version without it decide whether you build. If the only version that works depends on a rule the IRS has not finished writing, then what you have is a financing instrument with a project attached to it. From Dave’s video library Dave shows how to put a number on something that looks unmeasurable, and how little data a real decision needs. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [GM Walked Away From This Battery Plant Twenty Months Before It Opened →](/field-notes/2026-08-20-gm-walked-away-from-this-battery-plant) [Hadrian Raised $1.37 Billion for Factories. The Drawings Are Still Your Problem. →](/field-notes/2026-08-12-hadrian-factories-drawings-are-your-problem) [The Navy Wants Sixteen Times the Tomahawks. The Seven Years Is What Buys Them. →](/field-notes/2026-08-19-navy-tomahawk-seven-years-buys-the-rate) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Kardium's Cleanroom Builds 500 a Year. The Business Needs 30,000. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-09-03-kardium-cleanroom-500-a-year-needs-30000 [← All Field Notes](/field-notes) September 3, 2026 Medical Devices Kardium's cleanroom in Vancouver can build 500 Globe catheters a year. The company is spending $125 million to make that number 30,000. The distance between those two figures decides whether Kardium competes in pulsed field ablation, and the FDA approval it won twelve months ago barely touches it. MedTech Dive reported Monday that $31.25 million of the expansion comes from the Canadian government, through a Strategic Response Fund set up for industries hit by tariffs and market disruption. Kardium had already closed $250 million in July 2025 to commercialize the Globe, and $104 million the year before that. When the FDA approved the system in September 2025, the company said it would expand manufacturing and stand up teams for the launch. The $125 million is what that sentence costs. ## The approval came with dates on it Kardium's regulatory work was heavy and it was legible. The Globe went through premarket approval, the long path, with separate 510(k) clearances for the introducer sheath and the mapping software. The catheter carries a spherical array of 122 electrodes that maps the heart and ablates tissue in one pass. The pivotal data MedTech Dive reported at approval showed 78 percent freedom from atrial arrhythmia at one year in paroxysmal patients. Every piece of that work had a date and a submission behind it. A factory has none of that. Going from 500 units to 30,000 is a different plant. It needs tooling that doesn't exist yet, suppliers who have never shipped you volume, sterilization validations repeated at the new scale, and a cleanroom footprint built and qualified before one sellable unit comes off it. Meanwhile Boston Scientific sells Farapulse, Medtronic sells Pulseselect and Affera, and Johnson & Johnson sells Varipulse. Against that, a 500-unit line is a pilot. ## A shipping test I still think about At Galen we ran a shipping test on the surgical robot. The ship-test company dropped the shipping container off a raised forklift, and I'm not naming names. The robot didn't bang around inside at all. The container stopped when it hit the ground. The mass inside kept going, and one of the actuators tore itself off its mounts. I never would have thought about internal kinetic energy issues. There is video, and it was terrifying to watch. That one test sent us back to the container, the roll-in and roll-out process, and how the robot was secured inside. Weeks of work came out of a failure mode that was in nobody's requirements document. And it was one test, on one shipping lane. Multiply it by a consumable at 30,000 units a year, where every supplier qualification and every sterilization cycle at volume is another chance to find one of those. What I took from it is to treat the plant as a product with its own schedule and its own line in the raise. Clinical and regulatory work comes with dates. Manufacturing surprises don't, and a plan built only from the dated work quietly assumes they won't happen. ## Who pays for the plant The $250 million Kardium raised for commercialization didn't cover the plant. The $31.25 million that closed the gap came from a fund built to answer tariffs, and Adrian Dix, British Columbia's minister of jobs and economic growth, framed the support as a response to "unprecedented attacks on our sovereignty." The number that got the check signed was 1,000 skilled jobs in British Columbia. Industrial-policy money buys domestic capacity and jobs, on terms no clinical investor offers. It is worth finding before you need it. Write down the number of units you have to build in year three, then work out who besides your investors is paid to care about that number. In Kardium's case the answer was a provincial government counting jobs. Dave's take Money is runway, not a moat. It's a clock ticking against the milestones you promised investors, and $250 million sets that clock loud and short. What Kardium spent the first year of it on tells me more than the raise did: it went and found out what 30,000 units a year costs to build, then went and got that money too. Most of the roadmaps I read put the clearance date at the top. The build rate belongs there. From Dave’s video library Dave walks through the five numbers that tell you whether a business is working, and how they read together. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Boston Scientific's Plants Went Quiet. Its Filing Names Shipping, Not Devices. →](/field-notes/2026-08-26-boston-scientific-filing-names-shipping-not-devices) [GM Walked Away From This Battery Plant Twenty Months Before It Opened →](/field-notes/2026-08-20-gm-walked-away-from-this-battery-plant) [The Navy Wants Sixteen Times the Tomahawks. The Seven Years Is What Buys Them. →](/field-notes/2026-08-19-navy-tomahawk-seven-years-buys-the-rate) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # FDA Reclassified This Cancer Test Category. Fewer Than Five People Commented. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-09-04-fda-reclassified-cancer-tests-five-comments [← All Field Notes](/field-notes) September 4, 2026 Diagnostics An FDA order that changes who can compete in a cancer-test category takes effect on September 16. Fewer than five people commented on it. The comment period closed in August of last year, and everyone who wrote in was in favor. The order is Federal Register 2026-16727, published August 17. It covers in situ hybridization test systems indicated for use with an approved oncology drug. These are the probe-based assays that tell an oncologist whether a tumor carries the genetic change a particular drug goes after. The order moves them out of Class III and into Class II. Four product codes are affected: NYQ, MVD, OWE and PNK. A company bringing one of these tests to market no longer files a PMA. It files a 510(k) against a new regulation, 21 CFR 864.1890. If you already hold one of those PMAs, nothing happens to you on September 16. FDA says approved devices may keep marketing under the approval already issued, with no additional authorization required. Changes that could affect safety or effectiveness now go in as a 510(k) instead of a PMA supplement. ## FDA built the new door out of the incumbents' data FDA points to the PMA data already sitting in its files, to published literature on what it calls a longstanding and well-understood technology, and to these tests not having thrown significant postmarket safety signals. The agency expects the change to cut regulatory burden, shorten review, and bring more manufacturers into the category. The incumbents built that record to get through Class III. FDA used it to decide Class III was no longer necessary. I've been through a version of this from the other side. At Galen we were routed to a de novo, and near the end of the process FDA asked us what the new regulation describing our device ought to say. We conferred, wrote a paragraph, and sent it in. The regulation came back word for word what we had submitted. We set the standard the category had to match, and we knew the exact sentence we'd set it with, which is the part the ISH order takes away: the incumbents here didn't get to pick the words. ## Class II here is not a shortcut The new regulation carries nine design verification and validation special controls, and they're specific. Analytical sensitivity against normal karyotype specimens. Analytical specificity covering probe specificity, interference and cross-reactivity. Precision across multiple reagent lots, operators, instruments, sites and reader assessments, with data near the clinical thresholds. Robustness across the tolerance ranges of critical test parameters. Linearity. Specimen stability. Clinical performance in specimens from the intended population. Three labeling controls sit on top of that. So the evidence burden didn't go anywhere, it moved. A company used to work that burden out privately with a reviewer over the course of a PMA. Now any competitor can pull up 21 CFR 864.1890, read the list, and cost it out before deciding whether to enter. ## The docket ran for fifteen months FDA proposed the change in June 2025. Comments closed that August, the final order landed a year after that, and it takes effect thirty days later. Fifteen months of public notice, and fewer than five people commented on a rule resetting the entry cost for a whole class of oncology tests. Most of the few who did comment came from device companies themselves. Fifteen months of notice is the part worth carrying into other categories, because the pattern travels. Any company whose competitive story leans on a regulatory approval is exposed to the same move, and the warning shows up in a docket months before it shows up anywhere a founder is looking. The product codes covering your device are public. Every proposed rule that touches them is published before it takes effect. Dave's take When I argue that a de novo makes you the predicate, I've always attached a condition to it: the predicate position holds until the category gets crowded enough that defending the standard costs more than it returns. This rule is a faster route to that same crowding. FDA didn't wait for fast followers to show up, it opened the door itself, and it used the incumbents' own submissions as the reason. If your plan for years three through six assumes the pathway you walked is the pathway your competitors will have to walk, that assumption belongs on your risk register. From Dave’s video library Dave checks a claimed time-saving number for AI scribes against the study behind it, and finds the measured result is far smaller than the claim. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Every COVID-19 Test EUA Ends December 26. Authorized Was Never the Same as Cleared. →](/field-notes/2026-08-24-covid-test-eua-ends-authorized-is-not-cleared) [The FDA's TEMPO Pilot Isn't Reviewing These Devices. The Companies Wrote Their Own Limits. →](/field-notes/2026-08-23-fda-tempo-pilot-companies-wrote-their-own-limits) [The Medicare Coverage Decision Now Happens Before Your Pivotal Trial →](/field-notes/2026-08-15-medicare-coverage-decision-before-pivotal-trial) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Medtronic Paid $700 Million for a Surgical Robot It Can't Sell in the United States | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-09-05-medtronic-700m-robot-it-cant-sell-in-the-us [← All Field Notes](/field-notes) September 5, 2026 Surgical Robotics Medtronic committed about $700 million this month to a surgical robot it didn't build and can't sell in the United States. The money buys distribution rights to Sentire, a multi-port system from Cornerstone Robotics in Hong Kong, in select markets outside the US where the machine already carries approval. The obvious question was whether that signals thin confidence in Hugo, Medtronic's own robot. Chief executive Geoff Martha told MedTech Dive it was the opposite. He's right, though the reason has little to do with Hugo. Matt Anderson, who runs Medtronic's surgical business, framed the two-platform decision around the "varying needs and preferences" of surgeons, health systems and markets. Cornerstone built the whole Sentire system in-house, including an immersive console and a dual-console setup. It won its CE mark in May for general, gynecologic, thoracic and urologic procedures, and it holds approvals in China and Singapore. ## Medtronic wasn't short a robot Hugo is in more than 35 countries across six continents, with roughly 250 systems installed. Medtronic expects it to pass 50,000 procedures globally by the end of its fiscal year, and procedure volume is growing at more than twice the rate of the robotic surgery market. The FDA cleared Hugo for urologic surgery inside the past year, which made Medtronic the first large device company to put a soft-tissue robot against Intuitive's da Vinci in the US. Filings for general surgery and gynecology are in, and Johnson & Johnson got its Ottava authorization in July. Nobody in that group is short a platform, which is why the $700 million didn't buy a machine. It bought rooms. ## Single-digit penetration is the whole business Global penetration of robotic-assisted surgery still sits in the single digits. In emerging markets it runs around one percent. Those two numbers are the argument I've been making from a stage for years, and they have nothing to do with whether the robot is better. Manual laparoscopic nerve-sparing prostatectomy is one of the hardest operations in surgery, and the number of surgeons who can do it consistently by hand is in the dozens. The da Vinci grew that population by orders of magnitude. More surgeons, more patients. The comparison that matters is robot against open surgery, because that's what most of those patients get otherwise. Ninety-nine percent of a market is a price, workflow and training gap long before it's a technology gap. A flagship platform tuned for a large American academic center reaches a thin slice of that. Two platforms, approved in different jurisdictions and priced for different buyers, reach more of it. Medtronic can push both through a distribution organization it already pays for. ## What made Sentire buyable Cornerstone arrived at this deal with a CE mark, a China approval and a Singapore approval in hand. Those clearances turned an engineering program into something a strategic partner could start shipping on Monday. Without them, the same console is a demo, and demos are cheap in this category. Approvals are the part hard-tech founders keep underweighting. If you're building a platform against an entrenched incumbent, the useful question is which segment the incumbent's machine structurally doesn't fit, and whether you can clear a regulator there before anyone else does. Form factor and price follow the task, which is why a company with a working robot paid $700 million for access to a second one. Dave's take Cornerstone made the decision that mattered years before anyone sat at a deal table, when it picked which regulators to clear and in what order. I watch founders treat the regulatory pathway as a cost line and the product architecture as the strategy. They're the same decision. From Dave’s video library Dave on what happens when every sale still runs through the founder, and the point where that stops being a growth plan. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [ARPA-H Bet $175 Million on Autonomous Stroke Robots. The Hard Part Is the Room They Land In. →](/field-notes/2026-08-22-arpa-h-stroke-robots-the-room-they-land-in) [J&J Cleared AI on Its Lung Robot. The Fourth Update in Eighteen Months Matters More. →](/field-notes/2026-08-18-jj-monarch-fourth-update-in-eighteen-months) [The FDA Authorized a Blood Draw Robot. The Ratio in the Label Is the Product. →](/field-notes/2026-08-21-blood-draw-robot-ratio-in-the-label) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Stryker's Surgical Headset Is FDA Authorized. Its Own Label Keeps the Monitors On. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-09-06-stryker-headset-label-keeps-the-monitors-on [← All Field Notes](/field-notes) September 6, 2026 Medical Devices Stryker's new surgical headset does not replace the monitors in the operating room. Its own label says it can't. The FDA authorized SportSuite Vision on July 17 through the De Novo pathway. The indication that came with it requires the images a surgeon sees through an Apple Vision Pro to be used alongside the traditional monitors the rest of the room is watching. Stryker printed that requirement itself, at the foot of its own press release, and it's the reason the authorization exists. That release went out on September 1. Stryker said Chad Mather III, an orthopedic surgeon at Duke Health and a disclosed Stryker consultant, had completed the first hip arthroscopy using the system. He said it let him place the clinical information where he wanted it and reach it inside the sterile field: arthroscopic video, CT imaging, and Stryker's HipCheck and HipMap software in one view instead of on monitors scattered around the room. The indication covers two procedures. SportSuite Vision is authorized for the intraoperative display of arthroscopic video and medical imaging during femoroacetabular impingement and labral repair hip arthroscopy. That's the whole envelope. Not shoulders, not knees, and nothing that steers an instrument. A company with Stryker's balance sheet could have asked for much more. ## What a narrow indication buys I've argued for a while that a De Novo signal isn't the punishment founders take it for. If the agency thought you were high risk, you'd already be headed for a PMA. Routing you to De Novo means they think you sit inside a risk envelope they're willing to write a new regulation around. The catch is that most De Novo submissions don't land. Former FDA examiners have told me that roughly half of them arrive effectively dead, and that figure is what I was told rather than anything the agency publishes. The ones that go through are the ones that asked for a clearance they could get. So you trade scope. You give up claims you wanted on the first pass and you keep the ones your evidence already supports. Stryker was asking for a display class with no predicate anywhere. It was also putting a consumer headset inside a sterile field. Asking for every arthroscopic procedure in the catalog on top of that would have opened a much bigger evidence question about what happens when a surgeon's only view of the field runs through a video feed. Writing the monitors into the indication takes that question off the table. ## Being the predicate The payoff sits on the other side. A De Novo does more than authorize a device. It creates a regulation in the Code of Federal Regulations describing a class that didn't exist before, and everyone who follows has to be substantially equivalent to what that regulation says. I've been on the inside of that. In the last stretch of our De Novo at Galen, the FDA asked what we would like the new regulation to say. We conferred, wrote a paragraph, and sent it as a PDF. A few hours later they asked for it in Word. A day or two after that the clearance arrived, and the regulation describing our device was word for word what we had sent. Anyone entering that category has to match language my team wrote, and when we wanted to add a feature our own predicate was that same regulation. Stryker now holds the reference point for a surgeon-worn augmented reality display in the operating room. The next company that wants a headset in an OR has a path, and the path runs through a regulation built around Stryker's device and bounded by conditions Stryker chose to accept. When Stryker wants shoulders and knees, it files a 510(k) against its own device. ## The framing is the regulatory strategy All of that follows from one sentence: what the device is, stated so a regulator reads it as moderate risk. That sentence sets the scope you are allowed to ask for, which makes it a product decision, and founders keep handing it to the regulatory consultant. We learned that at Galen the expensive way. We framed the robot as a remote-control telemanipulator and put da Vinci up as the predicate. The FDA rejected the comparison, saying those predicates weren't appropriate for the device as we had described it. Framed instead as a stabilizer that improves a surgeon's access to anatomy at the end of a long instrument, the same machine went through as a De Novo, with no change to the hardware. The framing decides what evidence you owe. A claim about patient outcomes buys you a clinical trial. A claim about where information sits in a surgeon's visual field buys you engineering and usability testing. Those are two different companies with two different burn rates. Dave's take When I read a first clearance now, the first thing I look for is what the company gave up. Stryker's list is long and specific, which tells me someone with authority over the product was in the room the day that indication got written. At most companies I walk into, nobody was. From Dave’s video library Dave reads Duolingo's own filings against the headlines about its AI-first announcement, and shows what the documents commit the company to. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [The FDA Authorized a Blood Draw Robot. The Ratio in the Label Is the Product. →](/field-notes/2026-08-21-blood-draw-robot-ratio-in-the-label) [Every COVID-19 Test EUA Ends December 26. Authorized Was Never the Same as Cleared. →](/field-notes/2026-08-24-covid-test-eua-ends-authorized-is-not-cleared) [Medtronic Paid $700 Million for a Surgical Robot It Can't Sell in the United States →](/field-notes/2026-09-05-medtronic-700m-robot-it-cant-sell-in-the-us) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Cellares' Machine Met Every Release Spec. Bristol Myers Squibb Still Walked Away. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-09-07-cellares-met-every-release-spec-bms-still-walked [← All Field Notes](/field-notes) September 7, 2026 Commercialization Bristol Myers Squibb finished its evaluation and decided the machine could not build its product. That ended a $380 million capacity reservation and supply agreement signed with Cellares in 2024, under which the automated Cell Shuttle system was to make Breyanzi, Bristol Myers Squibb's approved CAR-T cell therapy. Cellares strongly disagrees with the characterization. Its Cell Shuttle, the company says, has already made GMP drug product in an FDA-regulated clinical program. The doses met every release specification, shipped on time, and went into patients. Both statements can be true at the same time. They describe two different tests. ## Clinical release and commercial qualification ask different questions A clinical batch has to conform. The process runs, the doses get tested against release specifications, and the ones that pass go to patients. That is a real bar, and on the public record the Cell Shuttle cleared it. Commercial qualification asks a longer question. Can the process hold that result at volume, at yield, and at a deviation rate the quality system can absorb? Can it hold through operator turnover, equipment changeover, and a second site, for as long as the therapy stays on the market? The word doing all the work in Bristol Myers Squibb's statement is commercial, and in a supply agreement the buyer defines that word. A machine that makes conforming doses one batch at a time in a clinical program has answered the first question and said nothing about the second. That distance between a conforming batch and a qualified process is the prototype-to-product gap, and it does not care that both parties are already running a GMP operation. Passing every release spec starts the conversation about commercial supply. It does not finish it, and the two bars get confused constantly because the paperwork looks similar from the outside. ## What a capacity reservation actually buys Confusing the two bars is expensive. About 100 positions at Cellares end on October 20, most of them in New Jersey, and the cuts hit software engineers, quality control and design staff, and manufacturing specialists. The company closed a $327 million Series D roughly two months before it announced those cuts, and says it has more than doubled the number of customers it serves since the start of the year. A $380 million capacity reservation is a customer buying the right to a supplier's factory, and the customer decides whether to use it. Money is runway, not a purchase order, and a $327 million round does not qualify a machine. A bigger anchor customer puts more of a supplier's build plan and headcount behind one company's qualification decision, and that exposure peaks on the day the agreement is signed. ## The requirements document you did not write This exposure is not particular to cell therapy. Any company selling a machine into somebody else's production line carries it. An industrial robot cell can pass a factory acceptance test at the integrator's shop and then fail a run-at-rate trial on the customer's floor, because the acceptance test measures what the integrator agreed to build, while the floor measures what the line actually demands. On any platform play, the question I keep returning to is whose document actually specifies the product. If a customer's quality organization writes the criteria that decide whether your machine is qualified, those criteria are your requirements, whether or not you have read them. Getting them on paper before the design is locked is a different project from getting them afterward. Only the first one can change the machine. Dave's take I don't know which company is right about the Cell Shuttle, and I am not sure the answer would help anyone. What I would want, sitting on either side of that agreement, is the qualification criteria written down and signed while the machine is still a drawing. Suppliers who wait for the customer's evaluation to tell them what commercial means find out at the worst possible moment, with the factory built and the headcount already hired. From Dave’s video library Dave walks through the five questions that surface why a customer actually left, and why the reason they hand you is rarely the whole one. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Kardium's Cleanroom Builds 500 a Year. The Business Needs 30,000. →](/field-notes/2026-09-03-kardium-cleanroom-500-a-year-needs-30000) [XPeng Raised $900 Million for a Humanoid. Its First Customer Is XPeng. →](/field-notes/2026-08-25-xpeng-humanoid-first-customer-is-xpeng) [Boston Scientific's Plants Went Quiet. Its Filing Names Shipping, Not Devices. →](/field-notes/2026-08-26-boston-scientific-filing-names-shipping-not-devices) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # The Army Bought Its First Production Laser. The Factory Was Funded in March. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-09-08-army-first-production-laser-factory-funded-in-march [← All Field Notes](/field-notes) September 8, 2026 Defense Hardware On March 3, AeroVironment committed $30 million to expanding three manufacturing buildings in Albuquerque. The announcement named no contract and no customer. It named 450 jobs, a $5 million check from the state of New Mexico, a $1 million check from the city, and a plan to build more directed energy systems. On September 2, six months later, the Army handed the company $464.8 million to put a laser weapon into production. The award is the Enduring High Energy Laser program, E-HEL, and AeroVironment calls it the first production contract for a high energy laser weapon in United States history. Its CEO, Wahid Nawabi, called it a defining moment for the future of modern defense. He's talking about September 2. What the March announcement shows is that the company had already decided, six months earlier and without a customer, to carry the cost of being able to build these at volume. ## The lasers were already working E-HEL builds on the Army Multi-Purpose High Energy Laser, prototypes already in service and tested at White Sands Missile Range. Soldiers have been shooting small drones out of the sky with them. What changed on September 2 was the size and shape of the order: dozens of systems delivered over years, on a 30-kilowatt unit that bolts to a Joint Light Tactical Vehicle or ships on a pallet. A prototype has to survive a range day. A production unit has to come off a line at a rate somebody else set and keep coming for years, which is a harder and far more expensive engineering problem. AeroVironment's John Garrity framed it in one line in the release: "E-HEL is not a future capability, it is a production-ready system, built on proven technology." ## The capacity decision came first Look at what went into Albuquerque in March. The money went to three existing sites in the Sandia Science and Technology Park, to capital equipment, and to 450 hires, alongside an incentive package the state structured as $5 million up front plus $6 million tied to performance. The release framed all of it against national security priorities in general terms and named no program, because in March the E-HEL contract didn't exist yet. Six million of the state's support only lands if the hiring follows, so AeroVironment took the schedule risk and New Mexico took almost none of it. Committing $30 million ahead of a program is a bet, and companies do build lines for contracts that get canceled. What makes this one reasonable is that once a prototype phase ends the buyer stops asking whether the device works and starts asking whether you'll still be delivering in year three. Founders I talk to treat manufacturing capacity as the reward for winning the order, so they wait. Then the order shows up, the buyer asks how many units land in the first year, and the honest answer starts with a lead time on tooling. That answer loses to the company that already spent the money, and it loses on a dimension the founder was never competing on. Capacity is the one thing you can't produce on demand, because it's made of buildings, machines that take a year to arrive, and people who have to be hired and trained before they're worth anything to you. Most of the distance from prototype to product is that, and little of it is engineering. ## The second road already has a buyer Committing capacity is easier to justify when more than one customer can absorb it, and counter-drone is one of the few defense categories where the civilian customer is already funded. On August 5 the Department of Homeland Security awarded a five-year, $1.5 billion counter-UAS contract across twelve companies, running through July 2031. Performance is written into the requirement: detect at two kilometers or better, mitigate at five hundred meters or better. Airports, stadiums and border crossings all buy on a different clock than the Army. That published floor gives a founder something a defense program rarely does. It states in writing what the civilian side will pay for, so you can design to it and know before committing tooling whether the second road is reachable from the first. A requirement you can read is what makes a second road testable, and most dual-use pitches I hear describe one with no requirement attached. That test is also my own gate: I won't coach a founder building an autonomous weapon aimed at people, and I will coach one whose platform has a credible civilian road. So I'd want to know which road a founder in this category designed the tooling around, because a line qualified for a military vehicle mount will not serve a stadium operator without being requalified, and requalifying gets more expensive every month it waits. Dave's take The part I'd want a founder to sit with is that this decision has to get made while it still looks unjustifiable, against a customer who hasn't committed to anything. AeroVironment made it in March with no program to point at. If the plan is to sort out manufacturing once somebody signs, you've already told your buyer where you rank. From Dave’s video library Dave covers what the fractional CPO role looks like week to week, and the six-month floor before it produces anything real. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [The Navy Wants Sixteen Times the Tomahawks. The Seven Years Is What Buys Them. →](/field-notes/2026-08-19-navy-tomahawk-seven-years-buys-the-rate) [Kardium's Cleanroom Builds 500 a Year. The Business Needs 30,000. →](/field-notes/2026-09-03-kardium-cleanroom-500-a-year-needs-30000) [GM Walked Away From This Battery Plant Twenty Months Before It Opened →](/field-notes/2026-08-20-gm-walked-away-from-this-battery-plant) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Nine Sites, $1.8 Million: How to Read a Hardware Company's Traction | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-09-10-nine-sites-1-8-million-reading-hardware-traction [← All Field Notes](/field-notes) September 10, 2026 Industrial Robotics Deployed is not a financial term. It shows up in decks because it sounds like revenue without having to behave like revenue, and nobody makes a private company prove the difference. Agility Robotics won't be private much longer. Its Digit humanoids have logged more than 65,000 operating hours across nine customer sites. Net sales last year were $1.8 million. ## Hours are engineering. Dollars are commerce. Both numbers come out of Agility's S-4. The company is merging with Churchill Capital Corp. XI at a $2.5 billion valuation, a deal expected to raise more than $620 million in gross proceeds. The Robot Report went through the filing on September 7. It also shows a $140 million operating loss and $111 million of operating spend, up from $71 million the year before. Divide the sales by the hours and you get under thirty dollars an hour, and that flatters it, because the hours are cumulative and the revenue covers one year. Most of those hours were never sold. They were spent proving a bipedal robot can survive a working warehouse, which is a real thing to prove and a different thing from a customer. Track the two separately and never let one stand in for the other. An hour of runtime tells you the engineering holds. A dollar tells you somebody watched what the hour produced and decided it beat the alternative on cost, or on injury rate, or on the third shift they can't staff. A company that lets runtime speak for demand will size its fleet to a market nobody has tested. ## One order carries the whole case The filing names more than $300 million in multi-year orders for Digit v5 from a single customer it doesn't identify. Set that beside $1.8 million in actual sales and the shape of the business is clear: almost all of the commercial case for a $2.5 billion company sits inside one purchase order. The deployments named in the filing are real, and the customers are not small. GXO, Toyota Motor Manufacturing Canada, Schaeffler, Mercado Libre. At GXO, Digit cleared more than 100,000 totes under a robots-as-a-service contract, which is about as close to a paying commercial customer as this category has produced. None of that is the $300 million. When one buyer carries the forecast, that buyer holds your pricing, your roadmap and your delivery dates, and you have nothing to negotiate back with. Losing them doesn't take a percentage off the plan, it takes the plan. That's expensive money you never raised and can't refinance, and it doesn't feel like a risk while you're taking it, because a large order looks like the best news the company has had. ## The forecast is a factory plan Agility projects 800 deployments in 2027, 7,000 in 2030 and 25,000 in 2035. That curve is a bill of materials, a supplier qualification schedule, a field service organization, a spare parts depot and a training program, each of which has to exist before the units it supports can ship. Eight hundred in 2027 is the number worth arguing about, because it arrives soonest and it's the only one anybody can hold the company to on this side of the merger. Everything past it is a slope drawn between a near-term commitment and a market size. The far end of the curve costs nothing to write down. The near end has to be built with money the company doesn't yet have from customers it has mostly not yet billed. The question I keep putting to my own numbers is which of them would survive being written into a document with a penalty attached. Deployed, piloted, engaged, in evaluation: those all survive a board meeting. Shipped, invoiced and collected survive a filing. Companies that never run the comparison tend to find out about a year late that what they were calling traction was a line item in somebody else's research budget. That comparison is why I ask a founder for an invoice before I spend time on a logo slide. Dave's take I'm not calling Agility's business thin. Sixty-five thousand hours in live warehouses is more commercial evidence than most of this category has, and the company put its numbers where anyone can check them. What I'd want to know before believing the curve is which of those nine sites would sign again at a price that covers the robot, because humanoids give a good demo, and a demo is the easiest thing in this business to mistake for demand. From Dave’s video library Dave covers how founders end up with months of engineering behind them and no evidence anyone will pay for it, which is the gap these numbers put a price on. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [XPeng Raised $900 Million for a Humanoid. Its First Customer Is XPeng. →](/field-notes/2026-08-25-xpeng-humanoid-first-customer-is-xpeng) [Cellares' Machine Met Every Release Spec. Bristol Myers Squibb Still Walked Away. →](/field-notes/2026-09-07-cellares-met-every-release-spec-bms-still-walked) [Kardium's Cleanroom Builds 500 a Year. The Business Needs 30,000. →](/field-notes/2026-09-03-kardium-cleanroom-500-a-year-needs-30000) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Axogen Paid $200 Million to Widen the Pool of Surgeons Who Can Repair a Nerve | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-09-11-axogen-200m-nerve-repair-without-the-microsurgeon [← All Field Notes](/field-notes) September 11, 2026 Medical Devices Axogen agreed on Thursday to pay $200 million in cash for BioCircuit Technologies, a private company in Atlanta. BioCircuit's one commercial product joins a severed nerve without stitches. Axogen already sells grafts and connectors into that operation. What it just bought is a repair that does not require a surgeon trained in microsuture. The device is NerveTape. The two companies call it the first FDA-cleared sutureless device for repairing transected nerves, and the announcement states the commercial point plainly: cutting the dependency on microsurgical skills puts nerve repair in reach of more surgeons and more care settings. The market read Thursday as a financing event. Axogen's stock fell more than 10% to $42.46 while the company priced a concurrent public offering of about $208.7 million, close to the entire purchase price. Truist Securities said the rationale makes sense and flagged the dilution from the equity raise as a point of contention for investors. ## What $200 million is buying Axogen's catalog runs on microsurgeons. Avance, the Axoguard connectors and protectors, the nerve cap: every one of them goes into an operation performed by someone trained to work under magnification. The number of nerves severed in a year is not what caps that business. The number of surgeons who can repair one is. NerveTape changes the second number. I have been making a version of this argument about surgical robots for years, usually into a room that wants to debate whether robot-assisted surgeons get better outcomes than the small group who can do the operation by hand. That comparison flatters a few dozen experts and ignores every patient who would otherwise have gotten open surgery. Manual laparoscopic nerve-sparing prostatectomy is one of the hardest operations in medicine, and the number of surgeons who can do it consistently by hand runs to the dozens, with no exact count published anywhere. The da Vinci put that operation in reach of thousands. More surgeons means more patients get offered the option. Until Thursday I had not seen anyone put a dollar figure on the same argument in nerve repair. Michael Dale, Axogen's chief executive, said the deal moves nerve repair toward being an expected standard of care. That is a market-size claim in the language of a mission statement. A procedure becomes standard when enough people can perform it, and Axogen already owns the harder half of the distribution problem: the surgeon relationships and hospital contracts it built selling grafts. The device arrives in a channel that is already there. ## The clause that says what Axogen would not buy Axogen was specific about what it is not taking. The deal does not close until BioCircuit spins out its electronics research and development business. BioCircuit describes its own expertise as biomaterials, precision nitinol design and manufacturing, and bioelectronics. Axogen is taking the first two. The spin-out is a product boundary, drawn by the buyer, in a contract. The cleared mechanical device has a market Axogen can name and a sales organization it already pays for, which is what makes it worth $200 million in cash at closing. The bioelectronics program has a research plan and a new owner. ## Who has to be in the room Nerves are incidental if you build work cells or diagnostics or anything else that needs a specialist present. Every hard-tech product has someone who has to be in the room for it to work, and how many of those people exist is the size of your market. A cell that only one integrator can commission sells at the speed of that integrator's calendar. An assay that needs a trained technician goes to the sites that already have one. Neither of those constraints appears anywhere on a bill of materials, so neither one gets engineered. Dave's take Axogen bought the right to sell nerve repair to surgeons who were never going to learn microsuture, and funded it almost entirely with new stock. When only a specialist can operate your product, that specialist is your ceiling, and more engineering on the device will not lift it. The cheapest question I know to put to a roadmap is which requirement on the user we could delete. From Dave’s video library Dave covers how to find out which part of the product decides whether anyone can buy it, before you spend months on the wrong part. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [The FDA Authorized a Blood Draw Robot. The Ratio in the Label Is the Product. →](/field-notes/2026-08-21-blood-draw-robot-ratio-in-the-label) [Medtronic Paid $700 Million for a Surgical Robot It Can't Sell in the United States →](/field-notes/2026-09-05-medtronic-700m-robot-it-cant-sell-in-the-us) [Kardium's Cleanroom Builds 500 a Year. The Business Needs 30,000. →](/field-notes/2026-09-03-kardium-cleanroom-500-a-year-needs-30000) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Teradyne Is Suing Over Three Universal Robots Patents. All Three Expire in March. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-09-12-teradyne-jaka-patents-expire-in-march [← All Field Notes](/field-notes) September 12, 2026 Commercialization Universal Robots bought one of JAKA's cobot arms and took a hard look at how it was built. On August 24 its parent, Teradyne Robotics, filed patent infringement proceedings against JAKA's German subsidiary at the Unified Patent Court in Copenhagen, naming three Universal Robots patents. All three were filed on March 2, 2007. All three expire on March 2, 2027. Twenty years of exclusivity run out in under six months. Whatever the court decides about JAKA, that is the date Teradyne has to be ready for. ## What the docket says The case is UPC-CFI-0003057/2026, filed in the Copenhagen Local Division and announced publicly three days later. One patent covers a programmable robot and its touchscreen interface. The other two cover joint construction, including a safety brake built from a solenoid, a ratchet and an annular friction member. All three carry a March 2006 priority date and are assigned to Universal Robots. The docket holds 128 documents and none of them are public yet. JAKA says it was never served. In a September 3 statement the company said its legal advisers had confirmed that no proceedings were open against it as of September 2. It pointed to two earlier freedom-to-operate analyses that it says found no infringement of Universal Robots patents, and it moved for injunctive relief to stop what it calls false and damaging public statements. JAKA also says it holds more than 300 patents of its own. The Robot Report notes the gap between filing and serving could explain the discrepancy. Nothing has been decided. ## Twenty years, ending in March The calendar is the one part of this case already settled. Universal Robots filed all three applications on March 2, 2007, claiming priority to March 2006. Google Patents lists the anticipated expiration for each one as March 2, 2027, and all three are still active. EP3530420, one of the two joint patents, was not granted until September 2023, which left it about three and a half years of enforceable life from grant to expiry. David Brandt, the chief technology officer at Universal Robots, put the underlying problem plainly when the case was announced. "You have to build the robot and ship it to a customer to have a robot," he told The Robot Report. "And, of course, some competitor can buy the robot, take it apart, and look at how we design stuff." A patent stops a competitor from selling in Europe, and leaves untouched everything that competitor learns from a machine already in the field. iRobot let its patent term run out without building much behind it. Exclusivity kept rival brands off the shelf for years, and all the while the factory assembling the product was learning everything about it. In December 2025 iRobot filed for Chapter 11, and the company that ended up with the equity was Shenzhen Picea Robotics, its own primary contract manufacturer. ## What JAKA still has to get past Universal Robots says it has sold more than 100,000 cobots, and it has spent more than a decade building software, accessories and an integrator base around the arms. Teradyne Robotics reported $100 million in the second quarter of 2026, its fifth consecutive quarter of growth and 33% above the same quarter a year earlier. Most of that came from Universal Robots. None of that expires in March. A cheaper arm with comparable joints still has to displace a machine the customer's integrator already knows how to program, and that kind of switching cost compounds while a patent term counts down. No examiner puts a term limit on that work, and it is what Teradyne will still have on March 3. That is why I read the lawsuit as housekeeping. Universal Robots spent its twenty years building an integrator base. The University of Nebraska Medical Center tech transfer program I worked on years ago spent its patent life waiting. We built a navigated sagittal saw for knee replacement, and I managed the patents that came out of it. The novel claim was about cutting planes, a real distinction from the burr-based platforms most of the field was building, and we won that argument at the patent office. Then the company stopped pushing the product. Stryker's Mako RPS now ships a navigated handheld saw that does the job better than ours ever did. Teradyne has something the Nebraska program never had, and Brandt names it without quite calling it an advantage. He worries aloud that unsafe copies could hurt somebody and sour regulators on the entire cobot category, which is a category Universal Robots helped create. A safety record takes years to build, cannot be reverse engineered off a bench, and carries no term limit either. Dave's take A patent is a clock, not a moat. March 2, 2027 has been readable on these three filings since the day each one was granted. What I want to know from any founder holding a portfolio like that is which advantage is still standing on March 3, and who spent the twenty years building it. From Dave’s video library A related argument in video form. I read Anthropic's own founder playbook and show where its fine print undercuts the headline claim that AI removed the hard part. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Nine Sites, $1.8 Million: How to Read a Hardware Company's Traction →](/field-notes/2026-09-10-nine-sites-1-8-million-reading-hardware-traction) [Cellares' Machine Met Every Release Spec. Bristol Myers Squibb Still Walked Away. →](/field-notes/2026-09-07-cellares-met-every-release-spec-bms-still-walked) [XPeng Raised $900 Million for a Humanoid. Its First Customer Is XPeng. →](/field-notes/2026-08-25-xpeng-humanoid-first-customer-is-xpeng) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # GM Is Building Battery Cells It Will Never Sell So It Can Judge the Ones It Buys | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-09-13-gm-900m-battery-cells-it-will-never-sell [← All Field Notes](/field-notes) September 13, 2026 Climate Hardware GM is finishing a battery plant in Warren, Michigan this year, and it won't sell a single cell that comes off the line. The Battery Cell Development Center is part of $900 million the company is putting into research and prototype facilities there. The campus runs to 700 acres and GM opened it in the 1950s. When construction wraps, the line turns out thousands of cells a year, all of them for development work. Production cells come from its suppliers. The Wallace Battery Cell Innovation Center opened on the same campus in 2022 and makes about a dozen full-sized prototype cells a week. The new building makes thousands. Neither number is production volume. The stretch between a working chemistry and gigascale manufacturing has a name in this industry, the valley of death, and GM just spent $900 million furnishing it. The company aims to commercialize its lithium-manganese-rich chemistry in 2028, and started prototyping sodium-ion cells in June. ## The two questions a cell line has to answer A dozen cells a week answers chemistry questions. Does this cathode hold its capacity through a winter, and what's left of it by cycle four hundred. Radu Theyyunni directs global propulsion virtual engineering at GM, and he told Latitude Media his team can tell within days how a change to the chemistry moves cost, range, performance and safety. Thousands a week answers whether you can make the same cell twice. Coating thickness drifts. One lot of separator film behaves differently from the lot before it. A dozen hand-built cells hide all of that, because somebody is watching every step. At thousands, nobody is, and the process has to hold the tolerance on its own. ## Buying the standing to check someone else's work GM buys its cells from LG, Panasonic, CATL and BYD, and the new building doesn't change that. Mo Gallegos runs the development center, and he explained to Latitude Media why GM built it anyway: understanding how a cell gets made from the ground up makes you “a better buyer and a better partner.” What a pilot line produces isn't cells. It's a written process definition, and the standing to read what a supplier tells you. Your contract manufacturer comes back with a yield number and a four-month tooling fix, and somebody on your side has to know whether both of those are true. ## What a ten-person company keeps in house Outsourcing most of the manufacturing is the right call for a small hardware company. The decision that matters is which piece of the process you run yourself, badly and at small volume, so you can buy the rest of it well. GM's answer is the whole cell, at a scale where the process starts misbehaving. For a ten-person device company the answer is smaller, and it isn't zero. One subassembly, built in house long enough to learn what breaks it, buys you the vocabulary for every supplier conversation that follows. The books work the same way. You'll never do your own accounting, and you still have to read the statement and catch the number that moved for a reason nobody mentioned. Learning that much accounting costs an evening. Learning the manufacturing version cost GM $900 million, and for a ten-person company it costs one stubborn subassembly. Dave's take A pilot line looks like waste right up until the first supplier call where the numbers don't add up. My rule on this is that a founder should be able to build, badly, at least one of the things they buy. Keep a piece of the making and you keep the ability to tell when you're being managed. From Dave’s video library Dave walks through five decisions that lock in the moment engineering starts, including the manufacturing model, and why the commercial conversation has to come first. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Kardium's Cleanroom Builds 500 a Year. The Business Needs 30,000. →](/field-notes/2026-09-03-kardium-cleanroom-500-a-year-needs-30000) [Cellares' Machine Met Every Release Spec. Bristol Myers Squibb Still Walked Away. →](/field-notes/2026-09-07-cellares-met-every-release-spec-bms-still-walked) [Two in Ten Storage Projects Now Treat the 30% Tax Credit as Optional →](/field-notes/2026-09-02-storage-projects-treat-tax-credit-as-optional) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # J&J Bought Into a $37 Billion Orthopedics Market. It's Leaving a $50 Billion One. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-09-14-jnj-bought-37-billion-market-leaving-50-billion [← All Field Notes](/field-notes) September 14, 2026 Medical Devices In April 2011, Johnson & Johnson's chief executive Bill Weldon explained why he was paying $21.3 billion for a Swiss implant maker called Synthes. Orthopedics, he said, was "a large and growing $37 billion global market." He called it a growth driver for the company. Last Friday, Bloomberg reported that Apollo Global Management is in talks to take that business off J&J's hands for something close to $20 billion. The market did not shrink. When J&J announced last October that it intended to separate orthopedics, the company described the category as a $50 billion-plus global market serving about seven million patients a year. By J&J's own accounting in both documents, the market grew by more than a third. The business J&J built inside it is now being discussed at less than the price of Synthes alone. ## What fifteen years did to the price J&J paid $21.3 billion for Synthes in 2011, or $19.3 billion after Synthes' cash. It merged the company with the DePuy orthopedics business it already owned. Apollo is discussing something close to $20 billion for the result. Revenue was $9.2 billion in fiscal 2024 and $9.3 billion in 2025. Weldon's case in 2011 rested on more than the market number. He pointed to aging populations, rising obesity, and demand from emerging markets, and the demand has behaved the way he said it would. Fifteen years of it produced a business J&J now calls low growth and a buyer prices below what Synthes alone cost. First-quarter orthopedics sales this year came in at $2.38 billion, up 6.3 percent from a year earlier. DePuy Synthes is the largest orthopedics company in the world and it will stay that way whoever ends up owning it. J&J is also still weighing a spinoff, which its finance chief has called the primary path. The Apollo talks may end without a deal. What neither path changes is the number the business now carries. ## The market never changed its mind That number is a verdict on a growth rate, and J&J has delivered that verdict twice in fifteen years, in opposite directions. In 2011 the size and growth of orthopedics were the reason to buy in. In October 2025 Tim Schmid, who runs J&J's medtech business, called orthopedics a lower-growth market and described the separation as shrinking in order to grow faster. Joaquin Duato was specific about where the money goes instead: cardiovascular and robotics. The market got bigger the whole time. J&J published both halves of its own answer in the same quarter. Cardiovascular grew 12.6 percent in the third quarter of 2025. Orthopedics grew 3.8 percent in that quarter and slipped slightly across the first nine months. Both of those growth rates live inside the same $50 billion market figure. That gap is where the $20 billion comes from. Growth follows the purchases that are changing, and in orthopedics the purchase has not changed much. A hip implant is bought today the way it was bought a decade ago, by a value analysis committee weighing it against the implant already on the shelf. Cardiovascular and robotics are bought by people whose procedures are still in motion, which is why J&J is keeping them. If you are building implants, disposables, or any other high-volume product in a settled category, the growth in your market slide is going somewhere, and you should be able to say where. Dave's take Revenue, costs, runway. Everything else is decoration, and market size is the most expensive decoration on the wall, because it looks like an argument. J&J bought that argument in 2011 for $21.3 billion. None of this makes DePuy Synthes a bad business. It makes it a business that does not belong in a portfolio being rebuilt around growth rate, and J&J said so in writing last October. From Dave’s video library Dave covers what the word unicorn actually means, and why growth rate became the only measure anyone applies to a business. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Axogen Paid $200 Million to Widen the Pool of Surgeons Who Can Repair a Nerve →](/field-notes/2026-09-11-axogen-200m-nerve-repair-without-the-microsurgeon) [Medtronic Paid $700 Million for a Surgical Robot It Can't Sell in the United States →](/field-notes/2026-09-05-medtronic-700m-robot-it-cant-sell-in-the-us) [Kardium's Cleanroom Builds 500 a Year. The Business Needs 30,000. →](/field-notes/2026-09-03-kardium-cleanroom-500-a-year-needs-30000) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # When No Law Fits Your Product, Regulators Reach for the Nearest One | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-09-15-no-law-fits-regulators-reach-for-the-nearest-one [← All Field Notes](/field-notes) September 15, 2026 Regulatory In 1986 the White House Office of Science and Technology Policy decided which law governs a living thing that somebody built. It picked the Toxic Substances Control Act, passed in 1976 to regulate industrial chemicals. The reasoning was that genomes are made of chemicals, and novel chemicals were already covered. Between 1987 and 2018, researchers filed more than 240 applications under that framework. Almost none were approved. Niko McCarty published that history this month in Works in Progress. Drones went the same way, landing under aviation authority because aviation authority was what existed, and self-driving cars are there now, measured against vehicle safety standards drafted around the assumption of a driver. In each case the technology arrived before anyone had written a law for it, and a regulator reached for the nearest one already on the books. The 240 applications are what happens next. ## Nobody wrote a new law The 1986 document is called the Coordinated Framework for the Regulation of Biotechnology, and its central finding was that no new statutory authority was needed. Writing in Frontiers in Bioengineering and Biotechnology last November, Leah Buchman and Emma Kovak count four statutes holding the whole field up: the pesticide law, the food and drug law, the chemicals law, and the Plant Protection Act. The framework was updated in 1992 and again in 2017 without anyone revising the logic underneath it. The EPA issued its biotechnology rule in 1997, and a nitrogen-fixing biofertilizer for alfalfa cleared that same year. Congress never voted on the assignment. It was an interpretation, written by a handful of civil servants, and it has held for forty years. ## What the match cost An engineered enzyme called FAST-PETase breaks down PET plastic in under twenty-four hours, where the natural version needs about six weeks. A modified strain of E. coli detects cadmium at five parts per billion. A Cornell variant of Gluconobacter pulls rare earth metals out of ore 73 percent faster than the unmodified organism. All three work, and none of them is something you can buy, which is where McCarty's title comes from: all the microbes you can't have. Other countries reached different answers from the same science. Israel approved explosive-sensing microbes in roughly a year. Brazil approved a release of engineered Aedes aegypti that cut local mosquito populations by 80 to 96 percent. ## Where the decision is still open Medical devices are the one category here with a way out, and the microbe applicants had no equivalent. When no existing classification matches, a company can ask the FDA to write a new one instead. I led the De Novo effort at Galen Robotics, where the framing decision came first and did most of the work. The hardware was identical whether we described it as a remote manipulator or as a device that steadies a surgeon's access to anatomy. Only one of those descriptions had a route through. A De Novo that lands means the regulation describing your category is the one your team helped draft, and everyone who follows has to match you. It is pretty easy to be substantially equivalent to yourself. Writing your own classification is only possible while the question is open. Once a product has been read under a statute a few times, that reading becomes the precedent the next reviewer starts from, and proposing a different one means arguing against the file. So the question I ask early is not what the pathway is. It is who decided that was the pathway, when they decided it, and whether anyone has tested the decision since. For a device that decision is often still live, and a pre-submission meeting is where you find out. For engineered microbes it was settled in 1986, and most categories look more like the microbes than like the devices. Dave's take Regulatory strategy starts on the day somebody writes one sentence describing what the product is. That sentence decides which statute a reviewer reaches for, and every document that follows has to live inside it. I would rather spend a month on that sentence than a year answering questions generated by the wrong one. From Dave’s video library Dave takes a government move against an AI system apart and separates what the rule actually does from what the coverage said it did. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [FDA Reclassified This Cancer Test Category. Fewer Than Five People Commented. →](/field-notes/2026-09-04-fda-reclassified-cancer-tests-five-comments) [The FDA Wants to Know How You'd Test a Device That Answers Differently Every Time →](/field-notes/2026-08-30-fda-genai-devices-answer-differently-every-time) [Every COVID-19 Test EUA Ends December 26. Authorized Was Never the Same as Cleared. →](/field-notes/2026-08-24-covid-test-eua-ends-authorized-is-not-cleared) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # The FDA Approved This Valve for Symptom-Free Patients. Medicare Started Paying Sixteen Months Later. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-09-16-fda-approved-it-medicare-paid-sixteen-months-later [← All Field Notes](/field-notes) September 16, 2026 Medical Devices The day after the FDA approved Edwards' Sapien 3 valves for severe aortic stenosis in patients who had no symptoms, a cardiologist could offer one of those patients a new valve. What that cardiologist couldn't do was bill Medicare for it. That approval came on May 1, 2025. Medicare finalized coverage for the same patients on September 11 of this year. The device didn't change in between, and neither did the label or the trial behind it. ## Two agencies, two questions The FDA wanted to know whether replacing the valve early was safe and effective in people who felt fine. EARLY TAVR answered it: 901 patients randomized between early replacement and the standard practice of watchful waiting. Over nearly four years the replacement arm cut the risk of a composite of death, stroke, and unplanned cardiovascular hospitalization by half, a hazard ratio of 0.50. By the two-year mark, more than 70 percent of the watchful-waiting patients had received a valve anyway, once their symptoms arrived. A different agency decides what Medicare pays for. The Centers for Medicare and Medicaid Services, CMS, asks whether a service is reasonable and necessary for Medicare beneficiaries, which isn't the FDA's question. Its reviewers read the same trial, on their own schedule, and acted on it this September. ## The coverage arrived with an obligation attached That September determination dropped the coverage-with-evidence-development requirement for symptomatic patients, a condition that had made continued data collection part of getting paid. It kept that same requirement for the asymptomatic patients it had just added, calling the evidence there promising but incomplete. So the new indication carries a registry with it, and what that costs depends on who is carrying it. TAVR is roughly 75 percent of Edwards' revenue, and Stifel analysts value the global market above $7 billion, so the data collection is a line item against a business the decision just widened. At Medtronic, TAVR is 5.5 percent of revenue and the decision barely moves the quarter. A startup with one product doesn't get either position. Its data collection comes out of its launch budget, in its launch year. ## The calendar was public the whole time CMS opened the proposal on June 15, held a thirty-day comment window that closed July 15, and finalized in September. The proposal also took out facility-level volume thresholds, replacing them with expectations about infrastructure and quality improvement. It cut the heart-team workup to a single in-person evaluation by a TAVR operator, with a second evaluation optional and allowed by telehealth or chart review. Those are workflow changes, and they decide where the procedure can happen. A hospital that could never clear the old volume threshold can now stand up a program, and a patient who would have traveled to a high-volume center can be treated closer to home. That growth comes from the volume rule coming out, not from the label. The label was never what stood between Edwards and those patients. A coverage decision stood there, and behind it a volume rule, and MedTech is just where that gap is easiest to see. The authority that lets you sell is rarely the authority that decides whether anyone can buy. In enterprise software it's the security review and the procurement desk, holding the line months after the contract is technically signable. In grid hardware it's the interconnection queue, plus the separate question of whether a project qualifies for the credit. In defense it's the program of record. None of those bodies is bound by the approval you already hold. Most founders I work with bring me a clearance plan with dates on it and a reimbursement plan that is one sentence in a deck. Dave's take In my experience an approval binds the agency that issued it and nobody else. Edwards paid for the trial that changed Medicare's mind and still waited sixteen months for the decision that turned approval into revenue. Put the second gate on the roadmap with real dates against it, because the runway you raised against the first one is the runway you'll be spending while you wait. From Dave’s video library Dave walks through how to tell real demand from encouragement, and which signals hold up when you test them. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Medtronic Paid $700 Million for a Surgical Robot It Can't Sell in the United States →](/field-notes/2026-09-05-medtronic-700m-robot-it-cant-sell-in-the-us) [Axogen Paid $200 Million to Widen the Pool of Surgeons Who Can Repair a Nerve →](/field-notes/2026-09-11-axogen-200m-nerve-repair-without-the-microsurgeon) [Kardium's Cleanroom Builds 500 a Year. The Business Needs 30,000. →](/field-notes/2026-09-03-kardium-cleanroom-500-a-year-needs-30000) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # A $400 Million Device Failed Its Trial. Its Engineers Bought It Back. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-09-17-400m-device-failed-its-trial-engineers-bought-it-back [← All Field Notes](/field-notes) September 17, 2026 Medical Devices Johnson & Johnson paid $400 million in November 2023 for a heart implant that had never been through a pivotal trial. The trial started in February 2024. By the middle of 2025 it was suspended, because the implant was unwrapping inside the heart and blood was leaking past it. Last week the people who designed it bought it back. The buyer is Jaguar LAA, a new company in Santa Rosa, California, formed by Santé Ventures with members of the Laminar management team. The deal was announced on September 11, 2026, and the terms were not disclosed. J&J called the sale a disciplined approach to portfolio management, and by its own standards that is accurate. What the phrase cannot do is separate a device that does not work from a device on its first design cycle. ## What the trial found The Laminar device closes the left atrial appendage by rotating it shut. That geometry leaves little hardware exposed in the left atrium afterward, which is what made the program worth $400 million next to Boston Scientific's Watchman and Abbott's Amulet. Then it relaxed after implant, and the seal leaked. A rotational closure that loosens is a fixable class of problem: it could be the materials, the retention features, the delivery technique, or some combination, and finding out which takes another design cycle and another trial. An investigational device exemption study exists to surface exactly this, and NCT06168942 surfaced it about a year and a half in. The trial did not show the device was unfixable. It showed the fix would cost more than J&J wanted to spend, and J&J had a portfolio review to run that number through. ## Who can fund the second design cycle Laminar arrived inside J&J MedTech by acquisition. Nothing else J&J was already shipping depended on it going forward, and it shared no platform and no sales channel with an approved product, so at review time it defended itself on its own numbers alone. Those numbers in the summer of 2025 were a suspended trial, an unproven fix, a fresh IDE, and two approved competitors already selling. Funding the next design cycle meant taking money from products J&J ships today and handing it to a device that would reach the market years from now, if the second design worked. Every quarter Laminar sat there, the certainty got surer and the maybe got a year older. Jaguar runs different math on one program. The engineers who watched the device come unwrapped are the ones being paid to fix it, and that work is not competing for money against anything else the company sells, because the company sells nothing else. Jaguar says it will build on the clinical work already done and carry the device through development and regulatory review. Getting that work funded took a new company and an outside raise against their own program. A founder selling to a strategic before the pivotal trial reads out is on the other side of that same transaction, handing the next design cycle to a capital process that has no particular reason to fund it. Before signing, find out who inside the acquirer owns the program when the first trial surprises everyone, and what else their budget has to cover. Dave's take Santé Ventures and a handful of the original team just paid for the right to finish something a $400 million buyer decided not to finish, and they got the better end of it. Trial data, a named failure mode, and the engineers who know why it happened are worth more in one room where fixing it is the only job than they were worth inside a portfolio with somewhere else to put the money. From Dave’s video library Dave works through what happens to each idea when a company splits its attention across several at once. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [J&J Bought Into a $37 Billion Orthopedics Market. It's Leaving a $50 Billion One. →](/field-notes/2026-09-14-jnj-bought-37-billion-market-leaving-50-billion) [Cellares' Machine Met Every Release Spec. Bristol Myers Squibb Still Walked Away. →](/field-notes/2026-09-07-cellares-met-every-release-spec-bms-still-walked) [Medtronic Paid $700 Million for a Surgical Robot It Can't Sell in the United States →](/field-notes/2026-09-05-medtronic-700m-robot-it-cant-sell-in-the-us) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # ISO 21423 Isn’t Published Yet. The Code That Speaks It Is Already Free. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-09-18-iso-21423-not-published-the-code-is-already-free [← All Field Notes](/field-notes) September 18, 2026 Industrial Robotics On Monday InOrbit.AI put its fleet management software, OpenRobOps, on the internet for free. Apache 2.0, no license fee, hosted by the Open Source Robotics Foundation. Inside it is the first working implementation of ISO 21423, the standard that will let an autonomous forklift from one vendor and a floor scrubber from another stop blocking each other in a warehouse aisle. The standard has not finished publication. The code that speaks it is already out. Standards do not usually arrive in that order. I started my internet career in 1989, and some of that early work was on the standards development for the OSI model. I gave keynote talks about the seven layers at trade shows, back when the industry still assumed OSI would replace TCP/IP. It did not. Traffic was already moving on TCP/IP while the OSI documents were still under negotiation, and the network went where the working code was. I sold that model from a stage for a few years, and I now think the seven layers were a teaching diagram that the running network never matched. OSI won the textbooks and lost the network. ## Who was in the room ISO 21423 sets the communication and data exchange rules between industrial mobile robots and the fleet managers that direct them. Close to 90 participants from about fifteen countries wrote it inside ISO Technical Committee 299, Working Group 10. The document builds on the MassRobotics interoperability standard and VDA 5050, and it runs on MQTT and JSON, which is the plumbing the people who have to implement it were already using. Buyers sat in that working group. General Motors backed the effort from the beginning, according to Steven DuMont, who manages advanced automation development there and runs the company's mixed-brand mobile robot fleets around the world. At Procter & Gamble, robotics and machine safety director Mark Lewandowski says his team took part because the industry needed the standard. Those are two of the largest floors in the world where robots from different vendors share space, and both companies helped write the document. ## What a free implementation does to a roadmap OpenRobOps ingests telemetry over networks that drop, tracks robots on a live map, keeps its configuration in Git, and runs from the edge or the cloud. Florian Pestoni founded InOrbit in 2017 and calls the problem it removes the build trap: robot makers spending years writing their own backends and databases when the robot is the product. Steve Cousins ran Willow Garage and now directs the Stanford Robotics Center. He compares the release to what ROS did for robot development. Ten companies ran a multi-vendor orchestration demo at Automate 2026, the first public showing of the standard. A mobile robot company can put a number on this. How many engineer-months sit in the dashboard, the telemetry pipeline and the multi-robot traffic logic? Writing that backend was defensible while every buyer needed one and nobody was giving it away. It is a different line item once a customer can download a compliant version this afternoon and ask what they are paying for in yours. General Motors and Procter & Gamble spent years in that working group, and companies that size reference the result in their contracts. Medical devices ran this same sequence with IEC 62304 for the software lifecycle and DICOM for imaging interchange. Both began as documents a buyer could cite, and both ended up as the line in the specification a vendor has to answer. Dave's take I have been on the losing side of this one. The document I keynoted was more complete than the one that beat it, and completeness turned out not to be what the industry was adopting. ISO 21423 is showing up with a document and a working implementation at the same time, so I would design to it now and stop watching the publication date. From Dave’s video library Dave on two federal files about electronic door handles, one from a parent locked out with a child inside, one from a driver who couldn't find the mechanical release after a head-on collision. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Teradyne Is Suing Over Three Universal Robots Patents. All Three Expire in March. →](/field-notes/2026-09-12-teradyne-jaka-patents-expire-in-march) [Nine Sites, $1.8 Million: How to Read a Hardware Company’s Traction →](/field-notes/2026-09-10-nine-sites-1-8-million-reading-hardware-traction) [Bedrock’s Excavators Dig With Nobody in the Cab. Bedrock Doesn’t Build Excavators. →](/field-notes/2026-08-28-bedrock-retrofit-excavators-nobody-in-the-cab) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Olympus Backed This Endoscopy Robot, Then Took an Option on Its Rival | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-09-19-olympus-backed-neptune-then-optioned-its-rival [← All Field Notes](/field-notes) September 19, 2026 Surgical Robotics Olympus invested in Neptune Medical in August 2024. Eleven months later it co-founded a company to build a competing robot, and took an option to buy that one outright. On September 18 Neptune's system cleared the FDA. The company Olympus can own outright is still early stage. ## A 510(k) for a colonoscopy robot The Triton 1 is a robotic endoscopy system, and the FDA cleared it through the 510(k) pathway for diagnostic and therapeutic work in the colon: screening, surveillance, endoscopic mucosal resection, endoscopic submucosal dissection. The pitch is stability. A colonoscope that wanders gives an incomplete look at the wall, and an incomplete look is how adenomas get missed. Neptune's evidence is the CARE 1 study, released in May: fifty patients, a 54.2 percent adenoma detection rate, no adverse events. The 35 percent it beats is the industry benchmark for adenoma detection. A fifty-patient study measured against a benchmark is enough to clear a 510(k), and enough to open a conversation with a gastroenterology chief. It isn't enough to carry a value analysis committee against an incumbent whose scopes are already in the room. ## Olympus bought into this market twice, on different terms Neptune raised $97 million in a Series D in August 2024. Sonder Capital and Olympus Corporation of the Americas were both in that round, and the company named Fred Moll chairman of the board in the same announcement. On July 25, 2025, Olympus and Revival Healthcare Capital co-founded Swan EndoSurgical to build an endoluminal robotic system for the GI tract. The initial combined investment was at least $65 million, and the structure runs up to $458 million if the milestones land. Olympus holds an option to acquire Swan at a predetermined value. In Neptune, Olympus owns a piece of a company it doesn't control, building a product that competes with its own scopes. In Swan, Olympus owns a path to the whole thing at a price it already agreed to. ## What the strategic investor gets A strategic investor gets what a financial investor doesn't: a working read on your technology, your timelines, and how hard the problem turns out to be. Olympus held that read for eleven months before it co-founded Swan. I have no way to know whether the two decisions were connected, and it changes little either way. The read was already theirs. You pay for that read in control. You never negotiated its terms, you can't switch it off yourself, and nothing in the term sheet says when Olympus starts building its own. That decision sits in their roadmap, and you don't get to read that one. The question before a strategic round closes isn't whether the partner is trustworthy. Most of them are, inside their own incentives. The question is what this investor learns about my company that they can't buy anywhere else, and what they could build with it. If the answer is that they learn how hard the problem is and how long it takes, you've funded someone else's build-versus-buy analysis with your own operating data. Neptune's counter is the calendar. Triton is cleared and Swan is still early stage, and a clearance is the one asset an option can't replicate. Every month Neptune spends putting systems in front of gastroenterologists is a month of procedure data and surgeon muscle memory that Olympus would have to buy back later. Dave's take Olympus did nothing improper here. It read its own incentives correctly and bought two instruments for two purposes. What I wouldn't do, in Neptune's seat or any other, is read an investment as a commitment. Olympus hasn't committed to anything yet, and the option is what lets them keep putting that decision off while Neptune does the expensive part. From Dave’s video library Dave works through a half-billion-dollar AI bill and where big spending stops buying an advantage. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Medtronic Paid $700 Million for a Surgical Robot It Can't Sell in the United States →](/field-notes/2026-09-05-medtronic-700m-robot-it-cant-sell-in-the-us) [A $400 Million Device Failed Its Trial. Its Engineers Bought It Back. →](/field-notes/2026-09-17-400m-device-failed-its-trial-engineers-bought-it-back) [Teradyne Is Suing Over Three Universal Robots Patents. All Three Expire in March. →](/field-notes/2026-09-12-teradyne-jaka-patents-expire-in-march) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # A Humanoid Holds Twice the Rare Earth of an EV. Getting It Back Was Decided at the Design Review. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-09-20-humanoid-rare-earth-decided-at-the-design-review [← All Field Notes](/field-notes) September 20, 2026 Industrial Robotics Taking a humanoid robot apart means getting through 10,000 to 15,000 components and recovering about four kilograms of neodymium magnet without ruining it. That is roughly double the rare-earth magnet in an electric car. Robert Belt put those numbers in The Robot Report this week, under the heading of recycling. They belong in a different document: the total cost of ownership model a fleet buyer runs before signing anything. ## What is inside one, and why it fights back A humanoid that has worked in a building carries navigation maps of that building, biometric logs, facial recognition recordings, and behavioral patterns of the people around it. All of that has to be destroyed to a standard somebody will sign, so the machine has to be alive enough to wipe before it is dead enough to scrap. The lithium packs keep that window narrow. A pack cut open inside a frame built for motion can go into thermal runaway. Belt notes that stored hydraulic and pneumatic energy turns components into projectiles. Very little of what survives can be sold on. A carbon-fiber limb that has carried load for three years looks identical to a new one, which is why Belt warns against putting reclaimed frames back into service at all. The magnets are the only part with real recovery value, and they are the hardest to get at. Crushing a humanoid the way you would crush a car cross-contaminates the neodymium with everything else in the machine, and cross-contaminated rare earth sells as scrap. Getting it out clean takes a trained person with hand tools, working around pinch points and flying fragments. Belt calls the work "a highly technical, high-stakes surgical endeavor." He is working with a recycling firm on humanoids, which is who gets handed the machine at the end. ## The decision was made at a design review Belt's fix is design for recycling: modular cartridges and standardized decoupling joints, a machine built to come apart. That is not a recycling proposal. It is an architecture proposal, and it has to win an argument years before the first unit ships. Decommissioning is unlike most operating costs in that respect. You can renegotiate a service contract and re-source a supplier. A bonded carbon-fiber assembly stays bonded. Whatever it costs to take the current generation of humanoids apart was set in design reviews that were not about disposal, and recycling expertise arriving later does not move that number much. ## Where the cost lands A capital asset gets depreciated against an assumed salvage value, and a leased asset gets priced against a residual. Either way, a buyer comparing a humanoid to the labor it replaces is running a total cost of ownership model with a terminal line in it. If retiring the machine costs money, that line is negative, and the payback the buyer was shown is wrong in a direction that surfaces in year four, after they have standardized on you. A demo has no year four in it. Humanoid companies are being judged right now on what the machine does in a video, and no video covers the day it comes out of service. Humanoids are not a special case here. Any product a customer eventually has to get rid of carries a cost that lands on somebody, and the design decides who. A portable medical device with a sealed battery has the same problem, and so does grid hardware full of coolant. Put the number on your own roadmap: what does it cost to make this thing go away, and has a customer or an investor ever seen that figure? Dave's take I don't think the humanoid companies are hiding the number, and it is more likely that nobody has needed it yet. The first invoice for taking one of these apart lands years after the design review that set its size, and that review is the only place the number moves. By the time the invoice arrives, the answer is bonded into every machine already built. From Dave’s video library Dave on the same gap in software: shipping something that works is cheap, and changing it later is the bill that arrives afterward. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Nine Sites, $1.8 Million: How to Read a Hardware Company’s Traction →](/field-notes/2026-09-10-nine-sites-1-8-million-reading-hardware-traction) [ISO 21423 Isn’t Published Yet. The Code That Speaks It Is Already Free. →](/field-notes/2026-09-18-iso-21423-not-published-the-code-is-already-free) [Kardium’s Cleanroom Builds 500 a Year. The Business Needs 30,000. →](/field-notes/2026-09-03-kardium-cleanroom-500-a-year-needs-30000) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Your FDA Clearance Covers One Configuration. Your Roadmap Is the Queue. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-09-21-your-fda-clearance-covers-one-configuration [← All Field Notes](/field-notes) September 21, 2026 Surgical Robotics LigaSure has sealed vessels in more than 45 million procedures across 65 countries. In June, Medtronic filed a 510(k) asking the FDA for permission to put it on the end of a robot arm. The clearance came through on September 16. The sealing technology is the same one surgeons have used for two decades. What's new is its position, and position is what the FDA clears. A clearance covers one configuration doing one job, which turns a platform company's roadmap into a queue of submissions that somebody has to put in order. ## Cleared for one combination The cleared instrument is the LigaSure RAS Maryland, and it runs only on the Valleylab FT10 generator attached to Medtronic's Hugo robot. Three pieces went into the file, and the clearance describes the combination. That jaw seals a seven-millimeter vessel in about two seconds, on the Hugo arm, through the FT10. The 45 million procedures behind LigaSure carry weight in that file. A predicate history tells the reviewer the sealing physics is understood, and says nothing about how the jaw behaves when a robot arm is holding it. The robot arm is what the June file had to cover, and a European file had covered it a year earlier. The CE mark came in July 2025, and the instrument launched across Europe that year, so surgeons outside the United States had the robotic sealer for more than a year before American surgeons could touch it. The evidence was the same in both places, and so was the hardware. What differed was which regulator was reading, and what each one had already let Hugo do. ## How much of a hospital Hugo is allowed into Hugo has been cleared in the United States since December 2025 for urologic procedures and nothing else. Last year American surgeons ran about 201,000 urologic procedures on da Vinci robots, against 1.25 million in general surgery and 468,000 in gynecology, so the clearance opened the smallest of the three doors. A urology-only robot is a real product and a narrow one. Every hospital conversation starts from a single service line. The economic case has to close on that line's volume, and the general surgeons who would fill the room are watching a machine they can't use yet. Medtronic filed 510(k)s for general surgery and gynecology in late April, and five months later the company is still waiting on both. The vessel sealer went in two months after those filings and cleared in three. ## What a founder gets to decide A company chooses two things in that queue: which file goes in first, and how the product is defined when it does. The review clock and the order the answers come back in belong to the agency. Framing decides what evidence the agency asks for. Define a device around a patient outcome and you've invited a clinical trial. Define the same hardware around what it does for a surgeon's hands and you've invited an engineering and usability file, which is cheaper and more in your control. That decision belongs to whoever owns the product, not to the regulatory consultant implementing against it. Clearance queues are the medtech version of something every platform business runs into. You sell one configuration, then find a gate between the platform working and the customer finishing a whole job on it. Sometimes the gate is a certification. Sometimes it's an integration, a driver, a security review, or a procurement category nobody has written yet. When I sequence a roadmap I keep coming back to one question: which configuration lets a single customer finish a single job end to end. Then I check whether that's the file that went in first, or whether the easy one went instead. Dave's take I read the Hugo queue as a product roadmap that happens to be written on FDA letterhead. Somebody chose which indication to chase first and which market got the sealer a year early, and both were commercial calls before they were regulatory ones. If you're building a platform and your regulatory consultant is choosing that order, you've put the wrong person in the seat. From Dave’s video library Dave traces the most-repeated statistic about AI pilots back to the report it came from, and finds a document that says something narrower than the headline. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Medtronic Paid $700 Million for a Surgical Robot It Can't Sell in the United States →](/field-notes/2026-09-05-medtronic-700m-robot-it-cant-sell-in-the-us) [Stryker's Surgical Headset Is FDA Authorized. Its Own Label Keeps the Monitors On. →](/field-notes/2026-09-06-stryker-headset-label-keeps-the-monitors-on) [The FDA Approved This Valve for Symptom-Free Patients. Medicare Started Paying Sixteen Months Later. →](/field-notes/2026-09-16-fda-approved-it-medicare-paid-sixteen-months-later) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # FDA Approved the First Sirolimus Balloon. A Subgroup Decides Who Buys It. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-09-22-first-sirolimus-balloon-approved-subgroup-picks-the-buyer [← All Field Notes](/field-notes) September 22, 2026 Medical Devices In 2003 Cordis won the first FDA approval for a sirolimus-eluting coronary stent. Drug-coated stents went on to become what cardiologists reach for when an artery narrows. This week the company collected another first. Selution SLR, a balloon coated with sirolimus, is now approved in the U.S. to treat in-stent restenosis, when an artery narrows again inside a stent placed there earlier. The only other coated balloon approved here for that job, Boston Scientific's paclitaxel-coated Agent, arrived in March 2024. Selution's pitch is that the patient leaves without another layer of metal. The approval rests on SELUTION4ISR, which randomized 418 patients to the balloon or to standard care. Standard care meant whatever the treating cardiologist would normally do, and 80% of the time that was another drug-eluting stent. At one year, target lesion failure (cardiac death, a heart attack in the treated vessel, or a repeat procedure on the lesion) hit 15.2% of balloon patients and 13.5% of controls. The balloon was allowed to come in up to 10 percentage points worse and still pass as noninferior. It came in 1.7 points worse. The control arm records the habit Selution has to break. The protocol let cardiologists treat those patients as they normally would, and four times in five they put in another stent. A balloon that asks interventional cardiologists to unlearn their training is a hard sell, and cath lab teams have been placing drug-eluting stents for more than two decades. ## What the subgroup says The FDA weighs the primary endpoint. A cath lab director deciding what to stock reads further down, to the single-layer numbers MedTech Dive reported. In patients whose artery held a single layer of stent, failure ran 14.2% on the balloon and 6.5% on standard care. In those patients the balloon did not meet noninferiority against a stent. At TCT last October, where the data first came out, Sanjit Jolly said on the panel that the wide margin "muddies the waters," and the panel said a head-to-head trial against stents would tell them more. The approval stands on a legitimate design, and the investigators built the control arm to look like American practice on purpose. That design also sets the limit of the claim. Cordis can say its balloon holds up against what cardiologists usually do. It can't say the balloon holds up against a fresh stent in the simplest restenosis, because there it didn't. Which patient, then, makes a cardiologist leave the stent on the shelf? The case for adding no more metal grows with every stent layer a vessel already carries, so expect the pitch to drift toward patients with two. The published results don't settle the balloon's advantage there. The trial's main comparison ran against usual care across everyone it enrolled. ## Pick the patient before the protocol That comparison was a product decision as much as a statistical one. It fixed which sentence Cordis could take to the FDA, and which question a buyer would still be asking afterward. An industrial robotics startup faces the same trade when it runs its pilot against a plant's average line: it can post a win the plant's best cell would have beaten, and the plant manager will ask about the best cell first. Cordis is already running its next test. SELUTION4DeNovo, a 960-patient IDE trial in small coronary vessels that have never held a stent, has finished enrolling. Its comparator, and its subgroups, will set what Cordis can claim in those vessels. Dave's take I push founders to write down the patient or customer they must win on before anyone drafts a protocol, then size the study to show it, even when that costs them a broader label. Approval tells you that you may sell. The subgroup tells you to whom. From Dave’s video library Dave takes apart a BCG study where AI made consultants 40% better on tasks inside its range and 19 points worse on one built to sit outside it, and what the average hides. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Your FDA Clearance Covers One Configuration. Your Roadmap Is the Queue. →](/field-notes/2026-09-21-your-fda-clearance-covers-one-configuration) [The FDA Approved This Valve for Symptom-Free Patients. Medicare Started Paying Sixteen Months Later. →](/field-notes/2026-09-16-fda-approved-it-medicare-paid-sixteen-months-later) [A $400 Million Device Failed Its Trial. Its Engineers Bought It Back. →](/field-notes/2026-09-17-400m-device-failed-its-trial-engineers-bought-it-back) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # GM Shipped Patriot Interceptor Parts in 22 Days. Lockheed Is Buying the Rate. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-09-23-gm-shipped-patriot-parts-in-22-days-lockheed-needs-the-rate [← All Field Notes](/field-notes) September 23, 2026 Defense Hardware Lockheed Martin signed a production contract with GM Defense on August 6. Twenty-two days later, the first batch of GM-made housings for the PAC-3 MSE arrived. The PAC-3 MSE is the interceptor a Patriot battery fires. Tim Cahill, who runs Lockheed's missiles business, called the turnaround extraordinary, and Defense News noted that parts like these traditionally take months or years. The announcement leaves out three things: how many housings were in the shipment, whether they have passed qualification, and how many GM can make in a month. Those three blanks are what Lockheed is short of. ## A date is one delivery The easy read is that a car company just outran the defense industrial base. But a first shipment proves only that GM can cast and machine a housing to a drawing, and GM has been casting and machining metal at automotive volume for generations, so that was never much in doubt. An interceptor line needs something harder from a supplier: the same part, to the same spec, every week, with the records that let it fly. A date is one delivery. A rate is a business. The Pentagon's own paper trail says the rate is the scarce thing. On August 5, Deputy Defense Secretary Steve Feinberg asked defense companies for "significantly faster, more aggressive delivery schedules" on 16 critical programs. PAC-3 MSE and THAAD were on the list, and the deadline was August 26. Defense News reported on September 21 that the Pentagon still hasn't said how many companies replied or what schedules they offered. The Center for Strategic and International Studies estimates production takes about 42 months for a PAC-3 MSE interceptor and 53 for a THAAD round. Set 22 days against 42 months. The first box doesn't move the 42 months unless the housings keep coming every month and every lot passes inspection. Lockheed is spending $8 billion to $9 billion to expand munitions production across more than 20 U.S. sites. That money builds Lockheed's own plants, and the plants still wait on parts from suppliers like GM. ## What a prime is buying from you Speed to first article is the number you control as a supplier, and it demos well. A prime's supply chain team wants two other answers from you: what rate you'll commit to hold, and how long until your part is qualified to go into the product. Those two answers decide whether you're a line item in their production plan or a pilot they talk about at a conference. GM sits on the other side of that table every day. A new part goes into a GM vehicle after a production run shows the process can make it, again and again, inside tolerance. A prototype that fits is where that approval starts. GM Defense is now being judged by the kind of rule its parent company has applied to its own suppliers for decades. A battery pack startup courting an automaker meets that approval process before it gets to negotiate price. A contract manufacturer quoting a medical device company meets its version in process validation. Whatever your vertical, you need to know what your version of that production run looks like, and you need to show it before the buyer asks. It's built from fixture design, incoming material inspection, the second shift, and a yield number over 30 days instead of three. A prime checks each of those before it puts your part on a schedule. Dave's take If you sell parts to a prime, put the monthly rate you'll commit to hold on your first slide and your qualification schedule on the second slide, and move the first-article date to the appendix. Lockheed has told us when the first box arrived. The announcement I'd want to read says how many housings GM ships in March. From Dave’s video library Dave makes the case against vanity dashboards and names the five numbers worth tracking instead. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [The Army Bought Its First Production Laser. The Factory Was Funded in March. →](/field-notes/2026-09-08-army-first-production-laser-factory-funded-in-march) [Kardium's Cleanroom Builds 500 a Year. The Business Needs 30,000. →](/field-notes/2026-09-03-kardium-cleanroom-500-a-year-needs-30000) [Nine Sites, $1.8 Million: How to Read a Hardware Company's Traction →](/field-notes/2026-09-10-nine-sites-1-8-million-reading-hardware-traction) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # CATL’s Licensee Pays for the Factory. CATL Plans to Sell It the Cells. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-09-24-catl-licensee-pays-for-the-factory-catl-sells-the-cells [← All Field Notes](/field-notes) September 24, 2026 Commercialization A technology license is a bet that somebody else will build a factory. Last week CATL placed one in Egypt. BME is a battery maker founded by the truck builder MCV and the automotive supplier Auto D. CATL will license its battery pack technology to BME, supply the production equipment and train BME's staff. BME funds and builds the plant. Caixin puts the opening investment above 2 billion Egyptian pounds, about $38 million, for a line that starts at 1 gigawatt-hour a year of packs for heavy commercial vehicles and could reach 5 in a second phase. Alongside the license sits a proposed arrangement for BME to buy battery cells from CATL. In Caixin's account, the structure lets CATL earn service profits and lock in cell supply orders. This is CATL's first license for pack technology. Its cell licensing already runs in the United States. Ford paid for its own lithium iron phosphate battery plant, CATL holds no shares in it, and CATL supplies the technology license and operating services. ## What CATL is paid for A pack is the assembly around the cells: modules, wiring, cooling, the battery management electronics and the enclosure. BME gets the know-how to build that assembly in Egypt, with a target of sourcing 40% of the components locally. The cells, if the proposed arrangement holds, come from CATL. So CATL collects for services while the plant goes up, and for cells on every pack the line ships once it runs. BME carries the $38 million and the job of finding truck makers who want the packs. ## What 24M was paid for 24M Technologies, cofounded by MIT professor Yet-Ming Chiang, spent years on a cheaper way to make lithium-ion cells and was once valued above $1 billion. Its process worked with the lithium-ion chemistry the industry already used, and 24M sold it as licenses to companies that would build their own factories. The largest bet came from Freyr, which signed licenses in December 2020 for Norway and in October 2021 for the United States. Today's episode of The Teardown, at the end of this note, tells that story in full. Freyr's plans moved. It cut jobs in Norway and pulled out of Giga Arctic, the full-scale plant planned for Mo i Rana. On November 4, 2024, it ended both licenses, paid 24M $3 million for services already provided, and forfeited 6,975,956 shares of 24M's Series G preferred stock. The same week it agreed to buy Trina Solar's 5-gigawatt solar module factory in Wilmer, Texas, for $340 million. This March, The Information reported that 24M would shut down and auction its property. I don't know what 24M's contracts said, and Chiang declined to speak on the record about the shutdown. The whole industry had a bad stretch, too: MIT Technology Review points to cooling U.S. EV demand and gutted Inflation Reduction Act provisions. But the termination shows what 24M held once Freyr stopped. It had been paid for services rendered, and anything it stood to earn from cells made under the license needed a Freyr factory that was no longer in Freyr's plans. ## The question at signing The two deals split on what the licensee keeps buying after the plant opens, and for CATL that's cells. A process company has only the process, so its revenue lives inside the licensee's capital plan. Freyr's capital went to solar panels. Anyone who licenses a method and ships no product of their own carries the same exposure, whether the method is an electrolyzer stack design or a surgical instrument licensed to a larger device maker. Sit where 24M sat in 2020, with Freyr's signature fresh, and the question is what you still own if your biggest licensee never builds. The answers get written on signing day or not at all: a date by which the line has to run, a minimum annual payment that doesn't wait for output, exclusivity that lapses when the licensee stalls so the territory can go to someone else. CATL's answer is a component the licensee's line can't run without. A process startup usually has no such part, so it needs those terms in the contract. Dave's take Licensing looks like the capital-light route for a hard-tech startup, and on the balance sheet it is. The capital still has to be spent, just by someone whose plans you don't control. Price the license for the year that someone changes their mind. From Dave’s video library In today's Teardown, Dave walks through 24M Technologies, which licensed a cheaper way to make lithium-ion cells, and what happened when its biggest licensee bought a solar factory instead. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [GM Is Building Battery Cells It Will Never Sell So It Can Judge the Ones It Buys →](/field-notes/2026-09-13-gm-900m-battery-cells-it-will-never-sell) [Cellares' Machine Met Every Release Spec. Bristol Myers Squibb Still Walked Away. →](/field-notes/2026-09-07-cellares-met-every-release-spec-bms-still-walked) [Teradyne Is Suing Over Three Universal Robots Patents. All Three Expire in March. →](/field-notes/2026-09-12-teradyne-jaka-patents-expire-in-march) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # The FDA Panel Said Yes to Grail’s Cancer Test. It Said No to the Word Early. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-09-25-galleri-panel-said-yes-to-the-test-no-to-early [← All Field Notes](/field-notes) September 25, 2026 Diagnostics On September 23, an FDA advisory panel voted that the benefits of Grail's Galleri blood test outweigh its risks. The same panel asked that Grail's marketing avoid describing the test as an early detection tool. Galleri belongs to a category the industry calls multi-cancer early detection. Grail went into the room selling earlier diagnosis, and the panel's yes covers a narrower product than the one Grail has been selling. ## What the panel voted for All ten voting members called the test safe. On whether the benefits outweigh the risks the vote was seven to two, with one abstention, and on effectiveness it was six to four. The recommendation covers prescription screening for adults 50 and older, added to mammograms and colonoscopies. The panel also wanted the marketing materials to tell people to keep having those screenings. The FDA's review of the NHS-Galleri study, reported by Targeted Oncology, found that the test caught 13.6% of stage I cancers and 60% of stage IV, so Galleri sees cancer best once it has spread. The panel noted those stage figures may carry some bias, but the trial's primary endpoint pointed the same way. The trial set out to cut the combined number of stage III and IV diagnoses, and it missed. Grail has since pointed to a drop in stage IV alone. Phil Castle, who runs the National Cancer Institute's cancer prevention division, told the panel the data were too preliminary to show any clinical benefit. ## The word was the product Cristian Tomasetti of City of Hope told NPR that an approval would validate the whole concept, and Grail did win something on Wednesday. But the panel approved the test and trimmed the claim, and for a screening product the claim is most of what the customer pays for. Nobody spends $949 on a blood draw for its specificity. They spend it on the hope of hearing about cancer while it can still be treated. In a 510(k), the most expensive word is usually "better," because a superiority claim can push a device out of the equivalence pathway it was counting on. Galleri is going through premarket approval instead, and the same logic applies: every word in Grail's marketing is a claim that has to ride on evidence Grail already holds. Grail's evidence supports finding cancer, and the panel didn't think it yet supports finding it early. ## Where the claim got decided The panel's objection to "early" traces back to two choices Grail made when it designed its studies. Both large trials ran prospectively on an earlier investigational version of the assay, and the performance figures the panel saw came from retesting stored plasma with the current Galleri. That is a legitimate method, and it is also the kind of evidence [an advisory panel, reading the file cold](/field-notes/2026-08-08-fda-advisory-panel-reads-your-file-cold), tends to discount. The endpoint was the second choice. Tomasetti argued that three years was too short to show a stage shift. He may be right, but Grail picked the window. For any company whose product is a measurement, the question I'd put on the table early is which sentence goes on the box, and whether the study about to be funded can carry that sentence on the version that will ship. A battery-health sensor that promises to flag a failing cell before it fails, or an inspection system sold on catching cracks before they spread, is selling a timing claim, and timing is the hardest thing to prove. Grail's stakes are higher because Congress has already approved Medicare coverage for multi-cancer tests the FDA approves, so whatever the label ends up saying will describe what Medicare pays for. Dave's take I'd still call this a good week for Grail, and a better one for Abbott and the other companies lining up behind it, who now know which word a panel won't accept on today's evidence. But founders spend years building the product and a few weeks writing the sentence that describes it, and the sentence is what gets approved. From Dave’s video library In this week's Failure Modes Friday, I walk through the Challenger booster seals that were burning on shuttle flights for years before the 1986 launch. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [FDA Reclassified This Cancer Test Category. Fewer Than Five People Commented. →](/field-notes/2026-09-04-fda-reclassified-cancer-tests-five-comments) [Your FDA Clearance Covers One Configuration. Your Roadmap Is the Queue. →](/field-notes/2026-09-21-your-fda-clearance-covers-one-configuration) [The FDA Approved This Valve for Symptom-Free Patients. Medicare Started Paying Sixteen Months Later. →](/field-notes/2026-09-16-fda-approved-it-medicare-paid-sixteen-months-later) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # The FDA’s Draft Surgical Robot Guidance Lets the Hardest Procedure Cover the Rest | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-09-26-fda-robot-guidance-hardest-procedure-covers-the-rest [← All Field Notes](/field-notes) September 26, 2026 Surgical Robotics On September 25 the FDA published its first draft guidance written specifically for surgical robots. The part a founder should read first is Appendix A. It describes umbrella and covered procedures. If a company shows its robot performing a hard, higher-risk operation, that data can often support authorization for easier procedures in the same specialty without new performance data for each one. Which operation a company proves first now sets how much of a specialty it can sell into. ## What one umbrella covers The colorectal table shows the spread. A low anterior resection with inter-sphincteric resection, a deep pelvic operation for rectal cancer, sits at the top as the umbrella for seven covered procedures, from the colectomies down to appendectomy. A right hemicolectomy, which removes the right side of the colon, covers three of those seven. In gastric surgery, a Roux-en-Y bypass covers seven, including sleeve gastrectomy, gallbladder removal and hiatal hernia repair. The reasoning is that the harder operation contains the same or very similar steps as the easier ones, and the agency calls the approach least burdensome. The draft sets limits. The umbrella approach adds surgical procedures; it does not stretch to a new indication such as treating a particular disease. Umbrella and covered procedures generally sit in the same specialty, though the draft leaves room for one umbrella to span more than one anatomical area later. There is no fixed list, and the FDA asks companies to bring their proposed pairings to a Q-Submission, its pre-submission feedback program. The guidance also leaves out remotely teleoperated robots, autonomous systems and robots that work through blood vessels, so companies building those still negotiate their evidence with the agency one device at a time. ## The cheap study buys one procedure The draft says a robot with no prior U.S. authorization often needs clinical data. The cheaper route is to run that study on the simplest procedure a startup can defend, because a simple procedure means a smaller, faster study. Under this draft, that study buys roughly the procedure it tested. The harder study costs more, takes longer and asks more of the surgeons running it, and if it succeeds it can carry a whole row of the table. Choosing between those two studies belongs to whoever owns the product. It turns on which hospital departments the company needs as buyers in year three, what those surgeons expect the robot to handle on day one, and whether the company has the cash and the clinical sites to survive the harder trial. ## Plan for the reviewer after the Q-Sub The draft sends the umbrella choice through a Q-Sub, and a founder should use that feedback without banking on it. At Galen, our Q-Sub described a 16-subject cadaver study run by residents, and the FDA wrote back that it sounded fine. In substantive review, the first question was whether we planned to market to residents. I recruited 16 attending surgeons and re-ran the study at Johns Hopkins in two weeks. A regulatory attorney later told me that agreement in a Q-Sub doesn't obligate the team that reviews the file months later. So pick an umbrella the pre-sub supports, and design the study so it holds up for a reviewer who never saw that letter. Outside surgery, the umbrella choice shows up as a reference customer, a qualification run or a pilot site: a first proof expected to carry the next several sales. When I work with a hardware company, I ask which of those it wants its first customer to pay for. This draft puts a count on that choice for surgical robots. In its colorectal example, the hard operation carries seven procedures and the easy one carries only itself. Dave's take I would rather see a surgical robotics company spend longer on an umbrella procedure than race to an authorization that fits one line of the OR schedule. The draft is open for comment for 60 days, and any company with a pivotal study still on the whiteboard should read Appendix A before it locks the protocol. From Dave’s video library Dave explains how to find the single riskiest assumption in a product and build the cheapest test that settles it. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Your FDA Clearance Covers One Configuration. Your Roadmap Is the Queue. →](/field-notes/2026-09-21-your-fda-clearance-covers-one-configuration) [Olympus Backed This Endoscopy Robot, Then Took an Option on Its Rival →](/field-notes/2026-09-19-olympus-backed-neptune-then-optioned-its-rival) [Medtronic Paid $700 Million for a Surgical Robot It Can't Sell in the United States →](/field-notes/2026-09-05-medtronic-700m-robot-it-cant-sell-in-the-us) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Robotaxis Proved the Driving. How to Count the Service Work Every Demo Leaves Out | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-09-27-robotaxis-proved-the-driving-count-the-service-loop [← All Field Notes](/field-notes) September 27, 2026 Commercialization Dan Keene spent years building a battery-swapping network for shared scooters. The operators who survived that market, he writes, were the ones who could charge, repair and reposition tens of thousands of vehicles a day. On Saturday he applied the same lesson to robotaxis in The Robot Report. His company, Aseon Labs, builds robotic stations that charge, clean and inspect autonomous cars, so he sells the fix for the problem he describes. His arithmetic holds up anyway, and it gives any founder with hardware in the field a way to count the work a demo leaves out. ## What happens between rides Keene starts from Waymo's own numbers: more than 220 million autonomous miles, around half a million paid rides a week, 14 cities. Uber operates in more than 15,000. When a Waymo car finishes a ride with a spilled latte on the back seat, it can drive up to 15 miles, empty, to a central depot to be cleaned, charged and inspected. He puts each of those resets at roughly two hours. Of the 86 million Waymo miles reported in California through 2025, 54% carried a passenger, according to state data he cites. Depot runs are one source of those empty miles, and he is clear they aren't the only one. The driving is autonomous. The servicing, in Keene's telling, is manual, and it happens in a building the operator had to find, lease and wire for power. That building is where the cost of each new city piles up. He cites California's regulator targeting six months just to energize a standard charging site, 1.9 years for a new circuit, 2.8 years for a substation upgrade and 8.9 years for a new substation. Those clocks start over in every city, and Phoenix, Zurich and Riyadh each need a different depot. ## Count the touches per unit of work The same count works for anything that does repeated work in the field. Pick the unit your customer pays for: a ride, a procedure, a pallet moved, a sample run. Then list every time a person or a machine has to handle the product before it's ready for the next unit. Charging, cleaning, inspection, calibration, restocking consumables, clearing a fault, moving it back into position. For each touch, write down who does it, where it happens and how long it takes. That list is your service loop. A demo shows none of this. It runs one unit, on a charged machine an engineer prepared the night before. Keene observes that every robotics demo ends with someone off camera carrying in a charging cable. That person belongs in the financial model, with a wage, a drive time and a building to work from. Surgical robotics runs the same loop, and hospitals count it. After every case the instruments go to sterile processing and the room gets turned over for the next patient. A robot that adds, say, twenty minutes of turnover to each case can cost a busy room its last case of the day, and the hospital's value analysis committee will run that math whether the company ran it first or not. ## The second number is time per site The touch count sets your margin. Time to stand up servicing at a new site sets how fast you can grow. Keene's answer is a standard servicing unit that arrives on a flatbed and goes live within a day. Whether his version works is open, and he says himself that his company is early. The principle under it holds for any fleet: if each new site needs custom construction, permits and a utility upgrade, growth moves at the speed of the slowest permit office, and more capital won't speed it up. Waymo has already handed part of its loop to other companies. Zeekr, a carmaker, builds its vans, and four outside operators (Moove, Avomo, Avis and Flexdrive) run its depots. Deciding which touches you own, which you contract out and which you design out of the product is a product decision. A sensor that keeps itself clean removes a touch. A battery that swaps in minutes shortens one. The cheapest place to make those calls is the design review, while the service loop is still a drawing and a change costs an engineering revision. After launch the same change costs a lease. Dave's take When I review a hardware roadmap, I ask to see the service loop drawn next to the product architecture, with a name and a time on every touch. The economics of a machine in the field have to work on the first site, and the loop is most of that math. A team that can't draw that loop is still selling the demo. From Dave’s video library Dave on why a working demo isn't demand, and why a paying buyer is the clock that decides when a build is ready. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Nine Sites, $1.8 Million: How to Read a Hardware Company's Traction →](/field-notes/2026-09-10-nine-sites-1-8-million-reading-hardware-traction) [A Humanoid Holds Twice the Rare Earth of an EV. Getting It Back Was Decided at the Design Review. →](/field-notes/2026-09-20-humanoid-rare-earth-decided-at-the-design-review) [GM Shipped Patriot Interceptor Parts in 22 Days. Lockheed Is Buying the Rate. →](/field-notes/2026-09-23-gm-shipped-patriot-parts-in-22-days-lockheed-needs-the-rate) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # The Navy Put a Startup’s Hypersonic Missile Into Production. Its Investors Paid for the Development First. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-09-28-castelion-investors-paid-the-navys-development-bill [← All Field Notes](/field-notes) September 28, 2026 Defense Hardware On September 14 Castelion announced a Navy delivery order worth up to $200 million to manufacture, test and deliver a first production lot of Blackbeard, a hypersonic strike missile for carrier fighters. The Navy had selected Blackbeard in February, seven months earlier. In 2024 it had cancelled HALO, its own program for an air-launched hypersonic anti-ship missile, citing budget constraints. The Navy's release calls the Blackbeard award one of the first successes of its rapid capabilities office. Castelion says it developed the missile with hundreds of millions of dollars of private investment, and that money is what let the Navy move this fast. ## What the investors paid for The award is a Phase III order under the Small Business Innovation Research program, the phase meant for turning earlier work into products the government buys. The Navy's release says the order uses a provision in the 2026 defense authorization act written for technology developed with private capital. That route lets the Navy move it straight into production. The release says it avoids millions of dollars in redundant non-recurring engineering, the one-time cost of designing and qualifying a product before the first unit ships. A government development program such as HALO carries that engineering on its own budget. Before this order the Navy made three smaller awards: nearly $50 million in February, when it selected Blackbeard for a program called the Multi-mission Affordable Capacity Effector; $23.4 million in June for 50 pre-production rounds and their shipping containers; and about $90 million in August to bring the missile to early operational capability. They are separate from the $200 million production order, and each was sized for a buyer checking a design that already existed. ## The unit cost came first Castelion's other advantage is a price it set before the buyer asked. USNI News reports that the company aims to keep Blackbeard near half a million dollars a round. The Pentagon said in July it was interested in up to 12,000 of them over five years. At half a million dollars each, that is about $6 billion. Castelion's CEO, Bryon Hargis, says the missile was engineered for the production line from its first day. Hospitals have enforced the same unit-cost discipline since COVID. Their capital committees stopped paying for equipment whose economics might work later, so a surgical robot's price has to make sense at its very first hospital. ## The stage your buyer pays for Defense has a stage where the buyer pays for work that might fail: development and qualification. A hospital has one too, often a clinical evaluation its own staff run before it buys. Its pace follows the buyer's budget cycle. Castelion paid for the Navy's version of that stage with private money, so by the time the Navy looked, the risky engineering was done and what remained was a purchase. Castelion's investors held the technical risk for that stage before any order existed. The company's revenue also rests on how much the military actually buys. The Pentagon described its 12,000 as an interest. The Navy's fiscal 2027 budget request funds 353 missiles under the MACE program, for $156 million. A founder weighing the same move should model the year where the customer buys the first lot and then waits. Dave's take The Navy's acquisition office earned its credit here by building a contract path that lets privately funded hardware go straight to a production order. The question I'd put to any hardware founder's board is which of the buyer's risky stages we're willing to pay for ourselves, and what that costs us in dilution. Castelion's backers answered it before anyone had ordered a missile. From Dave’s video library This week's Monday MedTech looks at an FDA panel vote on Grail's Galleri blood test and a European clearance for da Vinci 5 in heart surgery, two yeses that cover less than the technology promises. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [GM Shipped Patriot Interceptor Parts in 22 Days. Lockheed Is Buying the Rate. →](/field-notes/2026-09-23-gm-shipped-patriot-parts-in-22-days-lockheed-needs-the-rate) [The Army Bought Its First Production Laser. The Factory Was Funded in March. →](/field-notes/2026-09-08-army-first-production-laser-factory-funded-in-march) [Nine Sites, $1.8 Million: How to Read a Hardware Company's Traction →](/field-notes/2026-09-10-nine-sites-1-8-million-reading-hardware-traction) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Microbot’s FDA-Cleared Robot Now Meets the Hospital’s Value Analysis Committee Earlier | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-09-29-microbot-robot-cleared-hospital-committee-comes-first [← All Field Notes](/field-notes) September 29, 2026 Surgical Robotics Microbot Medical has changed the order in which it sells its LIBERTY robot. At an analyst meeting in New York last Thursday, the company said its reps now engage a hospital's value analysis committee earlier in each sale, and that the committee's review takes four to six months on its experience so far. LIBERTY is already FDA-cleared and went on full market release in April. The clock Microbot is now trying to start sooner is the hospital's. ## What the committee weighs A value analysis committee is the group inside a hospital or health system, drawn from clinicians, supply chain staff and finance, that decides whether a product goes on the shelf. It weighs clinical benefit against cost for the whole system, and the physician who asked for the product is one voice in the room. LIBERTY is a remotely operated robot for peripheral endovascular procedures, which are done with catheters threaded through blood vessels. Microbot describes it as the only FDA-cleared system of its kind that is single-use. It aims to reduce radiation exposure and physical strain, which gives the clinicians on the committee something to weigh. The finance people get something to weigh too. Every procedure uses up a robot, so the hospital's cost arrives case by case, and so does Microbot's. In the same update Microbot said it expects a second manufacturing line and more efficient processes to bring positive gross margins in 2027. Any committee reviewing LIBERTY now is looking at a per-case price set while the cost of building each unit is still coming down. ## Why Microbot moved the committee up Microbot's stated reason is that engaging committees early will speed adoption across health systems with several hospitals. The expansion behind that reason is fast. The company went from four sales territories at the end of March to eight by midyear and plans twelve by December. It has hired three more area sales managers and a regional sales director to push west, where it has its first California customer. A rep in one of those new territories who wins over a physician and only then opens the committee file adds four to six months before the first order can arrive. Starting the committee review alongside the clinical evaluation, as Microbot now does, lets the two finish closer together. ## The cost case starts in the design review A startup still in development can start on the committee's questions before it has anything to sell. The committee will want to know what the device costs per case, what it replaces, what it saves the hospital in time or staff, and what happens to the price as volume rises. Those answers depend on decisions made in the design review: the architecture, the disposables, the bill of materials. Each is cheapest to change there. After launch the levers left are price and manufacturing cost, and Microbot is working on manufacturing cost now. My own lesson came from a surgical robot aimed at procedures that paid hospitals less than the big robotic operations, so the robot's price had to come in lower before any hospital would take a meeting. Since COVID, the hospitals I've sold to have had no budget for science experiments. Today's Founder Tip Tuesday video covers Stripe's habit of reviewing big projects at 20 percent and 80 percent of the way done, and a review at 20 percent can still change the direction of the work. A friendly health system's supply chain lead can sit in on that early review, and the questions they raise become design requirements while there's still time to meet them. Dave's take I'd want the per-case cost model on the table at the first architecture review, when changing a disposable costs a drawing revision. Microbot is bringing its per-unit cost down after launch with a second manufacturing line, which is a far more expensive place to do that work. From Dave’s video library Dave on how Stripe reviewed big projects at 20 percent and again at 80 percent of the way done. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [The FDA Approved This Valve for Symptom-Free Patients. Medicare Started Paying Sixteen Months Later. →](/field-notes/2026-09-16-fda-approved-it-medicare-paid-sixteen-months-later) [Your FDA Clearance Covers One Configuration. Your Roadmap Is the Queue. →](/field-notes/2026-09-21-your-fda-clearance-covers-one-configuration) [Robotaxis Proved the Driving. How to Count the Service Work Every Demo Leaves Out →](/field-notes/2026-09-27-robotaxis-proved-the-driving-count-the-service-loop) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # The FDA Would Credential Generative AI Like a Clinician. You Still Answer for a Model You Don’t Own. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-09-30-fda-genai-devices-credentialed-model-you-dont-own [← All Field Notes](/field-notes) September 30, 2026 Regulatory Medical device software has been verified the same way for decades: write down the inputs, run them, and compare each output to a known answer. On August 18 the FDA's device center said that method probably won't work for generative AI, because the range of possible inputs and outputs is too large to test. Its new discussion paper floats a replacement modeled on how clinicians get credentialed. The proposal shifts much of the proof from the day of submission to every day after it, and onto a foundation model the startup didn't build. ## Credentials, then monitoring The paper, Considerations for the Regulation of Generative AI-Enabled Medical Devices, is not guidance and sets no requirements. Comments close October 19. It sketches a risk chart that weighs how much the software does on its own against how bad the harm is when someone relies on a wrong answer. Then it describes a two-part evaluation. The finished device, in the configuration it will ship in, gets benchmarked on clinical knowledge, safety behavior, robustness and knowing when to defer to a human. A clinical confirmation step checks real-world performance, with the rigor scaled to risk, anywhere from a retrospective look at patient data to a randomized trial. The agency also says it may accept more uncertainty before launch in exchange for more monitoring after. Its options include re-running the benchmark on a schedule and after defined changes, independent clinicians reviewing samples of real inputs and outputs, and watching for drift. The paper asks whether AI agents could do some of the supervising. ## The component nobody can freeze Most generative-AI devices will sit on a foundation model built by someone else. The paper's proposed help is a voluntary Foundation Model Master File. Model developers could give the agency confidential information on architecture, training data, known failure modes and update commitments, and a device maker could reference that file with the developer's permission. The paper is explicit that a master file authorizes nothing. The device maker stays responsible for a component it didn't build and can't fully inspect, from a supplier with no obligation to tell the FDA anything. Cooley, the law firm, reviewed the public comments filed so far and reports worry that some models are updated hundreds of times a day. The paper names a change to the underlying model as a trigger for re-running the benchmark. A fast-moving vendor could hand a device maker a stream of regulatory events it never scheduled. ## Who this lands on The proposal reaches fewer founders than the word AI implies. The AI already shipping in cleared devices, and in the surgical robotics research I work around, is image and video processing with no language model in the loop. Those teams can read this as someone else's problem. The paper is aimed at founders putting a chat interface, a summarizer or an agent in front of a clinician or a patient, and it raises the idea of ranking a patient-facing function as higher risk than the same function aimed at a clinician. For a founder putting a chatbot or an agent in front of patients or clinicians, the decisive work comes before any submission: four questions for the model vendor, with the answers written into the contract. Can we name the exact model version we validated? Can we keep running that version after the vendor moves on? Will we get notice before it changes? Can we re-run our own benchmark within days when it does? If any answer depends on the vendor's goodwill, the model belongs in the risk file as a single-source part. Hardware teams already know how to handle one. A sole-source chip with no end-of-life notice gets a mitigation, a second source or a redesign. I expect the same treatment to reach industrial robots and defense systems that run vision-language models, because their buyers can't list every input either. They'll ask for competency evidence and a monitoring plan, and the seller will need the same four answers. Dave's take The competency model is a sensible idea and worth supporting in the docket. The master-file section is the weak spot. A voluntary program works only if the large model developers volunteer, and nothing in the paper requires them to, so the only master file a startup can count on today is the one it writes into its own vendor contract. From Dave’s video library Dave traces the study behind the claim that AI beats lawyers at contract review, and finds the AI helped build the answer key it was graded against. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [The FDA’s Draft Surgical Robot Guidance Lets the Hardest Procedure Cover the Rest →](/field-notes/2026-09-26-fda-robot-guidance-hardest-procedure-covers-the-rest) [When No Law Fits Your Product, Regulators Reach for the Nearest One →](/field-notes/2026-09-15-no-law-fits-regulators-reach-for-the-nearest-one) [Your FDA Clearance Covers One Configuration. Your Roadmap Is the Queue. →](/field-notes/2026-09-21-your-fda-clearance-covers-one-configuration) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Terumo Is Buying Arsenal Medical Before Its Pivotal Trial Has Data. Medtronic Paid for the Proof. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-10-01-terumo-buys-arsenal-medical-before-pivotal-trial-data [← All Field Notes](/field-notes) October 1, 2026 Medical Devices Arsenal Medical agreed on Tuesday to sell itself to Terumo Neuro. NeoCast, the product Terumo wants, is investigational and can't be sold in the United States. The trial meant to win its approval started in August and isn't scheduled to finish collecting its primary data until January 2028. Terumo is buying more than a year ahead of that data. The trial it is buying into only has to match a product Medtronic already proved. NeoCast is a liquid that a physician injects through a catheter to block the middle meningeal artery. That artery is believed to feed chronic subdural hematoma, a collection of blood on the surface of the brain that mostly affects older patients. The usual treatment is surgery to drain the blood, and the hematoma can come back afterward. ## Medtronic proved the procedure Medtronic sponsored a randomized trial, EMBOLISE, in which some patients had surgery alone and others had surgery plus embolization with Onyx, Medtronic's own liquid embolic. Adding Onyx cut the rate of recurrence or reoperation by roughly 63%, NeuroNews International reported. The results ran in the New England Journal of Medicine in November 2024, and thirteen months later Onyx became the first product in the United States indicated for the procedure. ## What Arsenal's trial has to show Five months after that, in May 2026, the FDA approved Arsenal's pivotal trial, RADIANT. It will randomize about 360 patients two to one between NeoCast and Onyx, at up to 35 sites in the United States and Australia. The ClinicalTrials.gov listing calls the primary comparison non-inferiority: at 90 days, NeoCast patients have to do no worse than Onyx patients on a combined count of treatment failures, meaning hematoma that remains or returns, reoperation, or a major adverse event. Medtronic had no such comparator. Until December 2025 no embolic was approved for this use, so EMBOLISE had to beat surgery alone. RADIANT compares one embolic with another and takes the procedure's benefit as settled. Arsenal's pitch for NeoCast over Onyx is the formulation, which contains no solvent and no adhesive. CEO Upma Sharma has said many physicians still rely on older embolics that can be difficult to use and contain one or the other. NeoCast is designed to be delivered without pain, and in Arsenal's first-in-human study in hematoma, 10 of the 24 patients were treated without general anesthesia. The hematoma resolved completely in 18. ## What Terumo is buying Terms weren't disclosed. What Arsenal brings is NeoCast itself, a pivotal trial with FDA approval and its first patients enrolled, and two small first-in-human studies. One is the 24-patient hematoma study. The other used NeoCast to cut off blood supply to brain tumors before surgery. Of those two uses, the one that reached a pivotal trial and a buyer is the one where another company had already proved the therapy. Whatever Terumo paid, the open question it bought is whether NeoCast performs about as well as Onyx in 360 patients, and the human evidence so far in this use is one 24-patient feasibility study. Medtronic's open question was whether the procedure worked at all. ## Going first Going first cost Medtronic a trial and bought it a head start: Onyx has been on label since December 2025, and NeoCast's primary data isn't due until 2028. I've been the one going first. Galen Robotics' De Novo created a new regulation, and any later company can cite our device as its predicate. We kept the ability to clear our next feature against our own device. The next company gets to skip the argument we had to make. Dave's take When I weigh a first indication now, I start with who has already proved the therapy and whether the FDA will accept a comparison against their product. I'd still want to read Arsenal's terms before calling this a good exit for the people who built the material, because how much of the payment waits on the 2028 data isn't public. From Dave’s video library In today's Teardown, Dave looks at Pallet, which built more than 6,000 shelter beds for cities and counties and closed in September, citing a shift in funding and political priorities. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [Olympus Backed This Endoscopy Robot, Then Took an Option on Its Rival →](/field-notes/2026-09-19-olympus-backed-neptune-then-optioned-its-rival) [A $400 Million Device Failed Its Trial. Its Engineers Bought It Back. →](/field-notes/2026-09-17-400m-device-failed-its-trial-engineers-bought-it-back) [FDA Approved the First Sirolimus Balloon. A Subgroup Decides Who Buys It. →](/field-notes/2026-09-22-first-sirolimus-balloon-approved-subgroup-picks-the-buyer) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Integra’s Flooded Plant Shows What Inventory, a Second Source and Insurance Can’t Cover | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-10-02-integra-flood-what-inventory-second-source-insurance-cover [← All Field Notes](/field-notes) October 2, 2026 Medical Devices On July 29, Integra LifeSciences told investors that flooding at its Cincinnati manufacturing site wouldn't have a material effect on its 2026 guidance. Today it cut the guidance. The flood took about $7 million out of third-quarter revenue. Integra expects it to take another $15 million to $20 million out of the fourth, and the site won't be back to full manufacturing until the second quarter of 2027. The revised revenue range, $1.634 billion to $1.654 billion, also reflects new assumptions for the rest of the business, so not every dollar of the cut is the flood. The Cincinnati site makes and supports several products in Integra's specialty surgical portfolio, and the company had a plan in place for it. It drew down available inventory, switched on alternate sources it had already established for certain products, and leaned on property and business interruption insurance that it expects to offset a substantial portion of the earnings hit. Inventory, a second source and insurance are the three defenses a company that makes its product in one place gets to choose from. Each one protects something different, and none of them protects the customer who needs the product this month. ## Inventory buys weeks Finished goods on a shelf keep orders shipping while the plant is down, for as many weeks as the stock lasts. Integra's flood damaged inventory along with the equipment and the building, because product stored at the plant floods with the plant. Count weeks of cover per product, and count only stock held somewhere the same storm can't reach. ## A second source buys specific products Integra's alternate sources apply only "for certain products." Qualifying a second source for a medical device takes months of work before anything goes wrong. You audit the supplier, validate the process on their line, and test their output against your specification. That effort goes to the products that carry the revenue, or the ones patients can't wait for, and the rest of the catalog carries the gap. ## Insurance buys back the earnings Business interruption coverage replaces lost profit while the plant is down. It doesn't put product on a hospital's shelf. A surgeon who spends most of a year using a competitor's device while yours is short may stay with it after your plant reopens, and the policy pays nothing for that. ## Measure the outage from the customer's side That surgeon's decision runs on a clock none of the three defenses controls directly: for each product, the weeks between the plant going down and the hospital's shelf going empty. Offsite inventory lengthens the clock. A qualified second source stops it for the products it covers. Insurance leaves it running. Today's Failure Modes Friday is a case of a number calculated too late. After the Francis Scott Key Bridge collapsed in Baltimore, the NTSB ran a vulnerability assessment and found the bridge almost 30 times over the acceptable risk threshold for a ship strike. The bridge's owner had never run that assessment against the ships using the channel. The weeks-to-empty-shelf number for a product made in one place can be worked out today, in a spreadsheet. For each product I run, I want three answers: how many weeks of stock sit outside the plant, how many validated lines can make it right now, and which customers would switch to a competitor before we could ship again. Dave's take Integra's July estimate said the flood wouldn't move the year, and two months of recovery work changed it. I want the product-by-product version of that estimate made before the water comes in, while there's still time to close the gaps it turns up. From Dave’s video library Today's Failure Modes Friday: the NTSB found the Key Bridge almost 30 times over the acceptable risk threshold for a ship strike, a number nobody calculated before the collapse. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [GM Is Building Battery Cells It Will Never Sell So It Can Judge the Ones It Buys →](/field-notes/2026-09-13-gm-900m-battery-cells-it-will-never-sell) [GM Shipped Patriot Interceptor Parts in 22 Days. Lockheed Is Buying the Rate. →](/field-notes/2026-09-23-gm-shipped-patriot-parts-in-22-days-lockheed-needs-the-rate) [Microbot's FDA-Cleared Robot Now Meets the Hospital's Value Analysis Committee Earlier →](/field-notes/2026-09-29-microbot-robot-cleared-hospital-committee-comes-first) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Agility Robotics Built Its Name on Legs. Its Digit 5 Launch Showed Wheels. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-10-03-agility-built-on-legs-now-sketching-wheels [← All Field Notes](/field-notes) October 3, 2026 Industrial Robotics Near the end of Agility Robotics' launch video for Digit 5, after minutes of a two-legged robot carrying totes, a humanoid on wheels appears. Agility has spent years arguing that legs suit buildings made for people. Jonathan Hurst, its CEO and chief robot officer, has spent decades studying how legs work. Asked about the wheels by The Robot Report, he confirmed Agility is exploring them and said it was too early to know whether they become a wheeled Digit or a different robot. ## What the filings say about who Digit is for Agility is going public through a merger with Churchill Capital Corp. XI at a $2.5 billion valuation, a deal expected to bring in more than $620 million. The filing reports $1.8 million of 2025 revenue against a $140 million operating loss, according to The Robot Report. It also cites more than $300 million in multi-year orders for Digit 5. GeekWire read the fine print on that number. All of it comes from one undisclosed customer, on a three-year contract for 1,000 robots, and it depends on Agility hitting milestones. Whatever that customer's buildings look like, they will shape Digit for the next three years. Digit's published job is moving totes and tending machines in warehouses and factories, and the new version lifts up to 50 pounds. The Robot Report lays out the trade between the two designs. Legs earn their cost on stairs, curbs and clutter, while wheels are steadier, mechanically simpler and use less energy anywhere the ground is flat. Hurst expects legs to become the right solution for more applications as their safety and reliability improve, and in the meantime he wants approaches that put robots in more places sooner. I keep landing in the same place on humanoids. They earn their keep where people and machines share a building nobody can redesign. Everywhere else, the robot that wins is built around the task. A tote run on level concrete sits close to that line, so a company this committed to legs has good reason to sketch wheels. ## The knee went first Agility has already given up one signature feature. Digit 5 dropped the backward-bending knees that people recognized earlier Digits by. The Robot Report read that change as a company willing to rethink even the fundamentals of its design. The wheeled rendering puts the legs themselves up for the same test. ## The cheapest moment to argue with your own design Look at the design decision your team is proudest of: the legs, the form factor on the first slide of the deck, the architecture somebody wrote a dissertation on. Is the customer paying for it, or is it how the team sees itself? Agility has now reopened two features, the knee in the product and the legs in the video, while annual revenue is under $2 million. The 1,000 robots on that contract are legged Digits, so that design is committed. The next product isn't yet, and it never gets cheaper to change than it is now. Every Digit that ships puts another set of leg actuators into a customer's building, and somebody has to stock the spare parts and train the technicians to fix them for as long as that robot runs. One customer holding $300 million in milestone-dependent orders will steer the roadmap whether the company intends it to or not. If that customer's floors look like the rest of the market, the steering helps. If they don't, the company ends up building for one buyer. I read the wheels as Agility keeping a way open to sell to the buyers whose floors never needed legs. Dave's take Hurst said he has guesses and inclinations and is only starting. I'd rather watch a humanoid company argue with its own silhouette in public at $1.8 million of revenue than find out after the thousandth robot ships that the customer was paying to get totes moved and never cared how. From Dave’s video library Dave on the five decisions that lock in once engineering starts, platform architecture among them, and why the conversation about who buys has to come first. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [A Humanoid Holds Twice the Rare Earth of an EV. Getting It Back Was Decided at the Design Review. →](/field-notes/2026-09-20-humanoid-rare-earth-decided-at-the-design-review) [ISO 21423 Isn't Published Yet. The Code That Speaks It Is Already Free. →](/field-notes/2026-09-18-iso-21423-not-published-the-code-is-already-free) [Robotaxis Proved the Driving. How to Count the Service Work Every Demo Leaves Out →](/field-notes/2026-09-27-robotaxis-proved-the-driving-count-the-service-loop) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # The FDA Approved a Heart Valve That Grows With Children. Proving It Saves Surgeries Takes a Decade. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-10-04-fda-valve-grows-with-children-proof-takes-a-decade [← All Field Notes](/field-notes) October 4, 2026 Medical Devices Until this week, a toddler born with a malformed pulmonary valve faced a fixed sequence. Surgeons sewed in a replacement valve sized for a small child, then reopened the chest when the child outgrew it, often more than once. On October 1 the FDA approved a valve built to break that sequence. Edwards Lifesciences' Autus valve goes in at about 13 millimeters. As the child grows, it can be stretched toward 22, close to adult size, with a balloon passed through a catheter. About 4,200 babies a year in the U.S. are born with a congenital pulmonary valve abnormality, according to the FDA. The agency's announcement leads with what the valve should mean for families, and Michelle Tarver, who runs the agency's device center, put it as far fewer trips to the operating room. The approval itself shows the valve works at six months. Fewer surgeries across a childhood is a claim six months of data can't test, and the FDA says longer-term studies will assess how well expansion holds up and whether it cuts surgeries. MedTech Dive reports a 10-year follow-up is under way. ## What six months of data show The pivotal trial enrolled 62 children at 12 U.S. sites, and the FDA reported six-month results on the first 60. Every implant procedure succeeded, with no deaths, blood clots or strokes. Three valves had frame fractures and two had reduced leaflet movement, and none of those five children had symptoms. The Autus is also the first FDA-approved heart valve with polymer leaflets in place of animal tissue. Two of those children needed the feature the valve exists for. Both had early valve deterioration, and both had their valves expanded without surgery. That's a strong early signal from a sample of two. ## Edwards bought before the long data Edwards bought Autus Valve Technologies in February for $128.9 million up front, plus up to $132.5 million in milestone payments, months before the approval came through. That came on top of an $11.5 million option payment Edwards had made earlier. The valve carried Breakthrough Device designation and became the first pediatric approval from the FDA's Total Product Life Cycle Advisory Program. Edwards paid for a working mechanism and an early signal, and the evidence on surgeries avoided will arrive on its watch. I expect hospitals and payers to keep asking for the ten-year data. Last month I wrote about [a heart valve Medicare started paying for sixteen months after its FDA approval](/field-notes/2026-09-16-fda-approved-it-medicare-paid-sixteen-months-later). ## Two evidence plans, written on day one The Autus program runs on two evidence plans. The first was a pivotal trial read at six months, which got the valve approved. The second is a ten-year follow-up of the same children, which will show whether it cuts surgeries. Hardware outside medicine meets the same split whenever the buyer pays for something that happens years later, such as a grid battery's service life. Certification covers the product as it ships, and the long claim has to be proven in the field. I want both plans written down at the start of a program: one that gets the product through the gate, and one that proves the long benefit on its own clock. The second plan has to settle early what the units in the field record, who pays for the follow-up, and who owns the data, because none of that is easy to add once the product ships. Dave's take The Autus team built expansion into the first device, so every child in the ten-year follow-up can add to the evidence for fewer surgeries. That's the design choice I would copy: a first version whose units in the field are already collecting proof of the benefit you sell, with a budget for the years that proof takes. From Dave’s video library Dave walks through what a fractional CPO costs next to a full-time hire, what the work looks like week to week, and why it takes six months to pay off. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [The FDA Approved This Valve for Symptom-Free Patients. Medicare Started Paying Sixteen Months Later. →](/field-notes/2026-09-16-fda-approved-it-medicare-paid-sixteen-months-later) [Your FDA Clearance Covers One Configuration. Your Roadmap Is the Queue. →](/field-notes/2026-09-21-your-fda-clearance-covers-one-configuration) [FDA Approved the First Sirolimus Balloon. A Subgroup Decides Who Buys It. →](/field-notes/2026-09-22-first-sirolimus-balloon-approved-subgroup-picks-the-buyer) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Amprius Rents Its Battery Factories. Every New Plant Has to Be Qualified Again. | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/2026-10-05-amprius-rents-battery-factories-qualifies-every-plant [← All Field Notes](/field-notes) October 5, 2026 Defense Hardware Amprius designs silicon-anode battery cells for drones and aircraft, and it has chosen not to own the factories that build them at volume. Its cells come off partner lines in China and South Korea. Today it added one in Indianapolis. CAMINUS, formerly EnPower, will build Amprius-designed pouch cells there. Qualification builds are running now, and volume production is planned for the first quarter of 2027. The arrangement keeps factory capital off Amprius's books and leaves it with a job it has to repeat at every partner: getting each new line qualified to build its cells. ## Why Amprius keeps adding plants Amprius keeps the anode, the cell design, an R&D lab and a pilot line in Fremont. It describes contract manufacturing as a way to add capacity quickly with little capital of its own. The government is now paying for more of that capacity. On September 28 the Department of War awarded Amprius a grant of up to $75 million through its Industrial Base Analysis and Sustainment program. The money converts a US electric-vehicle battery line, run by a South Korean manufacturer, into a line for drone cells. Amprius puts the conversion at about $100 million, including in-kind contributions from both companies, and expects the line to reach 12 million cells a year by early 2028. Both new sites serve the same buyer. Drone programs need cells that meet the domestic sourcing rules in the National Defense Authorization Act. Amprius calls the converted line a primary domestic source, and CAMINUS says its plant supports a supply chain that conforms to the act's Section 842. The same cell design now has to be qualified on two more lines within about a year. ## Qualification restarts at every plant CAMINUS says cell formats will move into production in stages, ordered by each format's volume and its qualification status. Customers approve cells made on a particular line, with a particular coating process and formation recipe. A new plant means new builds, new test data and a new sign-off for each format it takes on. Under the deal, CAMINUS will do anything from partial assembly of components Amprius ships in, up to full cell production. Amprius's grant announcement warns that cell performance shown in testing or in limited quantities may not hold at commercial volume, which is the question every new line has to answer during qualification. It also warns that Amprius relies on partners to meet the grant's milestones, so the answer comes from plants it doesn't operate. ## What to count before you sign Medical devices work the same way. Move a cleared device to a new contract manufacturer and the process has to be validated again before the first unit ships. When a founder brings me a plan built on contract manufacturing, I ask for three numbers: how many months a new line takes to qualify for the strictest customer, what the builds and testing cost, and how many product variants can move at once. I also ask who owns the qualification records. The process parameters, build records and test data from one plant are what let the next plant start from a file. If they sit only in the partner's quality system, the product can't move without starting over, and the business runs at the partner's discretion. Ownership of those records belongs in the contract, next to the qualification schedule. Dave's take The CAMINUS terms show the habit I look for in any founder renting capacity: formats staged by qualification status, with the staging built into the agreement. Amprius saved the factory capital, and the staged schedule in Indianapolis is where it pays the difference. From Dave’s video library Dave's Monday MedTech episode on the FDA approval of Edwards' Autus valve, the first heart valve built to expand as a child grows. I’m here to help you scale. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Related Field Notes [CATL’s Licensee Pays for the Factory. CATL Plans to Sell It the Cells. →](/field-notes/2026-09-24-catl-licensee-pays-for-the-factory-catl-sells-the-cells) [Integra’s Flooded Plant Shows What Inventory, a Second Source and Insurance Can’t Cover →](/field-notes/2026-10-02-integra-flood-what-inventory-second-source-insurance-cover) [The Navy Put a Startup’s Hypersonic Missile Into Production. Its Investors Paid for the Development First. →](/field-notes/2026-09-28-castelion-investors-paid-the-navys-development-bill) Also from Dave Saunders: [Founders Who Finish →](/field-notes/founders-who-finish) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Field Notes — April 2026 Archive | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/archive/2026-04 Archive # April _2026._ Field Notes ## Industry roundups [Apr 30, 2026 Adagio’s VT Pivotal Anchors a Week of Specialist Interventional Wins](/field-notes/2026-04-30-adagio-vclas-vt-pivotal-specialist-interventional-wins) [Apr 28, 2026 CMS and FDA Launch RAPID Pathway as Medicare Coverage Compresses to Months for Breakthrough Devices](/field-notes/2026-04-28-cms-fda-rapid-medicare-coverage-breakthrough-devices) [Apr 27, 2026 J&J Wins CE Mark for an Ottava-Compatible Stapler Before the Robot Ships](/field-notes/2026-04-27-jj-ottava-ethicon-stapler-ce-mark-platform-moat) [Apr 26, 2026 Boston Scientific Cuts 2026 Guidance as Watchman and EP Categories Restructure](/field-notes/2026-04-26-boston-scientific-2026-guidance-cut-watchman-ep-warning) [Apr 25, 2026 SS Innovations Showcases Next-Generation Surgical Robotics Platforms at SMRSC 2026](/field-notes/2026-04-25-next-gen-surgical-robotics-platforms-smrsc-2026) [Apr 24, 2026 Intuitive Surgical Raises 2026 Outlook as Robotic Surgery Expands Into General Surgery and Gynecology](/field-notes/2026-04-24-robotic-surgery-procedure-growth-q1-2026-market-expansion) [Apr 23, 2026 AI-Integrated Surgical Navigation Reaches Neurosurgery as Medtronic Clears Stealth AXiS for Cranial and ENT](/field-notes/2026-04-23-ai-surgical-navigation-neurosurgery-cranial-ent) [Apr 22, 2026 Robotic Surgery Expands Into Endovascular Practice as Microbot Launches Liberty at SIR](/field-notes/2026-04-22-endovascular-robotics-commercial-launch-sir-2026) [Apr 21, 2026 CMR Surgical’s US Launch and the Cost of Getting Market Entry Wrong](/field-notes/2026-04-21-cmr-surgical-us-launch-execution-discipline) [Apr 20, 2026 The Surgical Robotics Market Is No Longer a One-Platform Decision](/field-notes/2026-04-20-surgical-robotics-competition-intensifies-2026) [← Back to archive](/field-notes/archive) --- # Field Notes — May 2026 Archive | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/archive/2026-05 Archive # May _2026._ Field Notes ## Industry roundups [May 31, 2026 A Narrow Indication Is the Regulatory Strategy for AI Medical Devices](/field-notes/2026-05-31-narrow-indication-ai-medical-device-clearance) [May 30, 2026 The Climate-Hardware IPO Wave Is a Demand Story, Not a Technology Story](/field-notes/2026-05-30-climate-hardware-ipo-wave-demand-not-technology) [May 29, 2026 Surgical Robotics’ Real Edge in 2026 Is Knowing When Not to Launch](/field-notes/2026-05-29-surgical-robotics-when-to-launch-cmr-versius) [May 28, 2026 The Neurosurgical Supply Shortage Reveals the Hidden Risk in Your DHF](/field-notes/2026-05-28-neurosurgical-supply-shortage-supplier-qualification) [May 27, 2026 MDUFA VI Just Reset the De Novo and Pre-Sub Mechanics](/field-notes/2026-05-27-mdufa-vi-de-novo-presub-credits-founder-mechanics) [May 26, 2026 Physical AI’s Big Checks Are Landing Above the Robot](/field-notes/2026-05-26-physical-ai-foundation-models-platform-bet) [May 25, 2026 Deal Team Six Rewrites the Defense Hardware Founder’s Path to Market](/field-notes/2026-05-25-pentagon-deal-team-six-defense-hardware-founder-path) [May 24, 2026 Software Cadence Is Becoming the Surgical Robotics Moat](/field-notes/2026-05-24-surgical-robotics-software-cadence-platform-moat) [May 23, 2026 Medtronic’s $650M SPR Deal and the Chronic Pain Platform Consolidation](/field-notes/2026-05-23-medtronic-spr-chronic-pain-neuromodulation-platform) [May 22, 2026 Guardant360 Liquid CDx and the Companion Diagnostic Platform Moat](/field-notes/2026-05-22-guardant-liquid-cdx-companion-diagnostic-platform-moat) [May 21, 2026 The Surgical Robotics Challenger Field Fills Out in 2026](/field-notes/2026-05-21-surgical-robotics-challenger-field-2026-launch-architecture) [May 20, 2026 Boston Scientific’s SEISMIQ IVL Data Shows What Pivotal Evidence Now Buys You](/field-notes/2026-05-20-boston-scientific-seismiq-ivl-fracture-pivotal-evidence) [May 19, 2026 Boston Scientific’s $1.5B MiRus Stake Marks MedTech’s Milestone Era](/field-notes/2026-05-19-boston-scientific-mirus-tavr-milestone-deal-structure) [May 18, 2026 The FDA Just Reframed Regulatory Pathway Risk for Hard-Tech Founders](/field-notes/2026-05-18-fda-commissioner-resignation-regulatory-pathway-risk) [May 17, 2026 Physical AI Crosses From VC Bet to Incumbent Install Base](/field-notes/2026-05-17-physical-ai-fanuc-google-incumbent-install-base) [May 16, 2026 Rocket Lab and RTX Confirm the Subcontractor Playbook for Defense Hardware](/field-notes/2026-05-16-rocket-lab-rtx-defense-hardware-subcontractor-playbook) [May 15, 2026 Hard-Tech Mega-Rounds Land on the Physical Layer of the AI Stack](/field-notes/2026-05-15-hard-tech-mega-rounds-physical-layer-capital) [May 14, 2026 AI Imaging and Robotic Navigation Converge in the Interventional Stack](/field-notes/2026-05-14-ai-imaging-robotic-navigation-interventional-stack) [May 13, 2026 Medtronic Splits Cardiovascular Into Two Operating Stack Units as Interventional Cardiology Therapies Takes Shape](/field-notes/2026-05-13-medtronic-ict-cardiovascular-stack-restructure) [May 12, 2026 IVL Platform Competition Splits Into Three Mechanisms for Calcium Modification](/field-notes/2026-05-12-ivl-platform-competition-calcium-modification-2026) [May 11, 2026 Neptune Triton GI Robot Validates the Specialist Surgical Robotics Path](/field-notes/2026-05-11-neptune-triton-gi-robot-specialist-surgical-robotics-path) [May 10, 2026 SS Innovations Perth-to-India Telesurgery Reframes Surgical Robotics Platform Economics](/field-notes/2026-05-10-telesurgery-platform-economics-surgical-robotics) [May 9, 2026 Interventional Medtech Q1 Earnings Reveal a Margin Tax Beneath the Topline](/field-notes/2026-05-09-interventional-medtech-q1-margin-tariff-compression) [May 8, 2026 AI-Embodied Surgical Robots Force a Regulatory Framework Reset](/field-notes/2026-05-08-ai-embodied-surgical-robots-regulatory-framework-reset) [May 7, 2026 BD Q2 Interventional Confirms the Surgical Consumables Layer Underneath the Robotics Build-Out](/field-notes/2026-05-07-bd-q2-interventional-surgical-consumables-strength) [May 6, 2026 J&J Ottava FORTE Pivotal Resets the Soft Tissue Surgical Robotics Stack](/field-notes/2026-05-06-jj-ottava-forte-pivotal-soft-tissue-robotics-stack) [May 5, 2026 Olympus-EndoRobotics Deal Reframes Distribution as a Surgical Robotics Channel Strategy](/field-notes/2026-05-05-olympus-endorobotics-distribution-channel-strategy) [May 4, 2026 Surgical Robotics Moves to ASCs as the Site-of-Care Shift Resets the Platform Spec](/field-notes/2026-05-04-surgical-robotics-asc-site-shift-platform-spec) [May 3, 2026 Stealth AXiS Goes Global as Integrated Platforms Redraw Surgical Robotics Competition](/field-notes/2026-05-03-stealth-axis-ce-mark-platform-stack-surgical-robotics) [May 2, 2026 Stryker’s Q1 Cyber Recovery Reframes Resilience as a Medtech Strategic Asset](/field-notes/2026-05-02-stryker-cyber-recovery-medtech-resilience) [May 1, 2026 CMR’s Versius Plus Gynecology Filing Anchors a Week of Indication Expansion](/field-notes/2026-05-01-cmr-versius-plus-gynecology-510k-indication-expansion) [← Back to archive](/field-notes/archive) --- # Field Notes — June 2026 Archive | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/archive/2026-06 Archive # June _2026._ Field Notes ## Industry roundups [June 30, 2026 Medtronic’s Hugo Filing Is an Access Play, Not a Specs Fight](/field-notes/2026-06-30-medtronic-hugo-general-surgery-access-play) [June 29, 2026 Aidoc’s Radiology AI Got a Breakthrough Tag, Not a Clearance](/field-notes/2026-06-29-breakthrough-designation-is-not-clearance) [June 28, 2026 Why the FDA’s 2,200 New Hires Won’t Speed Up Your Review](/field-notes/2026-06-28-fda-rehiring-wont-speed-your-review) [June 27, 2026 Tissium’s $68M Raise and the Power of a Narrow De Novo](/field-notes/2026-06-27-tissium-narrow-de-novo-nerve-repair) [June 26, 2026 Energy Dome’s CO2 Battery Wins Where Lithium Can’t](/field-notes/2026-06-26-google-energy-dome-co2-battery-long-duration) [June 25, 2026 Why a 3-Hour-Faster Blood Culture Test Is a Real Product](/field-notes/2026-06-25-faster-blood-culture-is-a-real-product) [June 24, 2026 What FDA’s QMSR Transition to ISO 13485 Means for Founders](/field-notes/2026-06-24-qmsr-iso-13485-quality-system) [June 23, 2026 Karl Storz Is Retiring a Surgical Robot That Worked](/field-notes/2026-06-23-karl-storz-retires-senhance-surgical-robot) [June 22, 2026 Medtronic’s Pulse Oximeter Fix Began in the Cohort](/field-notes/2026-06-22-medtronic-pulse-oximeter-validation-cohort) [June 21, 2026 When Beating the Predicate Breaks Your 510(k)](/field-notes/2026-06-21-when-better-specs-break-510k-equivalence) [June 20, 2026 The FDA’s Former Device Chief Is Advising Startups. What That Actually Buys You](/field-notes/2026-06-20-fda-device-chief-advising-medtech-startups) [June 19, 2026 Sanctuary AI’s Physical AI Claim Is One You Can Actually Check](/field-notes/2026-06-19-sanctuary-ai-validated-number-not-demo) [June 18, 2026 How to Read the Pentagon’s Drone-Budget Surge as a Hardware Founder](/field-notes/2026-06-18-how-to-read-pentagon-drone-budget) [June 17, 2026 How a 2018 De Novo Let Dexcom Put a Glucose Monitor on a Toddler](/field-notes/2026-06-17-dexcom-otc-cgm-kids-de-novo-dividend) [June 16, 2026 Why OraSure Built a Collection Kit, Not a Better STI Test](/field-notes/2026-06-16-orasure-sti-collection-kit-access-bottleneck) [June 15, 2026 GT Medical Raised $100M on Evidence, Not the Clearance It Already Had](/field-notes/2026-06-15-gt-medical-gammatile-evidence-unlocks-capital) [June 14, 2026 The Robot Is the Giveaway. The Instruments Are the Business.](/field-notes/2026-06-14-surgical-robot-instruments-are-the-business) [June 13, 2026 NEURA’s $1.4 Billion Humanoid Bet. I’d Bet Narrow.](/field-notes/2026-06-13-neura-1-4b-physical-ai-humanoid-form-factor) [June 12, 2026 Penumbra’s Thunderbolt Stroke Clearance Is a Platform Play](/field-notes/2026-06-12-penumbra-thunderbolt-stroke-platform-extension) [June 11, 2026 Battlefield Drone Tech Came Home. The Dual-Use Line Is the Lesson.](/field-notes/2026-06-11-dual-use-counter-drone-comes-home) [June 10, 2026 GM’s Sodium-Ion Bet: Why It’s Selling Cells, Not Systems](/field-notes/2026-06-10-gm-sodium-ion-grid-storage-product-boundary) [June 9, 2026 When a Quality System Catches the Problem and Still Can’t Fix It](/field-notes/2026-06-09-quality-system-capa-effectiveness) [June 8, 2026 The 2026 Hardware Funding Records Hide a Higher Bar, Not Easier Money](/field-notes/2026-06-08-hardware-funding-records-higher-bar) [June 7, 2026 Medicare’s RAPID Coverage Pathway Moves the Work Earlier for Device Founders](/field-notes/2026-06-07-rapid-coverage-pathway-reimbursement-cliff) [June 6, 2026 FDA 510(k) Exemptions Are Expanding. They Can Erase Your Moat](/field-notes/2026-06-06-fda-510k-exemptions-predicate-moat) [June 5, 2026 Closed-Loop Insulin Delivery: The Algorithm Is the Easy Part](/field-notes/2026-06-05-closed-loop-insulin-delivery-validation) [June 4, 2026 The Humanoid Robot Is the Most Photogenic Bet in Physical AI, and the Hardest to Ship](/field-notes/2026-06-04-humanoid-robots-narrow-form-factor-edge-ai) [June 2, 2026 The Pentagon’s Drone Gauntlet Is Really a Manufacturing Test](/field-notes/2026-06-02-pentagon-drone-gauntlet-manufacturing-readiness-test) [June 1, 2026 Spectral AI’s De Novo Clearance Is a Lesson in Non-Dilutive Capital](/field-notes/2026-06-01-spectral-ai-de-novo-barda-non-dilutive-capital) [← Back to archive](/field-notes/archive) --- # Field Notes — July 2026 Archive | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/archive/2026-07 Archive # July _2026._ Field Notes ## Industry roundups [Jul 31, 2026 Momentis Built Its Own Predicate, Then Spent It Three Times](/field-notes/2026-07-31-momentis-built-its-own-predicate) [Jul 30, 2026 ENvue’s Feeding Tube Robot Came Second. That’s the Product Decision.](/field-notes/2026-07-30-envue-feeding-tube-robot-came-second) [Jul 29, 2026 How the Indication Sentence Decides Your 510(k)](/field-notes/2026-07-29-indication-sentence-decides-your-510k) [Jul 28, 2026 The Pentagon Rewrote the Drone Supply Chain. That’s a Product Requirement.](/field-notes/2026-07-28-pentagon-rewrote-the-drone-supply-chain) [Jul 27, 2026 Sartorius Got Its Tariff Money Back. It Handed the Refund to Customers.](/field-notes/2026-07-27-sartorius-gave-the-tariff-refund-to-customers) [Jul 26, 2026 Vicarious Surgical Raised $425 Million and Never Reached a Submission](/field-notes/2026-07-26-vicarious-surgical-never-reached-a-submission) [Jul 25, 2026 Cepheid’s New GeneXpert Kept the Old Menu. That’s the Product.](/field-notes/2026-07-25-cepheid-genexpert-kept-the-menu) [Jul 24, 2026 The Bearings, Not the Algorithm: Where Hardware Scale-Up Stalls](/field-notes/2026-07-24-bearings-not-the-algorithm-hardware-scale-up) [Jul 23, 2026 J&J’s Ottava Robot Cleared FDA. Beating da Vinci Is a Different Problem](/field-notes/2026-07-23-ottava-cleared-fda-beating-da-vinci-is-different) [Jul 22, 2026 The Pulse Oximeter Worked. The Validation Didn't.](/field-notes/2026-07-22-pulse-oximeter-validation-not-the-device) [Jul 21, 2026 En Carta Signed Its Buyer Before It Built the Line](/field-notes/2026-07-21-en-carta-signed-the-buyer-before-the-line) [Jul 20, 2026 Teleoperation Is a Trap Only If the Human Is There by Accident](/field-notes/2026-07-20-teleoperation-trap-human-by-accident) [Jul 19, 2026 A Disposable Surgical Robot Is a Manufacturing Bet](/field-notes/2026-07-19-disposable-surgical-robot-is-a-manufacturing-bet) [Jul 18, 2026 In Defense Hardware, the Constraint Is Production, Not Design](/field-notes/2026-07-18-defense-hardware-constraint-is-production-not-design) [Jul 17, 2026 For a New Implant, a Medicare Payment Code Is the Market Gate](/field-notes/2026-07-17-medicare-payment-code-is-the-market-gate) [Jul 16, 2026 The FDA’s Review Clock Runs on FDA Days, Not Calendar Days](/field-notes/2026-07-16-fda-days-are-not-calendar-days) [Jul 15, 2026 In Orthopedic Robots, the Moat Is Never the Robot](/field-notes/2026-07-15-orthopedic-robots-moat-is-not-the-robot) [Jul 14, 2026 Console-Free Cryocatheter: Cutting the Console Is the Product](/field-notes/2026-07-14-console-free-cryocatheter-is-the-product) [Jul 13, 2026 Europe’s Biggest Fusion Raise Buys Runway, Not Fusion](/field-notes/2026-07-13-fusion-mega-raise-is-runway-not-a-moat) [Jul 12, 2026 New Robot Safety Standards Are a Market-Access Gate, Not Paperwork](/field-notes/2026-07-12-robot-safety-standards-market-access-gate) [Jul 11, 2026 Non-Surgical Brain Interfaces Win on Access. That Is Not the Hard Part.](/field-notes/2026-07-11-non-surgical-brain-interfaces-access) [Jul 10, 2026 Eighty Percent of Factories Have No Robots. That Is a Product Problem.](/field-notes/2026-07-10-eighty-percent-of-factories-have-no-robots) [Jul 9, 2026 A Longer FDA Approval Clock Is a Runway Problem](/field-notes/2026-07-09-fda-approval-clock-is-a-runway-problem) [Jul 8, 2026 The First Patient-Facing Medical LLM Cleared FDA on a 510(k). The Scope Is Why.](/field-notes/2026-07-08-first-patient-facing-medical-llm-510k) [Jul 7, 2026 The First PTSD Neurostimulation Device Cleared FDA. Why Veterans Get It First](/field-notes/2026-07-07-ptsd-neurostim-de-novo-clearance-vs-market) [Jul 6, 2026 The Pentagon’s New Drone Integrator and Your DoD Buyer Map](/field-notes/2026-07-06-pentagon-drone-integrator-dod-buyer-map) [Jul 5, 2026 Merck’s $11.3B Bio-Techne Deal and the Premium on Being Indispensable](/field-notes/2026-07-05-merck-bio-techne-11b-tools-acquisition) [Jul 4, 2026 Launch on the Narrowest Indication You Can Defend](/field-notes/2026-07-04-launch-narrow-fda-indication-expansion) [Jul 3, 2026 Quantum Systems Raised $1.2 Billion. The Number to Study Is 19,000.](/field-notes/2026-07-03-quantum-systems-1-2b-drone-round) [Jul 2, 2026 A Cleared Device Is Not a Business: Zimmer Buys Pacira’s iovera](/field-notes/2026-07-02-cleared-device-is-not-a-business) [Jul 1, 2026 Physical AI’s Hard Problem Isn’t Vision. It’s Knowing Where the Robot Is.](/field-notes/2026-07-01-physical-ai-state-recovery-not-vision) [← Back to archive](/field-notes/archive) --- # Field Notes — August 2026 Archive | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/archive/2026-08 Archive # July _2026._ Field Notes ## Industry roundups [Aug 31, 2026 Oura Says It Isn't a Medical Device. The Ad Compared It to a Sleep Lab.](/field-notes/2026-08-31-oura-not-a-medical-device-ad-cited-the-sleep-lab) [Aug 30, 2026 The FDA Wants to Know How You'd Test a Device That Answers Differently Every Time](/field-notes/2026-08-30-fda-genai-devices-answer-differently-every-time) [Aug 29, 2026 Medtronic Pulled These Capsules Twice for the Same Failure. Only the Cause Changed.](/field-notes/2026-08-29-medtronic-bravo-recalled-twice-same-failure) [Aug 28, 2026 Bedrock’s Excavators Dig With Nobody in the Cab. Bedrock Doesn’t Build Excavators.](/field-notes/2026-08-28-bedrock-retrofit-excavators-nobody-in-the-cab) [Aug 27, 2026 Uber's Software Decided Who Kept Earning. The Missing Human Cost 825 Million Euros.](/field-notes/2026-08-27-uber-825-million-for-the-missing-human) [Aug 26, 2026 Boston Scientific's Plants Went Quiet. Its Filing Names Shipping, Not Devices.](/field-notes/2026-08-26-boston-scientific-filing-names-shipping-not-devices) [Aug 25, 2026 XPeng Raised $900 Million for a Humanoid. Its First Customer Is XPeng.](/field-notes/2026-08-25-xpeng-humanoid-first-customer-is-xpeng) [Aug 24, 2026 Every COVID-19 Test EUA Ends December 26. Authorized Was Never the Same as Cleared.](/field-notes/2026-08-24-covid-test-eua-ends-authorized-is-not-cleared) [Aug 23, 2026 The FDA's TEMPO Pilot Isn't Reviewing These Devices. The Companies Wrote Their Own Limits.](/field-notes/2026-08-23-fda-tempo-pilot-companies-wrote-their-own-limits) [Aug 22, 2026 ARPA-H Bet $175 Million on Autonomous Stroke Robots. The Hard Part Is the Room They Land In.](/field-notes/2026-08-22-arpa-h-stroke-robots-the-room-they-land-in) [Aug 21, 2026 The FDA Authorized a Blood Draw Robot. The Ratio in the Label Is the Product.](/field-notes/2026-08-21-blood-draw-robot-ratio-in-the-label) [Aug 20, 2026 GM Walked Away From This Battery Plant Twenty Months Before It Opened](/field-notes/2026-08-20-gm-walked-away-from-this-battery-plant) [Aug 19, 2026 The Navy Wants Sixteen Times the Tomahawks. The Seven Years Is What Buys Them.](/field-notes/2026-08-19-navy-tomahawk-seven-years-buys-the-rate) [Aug 18, 2026 J&J Cleared AI on Its Lung Robot. The Fourth Update in Eighteen Months Matters More.](/field-notes/2026-08-18-jj-monarch-fourth-update-in-eighteen-months) [Aug 17, 2026 Abbott Makes the Glucose Number. Google Charges for What to Do About It.](/field-notes/2026-08-17-abbott-lingo-google-owns-the-coaching-layer) [Aug 16, 2026 H.B. Fuller Already Made Surgical Glue. It's Paying £715 Million for the Evidence.](/field-notes/2026-08-16-hb-fuller-715-million-for-the-evidence) [Aug 15, 2026 The Medicare Coverage Decision Now Happens Before Your Pivotal Trial](/field-notes/2026-08-15-medicare-coverage-decision-before-pivotal-trial) [Aug 14, 2026 Cook Medical's Breach Never Touched the Device. It Went Through a Person.](/field-notes/2026-08-14-cook-medical-breach-went-through-a-person) [Aug 13, 2026 Joby Is Paying $500 Million to Move Its Defense Work Out of the Building](/field-notes/2026-08-13-joby-moved-defense-work-out-of-the-building) [Aug 12, 2026 Hadrian Raised $1.37 Billion for Factories. The Drawings Are Still Your Problem.](/field-notes/2026-08-12-hadrian-factories-drawings-are-your-problem) [Aug 11, 2026 iRhythm Paid $287.5 Million to Stop Waiting on Its Own FDA Clearance](/field-notes/2026-08-11-irhythm-vitalconnect-bought-out-of-the-clearance-clock) [Aug 10, 2026 Caristo Built a Heart Test for the Patients Who Get Sent Home. The Billing Codes Came First.](/field-notes/2026-08-10-caristo-heart-test-billing-codes-came-first) [Aug 9, 2026 HII Put $900 Million on the Table for Shipyard Robots. The Qualification Comes First.](/field-notes/2026-08-09-hii-shipyard-robots-qualification-comes-first) [Aug 8, 2026 An FDA Advisory Panel Reads Your File Cold. Grail Finds Out September 23.](/field-notes/2026-08-08-fda-advisory-panel-reads-your-file-cold) [Aug 8, 2026 An FDA Advisory Panel Reads Your File Cold. Grail Finds Out September 23.](/field-notes/2026-08-08-fda-advisory-panel-reads-your-file-cold) [Aug 7, 2026 Solventum Is Separating $1.4 Billion of Software. It Ran on a Different Clock.](/field-notes/2026-08-07-solventum-software-runs-on-a-different-clock) [Aug 6, 2026 BD's Needle Was Out of Tolerance for Twenty Months. Part of the Fix Is a Label.](/field-notes/2026-08-06-bd-needle-out-of-tolerance-fix-is-a-label) [Aug 5, 2026 The Pentagon Wants Four Times the THAADs. The Shopping List Says Muzzle Covers.](/field-notes/2026-08-05-pentagon-interceptor-surge-buys-muzzle-covers) [Aug 4, 2026 Reimagine Wants Your Line Worker to Teach the Robot. Walden Wants to Do It for You.](/field-notes/2026-08-04-who-teaches-the-robot-reimagine-vs-walden) [Aug 3, 2026 Walter Reed Cleared the Implant. Now the Hospital Is the Manufacturer.](/field-notes/2026-08-03-walter-reed-hospital-is-now-the-manufacturer) [Aug 2, 2026 MMI Didn't Build a Robot for Alzheimer's. Two Surgeons Decided It Was One Anyway](/field-notes/2026-08-02-mmi-symani-alzheimers-who-defines-your-market) [Aug 1, 2026 Peak Energy Picked the Weaker Chemistry. Then It Deleted the Cooling System.](/field-notes/2026-08-01-peak-energy-deleted-the-cooling-system) [← Back to archive](/field-notes/archive) --- # Field Notes — September 2026 Archive | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/archive/2026-09 Archive # July _2026._ Field Notes ## Industry roundups [Sep 30, 2026 The FDA Would Credential Generative AI Like a Clinician. You Still Answer for a Model You Don’t Own.](/field-notes/2026-09-30-fda-genai-devices-credentialed-model-you-dont-own) [Sep 29, 2026 Microbot’s FDA-Cleared Robot Now Meets the Hospital’s Value Analysis Committee Earlier](/field-notes/2026-09-29-microbot-robot-cleared-hospital-committee-comes-first) [Sep 28, 2026 The Navy Put a Startup’s Hypersonic Missile Into Production. Its Investors Paid for the Development First.](/field-notes/2026-09-28-castelion-investors-paid-the-navys-development-bill) [Sep 27, 2026 Robotaxis Proved the Driving. How to Count the Service Work Every Demo Leaves Out](/field-notes/2026-09-27-robotaxis-proved-the-driving-count-the-service-loop) [Sep 26, 2026 The FDA’s Draft Surgical Robot Guidance Lets the Hardest Procedure Cover the Rest](/field-notes/2026-09-26-fda-robot-guidance-hardest-procedure-covers-the-rest) [Sep 25, 2026 The FDA Panel Said Yes to Grail’s Cancer Test. It Said No to the Word Early.](/field-notes/2026-09-25-galleri-panel-said-yes-to-the-test-no-to-early) [Sep 24, 2026 CATL’s Licensee Pays for the Factory. CATL Plans to Sell It the Cells.](/field-notes/2026-09-24-catl-licensee-pays-for-the-factory-catl-sells-the-cells) [Sep 23, 2026 GM Shipped Patriot Interceptor Parts in 22 Days. Lockheed Is Buying the Rate.](/field-notes/2026-09-23-gm-shipped-patriot-parts-in-22-days-lockheed-needs-the-rate) [Sep 22, 2026 FDA Approved the First Sirolimus Balloon. A Subgroup Decides Who Buys It.](/field-notes/2026-09-22-first-sirolimus-balloon-approved-subgroup-picks-the-buyer) [Sep 21, 2026 Your FDA Clearance Covers One Configuration. Your Roadmap Is the Queue.](/field-notes/2026-09-21-your-fda-clearance-covers-one-configuration) [Sep 20, 2026 A Humanoid Holds Twice the Rare Earth of an EV. Getting It Back Was Decided at the Design Review.](/field-notes/2026-09-20-humanoid-rare-earth-decided-at-the-design-review) [Sep 19, 2026 Olympus Backed This Endoscopy Robot, Then Took an Option on Its Rival](/field-notes/2026-09-19-olympus-backed-neptune-then-optioned-its-rival) [Sep 18, 2026 ISO 21423 Isn’t Published Yet. The Code That Speaks It Is Already Free.](/field-notes/2026-09-18-iso-21423-not-published-the-code-is-already-free) [Sep 17, 2026 A $400 Million Device Failed Its Trial. Its Engineers Bought It Back.](/field-notes/2026-09-17-400m-device-failed-its-trial-engineers-bought-it-back) [Sep 16, 2026 The FDA Approved This Valve for Symptom-Free Patients. Medicare Started Paying Sixteen Months Later.](/field-notes/2026-09-16-fda-approved-it-medicare-paid-sixteen-months-later) [Sep 15, 2026 When No Law Fits Your Product, Regulators Reach for the Nearest One](/field-notes/2026-09-15-no-law-fits-regulators-reach-for-the-nearest-one) [Sep 14, 2026 J&J Bought Into a $37 Billion Orthopedics Market. It's Leaving a $50 Billion One.](/field-notes/2026-09-14-jnj-bought-37-billion-market-leaving-50-billion) [Sep 13, 2026 GM Is Building Battery Cells It Will Never Sell So It Can Judge the Ones It Buys](/field-notes/2026-09-13-gm-900m-battery-cells-it-will-never-sell) [Sep 12, 2026 Teradyne Is Suing Over Three Universal Robots Patents. All Three Expire in March.](/field-notes/2026-09-12-teradyne-jaka-patents-expire-in-march) [Sep 11, 2026 Axogen Paid $200 Million to Widen the Pool of Surgeons Who Can Repair a Nerve](/field-notes/2026-09-11-axogen-200m-nerve-repair-without-the-microsurgeon) [Sep 10, 2026 Nine Sites, $1.8 Million: How to Read a Hardware Company's Traction](/field-notes/2026-09-10-nine-sites-1-8-million-reading-hardware-traction) [Sep 8, 2026 The Army Bought Its First Production Laser. The Factory Was Funded in March.](/field-notes/2026-09-08-army-first-production-laser-factory-funded-in-march) [Sep 7, 2026 Cellares' Machine Met Every Release Spec. Bristol Myers Squibb Still Walked Away.](/field-notes/2026-09-07-cellares-met-every-release-spec-bms-still-walked) [Sep 6, 2026 Stryker's Surgical Headset Is FDA Authorized. Its Own Label Keeps the Monitors On.](/field-notes/2026-09-06-stryker-headset-label-keeps-the-monitors-on) [Sep 5, 2026 Medtronic Paid $700 Million for a Surgical Robot It Can't Sell in the United States](/field-notes/2026-09-05-medtronic-700m-robot-it-cant-sell-in-the-us) [Sep 4, 2026 FDA Reclassified This Cancer Test Category. Fewer Than Five People Commented.](/field-notes/2026-09-04-fda-reclassified-cancer-tests-five-comments) [Sep 3, 2026 Kardium's Cleanroom Builds 500 a Year. The Business Needs 30,000.](/field-notes/2026-09-03-kardium-cleanroom-500-a-year-needs-30000) [Sep 2, 2026 Two in Ten Storage Projects Now Treat the 30% Tax Credit as Optional](/field-notes/2026-09-02-storage-projects-treat-tax-credit-as-optional) [Sep 1, 2026 Boston Scientific Is Pulling Catheters Its Complaint Data Never Flagged](/field-notes/2026-09-01-boston-scientific-catheters-complaint-data-never-flagged) [← Back to archive](/field-notes/archive) --- # Field Notes — October 2026 Archive | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/archive/2026-10 Archive # July _2026._ Field Notes ## Industry roundups [Oct 5, 2026 Amprius Rents Its Battery Factories. Every New Plant Has to Be Qualified Again.](/field-notes/2026-10-05-amprius-rents-battery-factories-qualifies-every-plant) [Oct 4, 2026 The FDA Approved a Heart Valve That Grows With Children. Proving It Saves Surgeries Takes a Decade.](/field-notes/2026-10-04-fda-valve-grows-with-children-proof-takes-a-decade) [Oct 3, 2026 Agility Robotics Built Its Name on Legs. Its Digit 5 Launch Showed Wheels.](/field-notes/2026-10-03-agility-built-on-legs-now-sketching-wheels) [Oct 2, 2026 Integra’s Flooded Plant Shows What Inventory, a Second Source and Insurance Can’t Cover](/field-notes/2026-10-02-integra-flood-what-inventory-second-source-insurance-cover) [Oct 1, 2026 Terumo Is Buying Arsenal Medical Before Its Pivotal Trial Has Data. Medtronic Paid for the Proof.](/field-notes/2026-10-01-terumo-buys-arsenal-medical-before-pivotal-trial-data) [← Back to archive](/field-notes/archive) --- # Field Notes Archive | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/archive Archive # Every edition, _in order._ Browse Field Notes by month. Industry roundups, book pages, and newsletter pages are all indexed together within each month. [October 2026 5 entries](/field-notes/archive/2026-10) [Oct 5 Amprius Rents Its Battery Factories. Every New Plant Has to Be Qualified Again.](/field-notes/2026-10-05-amprius-rents-battery-factories-qualifies-every-plant) [Oct 4 The FDA Approved a Heart Valve That Grows With Children. Proving It Saves Surgeries Takes a Decade.](/field-notes/2026-10-04-fda-valve-grows-with-children-proof-takes-a-decade) [Oct 3 Agility Robotics Built Its Name on Legs. Its Digit 5 Launch Showed Wheels.](/field-notes/2026-10-03-agility-built-on-legs-now-sketching-wheels) [Oct 2 Integra’s Flooded Plant Shows What Inventory, a Second Source and Insurance Can’t Cover](/field-notes/2026-10-02-integra-flood-what-inventory-second-source-insurance-cover) [Oct 1 Terumo Is Buying Arsenal Medical Before Its Pivotal Trial Has Data. Medtronic Paid for the Proof.](/field-notes/2026-10-01-terumo-buys-arsenal-medical-before-pivotal-trial-data) [September 2026 29 entries](/field-notes/archive/2026-09) [Sep 30 The FDA Would Credential Generative AI Like a Clinician. You Still Answer for a Model You Don’t Own.](/field-notes/2026-09-30-fda-genai-devices-credentialed-model-you-dont-own) [Sep 29 Microbot’s FDA-Cleared Robot Now Meets the Hospital’s Value Analysis Committee Earlier](/field-notes/2026-09-29-microbot-robot-cleared-hospital-committee-comes-first) [Sep 28 The Navy Put a Startup’s Hypersonic Missile Into Production. Its Investors Paid for the Development First.](/field-notes/2026-09-28-castelion-investors-paid-the-navys-development-bill) [Sep 27 Robotaxis Proved the Driving. How to Count the Service Work Every Demo Leaves Out](/field-notes/2026-09-27-robotaxis-proved-the-driving-count-the-service-loop) [Sep 26 The FDA’s Draft Surgical Robot Guidance Lets the Hardest Procedure Cover the Rest](/field-notes/2026-09-26-fda-robot-guidance-hardest-procedure-covers-the-rest) [Sep 25 The FDA Panel Said Yes to Grail’s Cancer Test. It Said No to the Word Early.](/field-notes/2026-09-25-galleri-panel-said-yes-to-the-test-no-to-early) [Sep 24 CATL’s Licensee Pays for the Factory. CATL Plans to Sell It the Cells.](/field-notes/2026-09-24-catl-licensee-pays-for-the-factory-catl-sells-the-cells) [Sep 23 GM Shipped Patriot Interceptor Parts in 22 Days. Lockheed Is Buying the Rate.](/field-notes/2026-09-23-gm-shipped-patriot-parts-in-22-days-lockheed-needs-the-rate) [Sep 22 FDA Approved the First Sirolimus Balloon. A Subgroup Decides Who Buys It.](/field-notes/2026-09-22-first-sirolimus-balloon-approved-subgroup-picks-the-buyer) [Sep 21 Your FDA Clearance Covers One Configuration. Your Roadmap Is the Queue.](/field-notes/2026-09-21-your-fda-clearance-covers-one-configuration) [Sep 20 A Humanoid Holds Twice the Rare Earth of an EV. Getting It Back Was Decided at the Design Review.](/field-notes/2026-09-20-humanoid-rare-earth-decided-at-the-design-review) [Sep 19 Olympus Backed This Endoscopy Robot, Then Took an Option on Its Rival](/field-notes/2026-09-19-olympus-backed-neptune-then-optioned-its-rival) [Sep 18 ISO 21423 Isn’t Published Yet. The Code That Speaks It Is Already Free.](/field-notes/2026-09-18-iso-21423-not-published-the-code-is-already-free) [Sep 17 A $400 Million Device Failed Its Trial. Its Engineers Bought It Back.](/field-notes/2026-09-17-400m-device-failed-its-trial-engineers-bought-it-back) [Sep 16 The FDA Approved This Valve for Symptom-Free Patients. Medicare Started Paying Sixteen Months Later.](/field-notes/2026-09-16-fda-approved-it-medicare-paid-sixteen-months-later) [Sep 15 When No Law Fits Your Product, Regulators Reach for the Nearest One](/field-notes/2026-09-15-no-law-fits-regulators-reach-for-the-nearest-one) [Sep 14 J&J Bought Into a $37 Billion Orthopedics Market. It's Leaving a $50 Billion One.](/field-notes/2026-09-14-jnj-bought-37-billion-market-leaving-50-billion) [Sep 13 GM Is Building Battery Cells It Will Never Sell So It Can Judge the Ones It Buys](/field-notes/2026-09-13-gm-900m-battery-cells-it-will-never-sell) [Sep 12 Teradyne Is Suing Over Three Universal Robots Patents. All Three Expire in March.](/field-notes/2026-09-12-teradyne-jaka-patents-expire-in-march) [Sep 11 Axogen Paid $200 Million to Widen the Pool of Surgeons Who Can Repair a Nerve](/field-notes/2026-09-11-axogen-200m-nerve-repair-without-the-microsurgeon) [Sep 10 Nine Sites, $1.8 Million: How to Read a Hardware Company's Traction](/field-notes/2026-09-10-nine-sites-1-8-million-reading-hardware-traction) [Sep 8 The Army Bought Its First Production Laser. The Factory Was Funded in March.](/field-notes/2026-09-08-army-first-production-laser-factory-funded-in-march) [Sep 7 Cellares' Machine Met Every Release Spec. Bristol Myers Squibb Still Walked Away.](/field-notes/2026-09-07-cellares-met-every-release-spec-bms-still-walked) [Sep 6 Stryker's Surgical Headset Is FDA Authorized. Its Own Label Keeps the Monitors On.](/field-notes/2026-09-06-stryker-headset-label-keeps-the-monitors-on) [Sep 5 Medtronic Paid $700 Million for a Surgical Robot It Can't Sell in the United States](/field-notes/2026-09-05-medtronic-700m-robot-it-cant-sell-in-the-us) [Sep 4 FDA Reclassified This Cancer Test Category. Fewer Than Five People Commented.](/field-notes/2026-09-04-fda-reclassified-cancer-tests-five-comments) [Sep 3 Kardium's Cleanroom Builds 500 a Year. The Business Needs 30,000.](/field-notes/2026-09-03-kardium-cleanroom-500-a-year-needs-30000) [Sep 2 Two in Ten Storage Projects Now Treat the 30% Tax Credit as Optional](/field-notes/2026-09-02-storage-projects-treat-tax-credit-as-optional) [Sep 1 Boston Scientific Is Pulling Catheters Its Complaint Data Never Flagged](/field-notes/2026-09-01-boston-scientific-catheters-complaint-data-never-flagged) [August 2026 32 entries](/field-notes/archive/2026-08) [Aug 31 Oura Says It Isn't a Medical Device. The Ad Compared It to a Sleep Lab.](/field-notes/2026-08-31-oura-not-a-medical-device-ad-cited-the-sleep-lab) [Aug 30 The FDA Wants to Know How You'd Test a Device That Answers Differently Every Time](/field-notes/2026-08-30-fda-genai-devices-answer-differently-every-time) [Aug 29 Medtronic Pulled These Capsules Twice for the Same Failure. Only the Cause Changed.](/field-notes/2026-08-29-medtronic-bravo-recalled-twice-same-failure) [Aug 28 Bedrock’s Excavators Dig With Nobody in the Cab. Bedrock Doesn’t Build Excavators.](/field-notes/2026-08-28-bedrock-retrofit-excavators-nobody-in-the-cab) [Aug 27 Uber's Software Decided Who Kept Earning. The Missing Human Cost 825 Million Euros.](/field-notes/2026-08-27-uber-825-million-for-the-missing-human) [Aug 26 Boston Scientific's Plants Went Quiet. Its Filing Names Shipping, Not Devices.](/field-notes/2026-08-26-boston-scientific-filing-names-shipping-not-devices) [Aug 25 XPeng Raised $900 Million for a Humanoid. Its First Customer Is XPeng.](/field-notes/2026-08-25-xpeng-humanoid-first-customer-is-xpeng) [Aug 24 Every COVID-19 Test EUA Ends December 26. Authorized Was Never the Same as Cleared.](/field-notes/2026-08-24-covid-test-eua-ends-authorized-is-not-cleared) [Aug 23 The FDA's TEMPO Pilot Isn't Reviewing These Devices. The Companies Wrote Their Own Limits.](/field-notes/2026-08-23-fda-tempo-pilot-companies-wrote-their-own-limits) [Aug 22 ARPA-H Bet $175 Million on Autonomous Stroke Robots. The Hard Part Is the Room They Land In.](/field-notes/2026-08-22-arpa-h-stroke-robots-the-room-they-land-in) [Aug 21 The FDA Authorized a Blood Draw Robot. The Ratio in the Label Is the Product.](/field-notes/2026-08-21-blood-draw-robot-ratio-in-the-label) [Aug 20 GM Walked Away From This Battery Plant Twenty Months Before It Opened](/field-notes/2026-08-20-gm-walked-away-from-this-battery-plant) [Aug 19 The Navy Wants Sixteen Times the Tomahawks. The Seven Years Is What Buys Them.](/field-notes/2026-08-19-navy-tomahawk-seven-years-buys-the-rate) [Aug 18 J&J Cleared AI on Its Lung Robot. The Fourth Update in Eighteen Months Matters More.](/field-notes/2026-08-18-jj-monarch-fourth-update-in-eighteen-months) [Aug 17 Abbott Makes the Glucose Number. Google Charges for What to Do About It.](/field-notes/2026-08-17-abbott-lingo-google-owns-the-coaching-layer) [Aug 16 H.B. Fuller Already Made Surgical Glue. It's Paying £715 Million for the Evidence.](/field-notes/2026-08-16-hb-fuller-715-million-for-the-evidence) [Aug 15 The Medicare Coverage Decision Now Happens Before Your Pivotal Trial](/field-notes/2026-08-15-medicare-coverage-decision-before-pivotal-trial) [Aug 14 Cook Medical's Breach Never Touched the Device. It Went Through a Person.](/field-notes/2026-08-14-cook-medical-breach-went-through-a-person) [Aug 13 Joby Is Paying $500 Million to Move Its Defense Work Out of the Building](/field-notes/2026-08-13-joby-moved-defense-work-out-of-the-building) [Aug 12 Hadrian Raised $1.37 Billion for Factories. The Drawings Are Still Your Problem.](/field-notes/2026-08-12-hadrian-factories-drawings-are-your-problem) [Aug 11 iRhythm Paid $287.5 Million to Stop Waiting on Its Own FDA Clearance](/field-notes/2026-08-11-irhythm-vitalconnect-bought-out-of-the-clearance-clock) [Aug 10 Caristo Built a Heart Test for the Patients Who Get Sent Home. The Billing Codes Came First.](/field-notes/2026-08-10-caristo-heart-test-billing-codes-came-first) [Aug 9 HII Put $900 Million on the Table for Shipyard Robots. The Qualification Comes First.](/field-notes/2026-08-09-hii-shipyard-robots-qualification-comes-first) [Aug 8 An FDA Advisory Panel Reads Your File Cold. Grail Finds Out September 23.](/field-notes/2026-08-08-fda-advisory-panel-reads-your-file-cold) [Aug 8 An FDA Advisory Panel Reads Your File Cold. Grail Finds Out September 23.](/field-notes/2026-08-08-fda-advisory-panel-reads-your-file-cold) [Aug 7 Solventum Is Separating $1.4 Billion of Software. It Ran on a Different Clock.](/field-notes/2026-08-07-solventum-software-runs-on-a-different-clock) [Aug 6 BD's Needle Was Out of Tolerance for Twenty Months. Part of the Fix Is a Label.](/field-notes/2026-08-06-bd-needle-out-of-tolerance-fix-is-a-label) [Aug 5 The Pentagon Wants Four Times the THAADs. The Shopping List Says Muzzle Covers.](/field-notes/2026-08-05-pentagon-interceptor-surge-buys-muzzle-covers) [Aug 4 Reimagine Wants Your Line Worker to Teach the Robot. Walden Wants to Do It for You.](/field-notes/2026-08-04-who-teaches-the-robot-reimagine-vs-walden) [Aug 3 Walter Reed Cleared the Implant. Now the Hospital Is the Manufacturer.](/field-notes/2026-08-03-walter-reed-hospital-is-now-the-manufacturer) [Aug 2 MMI Didn't Build a Robot for Alzheimer's. Two Surgeons Decided It Was One Anyway](/field-notes/2026-08-02-mmi-symani-alzheimers-who-defines-your-market) [Aug 1 Peak Energy Picked the Weaker Chemistry. Then It Deleted the Cooling System.](/field-notes/2026-08-01-peak-energy-deleted-the-cooling-system) [July 2026 31 entries](/field-notes/archive/2026-07) [Jul 31 Momentis Built Its Own Predicate, Then Spent It Three Times](/field-notes/2026-07-31-momentis-built-its-own-predicate) [Jul 30 ENvue’s Feeding Tube Robot Came Second. That’s the Product Decision.](/field-notes/2026-07-30-envue-feeding-tube-robot-came-second) [Jul 29 How the Indication Sentence Decides Your 510(k)](/field-notes/2026-07-29-indication-sentence-decides-your-510k) [Jul 28 The Pentagon Rewrote the Drone Supply Chain. That’s a Product Requirement.](/field-notes/2026-07-28-pentagon-rewrote-the-drone-supply-chain) [Jul 27 Sartorius Got Its Tariff Money Back. It Handed the Refund to Customers.](/field-notes/2026-07-27-sartorius-gave-the-tariff-refund-to-customers) [Jul 26 Vicarious Surgical Raised $425 Million and Never Reached a Submission](/field-notes/2026-07-26-vicarious-surgical-never-reached-a-submission) [Jul 25 Cepheid’s New GeneXpert Kept the Old Menu. That’s the Product.](/field-notes/2026-07-25-cepheid-genexpert-kept-the-menu) [Jul 24 The Bearings, Not the Algorithm: Where Hardware Scale-Up Stalls](/field-notes/2026-07-24-bearings-not-the-algorithm-hardware-scale-up) [Jul 23 J&J’s Ottava Robot Cleared FDA. Beating da Vinci Is a Different Problem](/field-notes/2026-07-23-ottava-cleared-fda-beating-da-vinci-is-different) [Jul 22 The Pulse Oximeter Worked. The Validation Didn't.](/field-notes/2026-07-22-pulse-oximeter-validation-not-the-device) [Jul 21 En Carta Signed Its Buyer Before It Built the Line](/field-notes/2026-07-21-en-carta-signed-the-buyer-before-the-line) [Jul 20 Teleoperation Is a Trap Only If the Human Is There by Accident](/field-notes/2026-07-20-teleoperation-trap-human-by-accident) [Jul 19 A Disposable Surgical Robot Is a Manufacturing Bet](/field-notes/2026-07-19-disposable-surgical-robot-is-a-manufacturing-bet) [Jul 18 In Defense Hardware, the Constraint Is Production, Not Design](/field-notes/2026-07-18-defense-hardware-constraint-is-production-not-design) [Jul 17 For a New Implant, a Medicare Payment Code Is the Market Gate](/field-notes/2026-07-17-medicare-payment-code-is-the-market-gate) [Jul 16 The FDA’s Review Clock Runs on FDA Days, Not Calendar Days](/field-notes/2026-07-16-fda-days-are-not-calendar-days) [Jul 15 In Orthopedic Robots, the Moat Is Never the Robot](/field-notes/2026-07-15-orthopedic-robots-moat-is-not-the-robot) [Jul 14 Console-Free Cryocatheter: Cutting the Console Is the Product](/field-notes/2026-07-14-console-free-cryocatheter-is-the-product) [Jul 13 Europe’s Biggest Fusion Raise Buys Runway, Not Fusion](/field-notes/2026-07-13-fusion-mega-raise-is-runway-not-a-moat) [Jul 12 New Robot Safety Standards Are a Market-Access Gate, Not Paperwork](/field-notes/2026-07-12-robot-safety-standards-market-access-gate) [Jul 11 Non-Surgical Brain Interfaces Win on Access. That Is Not the Hard Part.](/field-notes/2026-07-11-non-surgical-brain-interfaces-access) [Jul 10 Eighty Percent of Factories Have No Robots. That Is a Product Problem.](/field-notes/2026-07-10-eighty-percent-of-factories-have-no-robots) [Jul 9 A Longer FDA Approval Clock Is a Runway Problem](/field-notes/2026-07-09-fda-approval-clock-is-a-runway-problem) [Jul 8 The First Patient-Facing Medical LLM Cleared FDA on a 510(k). The Scope Is Why.](/field-notes/2026-07-08-first-patient-facing-medical-llm-510k) [Jul 7 The First PTSD Neurostimulation Device Cleared FDA. Why Veterans Get It First](/field-notes/2026-07-07-ptsd-neurostim-de-novo-clearance-vs-market) [Jul 6 The Pentagon’s New Drone Integrator and Your DoD Buyer Map](/field-notes/2026-07-06-pentagon-drone-integrator-dod-buyer-map) [Jul 5 Merck’s $11.3B Bio-Techne Deal and the Premium on Being Indispensable](/field-notes/2026-07-05-merck-bio-techne-11b-tools-acquisition) [Jul 4 Launch on the Narrowest Indication You Can Defend](/field-notes/2026-07-04-launch-narrow-fda-indication-expansion) [Jul 3 Quantum Systems Raised $1.2 Billion. The Number to Study Is 19,000.](/field-notes/2026-07-03-quantum-systems-1-2b-drone-round) [Jul 2 A Cleared Device Is Not a Business: Zimmer Buys Pacira’s iovera](/field-notes/2026-07-02-cleared-device-is-not-a-business) [Jul 1 Physical AI’s Hard Problem Isn’t Vision. It’s Knowing Where the Robot Is.](/field-notes/2026-07-01-physical-ai-state-recovery-not-vision) [June 2026 29 entries](/field-notes/archive/2026-06) [June 30 Medtronic’s Hugo Filing Is an Access Play, Not a Specs Fight](/field-notes/2026-06-30-medtronic-hugo-general-surgery-access-play) [June 29 Aidoc’s Radiology AI Got a Breakthrough Tag, Not a Clearance](/field-notes/2026-06-29-breakthrough-designation-is-not-clearance) [June 28 Why the FDA’s 2,200 New Hires Won’t Speed Up Your Review](/field-notes/2026-06-28-fda-rehiring-wont-speed-your-review) [June 27 Tissium’s $68M Raise and the Power of a Narrow De Novo](/field-notes/2026-06-27-tissium-narrow-de-novo-nerve-repair) [June 26 Energy Dome’s CO2 Battery Wins Where Lithium Can’t](/field-notes/2026-06-26-google-energy-dome-co2-battery-long-duration) [June 25 Why a 3-Hour-Faster Blood Culture Test Is a Real Product](/field-notes/2026-06-25-faster-blood-culture-is-a-real-product) [June 24 What FDA’s QMSR Transition to ISO 13485 Means for Founders](/field-notes/2026-06-24-qmsr-iso-13485-quality-system) [June 23 Karl Storz Is Retiring a Surgical Robot That Worked](/field-notes/2026-06-23-karl-storz-retires-senhance-surgical-robot) [June 22 Medtronic’s Pulse Oximeter Fix Began in the Cohort](/field-notes/2026-06-22-medtronic-pulse-oximeter-validation-cohort) [June 21 When Beating the Predicate Breaks Your 510(k)](/field-notes/2026-06-21-when-better-specs-break-510k-equivalence) [June 20 The FDA’s Former Device Chief Is Advising Startups. What That Actually Buys You](/field-notes/2026-06-20-fda-device-chief-advising-medtech-startups) [June 19 Sanctuary AI’s Physical AI Claim Is One You Can Actually Check](/field-notes/2026-06-19-sanctuary-ai-validated-number-not-demo) [June 18 How to Read the Pentagon’s Drone-Budget Surge as a Hardware Founder](/field-notes/2026-06-18-how-to-read-pentagon-drone-budget) [June 17 How a 2018 De Novo Let Dexcom Put a Glucose Monitor on a Toddler](/field-notes/2026-06-17-dexcom-otc-cgm-kids-de-novo-dividend) [June 16 Why OraSure Built a Collection Kit, Not a Better STI Test](/field-notes/2026-06-16-orasure-sti-collection-kit-access-bottleneck) [June 15 GT Medical Raised $100M on Evidence, Not the Clearance It Already Had](/field-notes/2026-06-15-gt-medical-gammatile-evidence-unlocks-capital) [June 14 The Robot Is the Giveaway. The Instruments Are the Business.](/field-notes/2026-06-14-surgical-robot-instruments-are-the-business) [June 13 NEURA’s $1.4 Billion Humanoid Bet. I’d Bet Narrow.](/field-notes/2026-06-13-neura-1-4b-physical-ai-humanoid-form-factor) [June 12 Penumbra’s Thunderbolt Stroke Clearance Is a Platform Play](/field-notes/2026-06-12-penumbra-thunderbolt-stroke-platform-extension) [June 11 Battlefield Drone Tech Came Home. The Dual-Use Line Is the Lesson.](/field-notes/2026-06-11-dual-use-counter-drone-comes-home) [June 10 GM’s Sodium-Ion Bet: Why It’s Selling Cells, Not Systems](/field-notes/2026-06-10-gm-sodium-ion-grid-storage-product-boundary) [June 9 When a Quality System Catches the Problem and Still Can’t Fix It](/field-notes/2026-06-09-quality-system-capa-effectiveness) [June 8 The 2026 Hardware Funding Records Hide a Higher Bar, Not Easier Money](/field-notes/2026-06-08-hardware-funding-records-higher-bar) [June 7 Medicare’s RAPID Coverage Pathway Moves the Work Earlier for Device Founders](/field-notes/2026-06-07-rapid-coverage-pathway-reimbursement-cliff) [June 6 FDA 510(k) Exemptions Are Expanding. They Can Erase Your Moat](/field-notes/2026-06-06-fda-510k-exemptions-predicate-moat) [June 5 Closed-Loop Insulin Delivery: The Algorithm Is the Easy Part](/field-notes/2026-06-05-closed-loop-insulin-delivery-validation) [June 4 The Humanoid Robot Is the Most Photogenic Bet in Physical AI, and the Hardest to Ship](/field-notes/2026-06-04-humanoid-robots-narrow-form-factor-edge-ai) [June 2 The Pentagon’s Drone Gauntlet Is Really a Manufacturing Test](/field-notes/2026-06-02-pentagon-drone-gauntlet-manufacturing-readiness-test) [June 1 Spectral AI’s De Novo Clearance Is a Lesson in Non-Dilutive Capital](/field-notes/2026-06-01-spectral-ai-de-novo-barda-non-dilutive-capital) [May 2026 31 entries](/field-notes/archive/2026-05) [May 31 A Narrow Indication Is the Regulatory Strategy for AI Medical Devices](/field-notes/2026-05-31-narrow-indication-ai-medical-device-clearance) [May 30 The Climate-Hardware IPO Wave Is a Demand Story, Not a Technology Story](/field-notes/2026-05-30-climate-hardware-ipo-wave-demand-not-technology) [May 29 Surgical Robotics’ Real Edge in 2026 Is Knowing When Not to Launch](/field-notes/2026-05-29-surgical-robotics-when-to-launch-cmr-versius) [May 28 The Neurosurgical Supply Shortage Reveals the Hidden Risk in Your DHF](/field-notes/2026-05-28-neurosurgical-supply-shortage-supplier-qualification) [May 27 MDUFA VI Just Reset the De Novo and Pre-Sub Mechanics](/field-notes/2026-05-27-mdufa-vi-de-novo-presub-credits-founder-mechanics) [May 26 Physical AI’s Big Checks Are Landing Above the Robot](/field-notes/2026-05-26-physical-ai-foundation-models-platform-bet) [May 25 Deal Team Six Rewrites the Defense Hardware Founder’s Path to Market](/field-notes/2026-05-25-pentagon-deal-team-six-defense-hardware-founder-path) [May 24 Software Cadence Is Becoming the Surgical Robotics Moat](/field-notes/2026-05-24-surgical-robotics-software-cadence-platform-moat) [May 23 Medtronic’s $650M SPR Deal and the Chronic Pain Platform Consolidation](/field-notes/2026-05-23-medtronic-spr-chronic-pain-neuromodulation-platform) [May 22 Guardant360 Liquid CDx and the Companion Diagnostic Platform Moat](/field-notes/2026-05-22-guardant-liquid-cdx-companion-diagnostic-platform-moat) [May 21 The Surgical Robotics Challenger Field Fills Out in 2026](/field-notes/2026-05-21-surgical-robotics-challenger-field-2026-launch-architecture) [May 20 Boston Scientific’s SEISMIQ IVL Data Shows What Pivotal Evidence Now Buys You](/field-notes/2026-05-20-boston-scientific-seismiq-ivl-fracture-pivotal-evidence) [May 19 Boston Scientific’s $1.5B MiRus Stake Marks MedTech’s Milestone Era](/field-notes/2026-05-19-boston-scientific-mirus-tavr-milestone-deal-structure) [May 18 The FDA Just Reframed Regulatory Pathway Risk for Hard-Tech Founders](/field-notes/2026-05-18-fda-commissioner-resignation-regulatory-pathway-risk) [May 17 Physical AI Crosses From VC Bet to Incumbent Install Base](/field-notes/2026-05-17-physical-ai-fanuc-google-incumbent-install-base) [May 16 Rocket Lab and RTX Confirm the Subcontractor Playbook for Defense Hardware](/field-notes/2026-05-16-rocket-lab-rtx-defense-hardware-subcontractor-playbook) [May 15 Hard-Tech Mega-Rounds Land on the Physical Layer of the AI Stack](/field-notes/2026-05-15-hard-tech-mega-rounds-physical-layer-capital) [May 14 AI Imaging and Robotic Navigation Converge in the Interventional Stack](/field-notes/2026-05-14-ai-imaging-robotic-navigation-interventional-stack) [May 13 Medtronic Splits Cardiovascular Into Two Operating Stack Units as Interventional Cardiology Therapies Takes Shape](/field-notes/2026-05-13-medtronic-ict-cardiovascular-stack-restructure) [May 12 IVL Platform Competition Splits Into Three Mechanisms for Calcium Modification](/field-notes/2026-05-12-ivl-platform-competition-calcium-modification-2026) [May 11 Neptune Triton GI Robot Validates the Specialist Surgical Robotics Path](/field-notes/2026-05-11-neptune-triton-gi-robot-specialist-surgical-robotics-path) [May 10 SS Innovations Perth-to-India Telesurgery Reframes Surgical Robotics Platform Economics](/field-notes/2026-05-10-telesurgery-platform-economics-surgical-robotics) [May 9 Interventional Medtech Q1 Earnings Reveal a Margin Tax Beneath the Topline](/field-notes/2026-05-09-interventional-medtech-q1-margin-tariff-compression) [May 8 AI-Embodied Surgical Robots Force a Regulatory Framework Reset](/field-notes/2026-05-08-ai-embodied-surgical-robots-regulatory-framework-reset) [May 7 BD Q2 Interventional Confirms the Surgical Consumables Layer Underneath the Robotics Build-Out](/field-notes/2026-05-07-bd-q2-interventional-surgical-consumables-strength) [May 6 J&J Ottava FORTE Pivotal Resets the Soft Tissue Surgical Robotics Stack](/field-notes/2026-05-06-jj-ottava-forte-pivotal-soft-tissue-robotics-stack) [May 5 Olympus-EndoRobotics Deal Reframes Distribution as a Surgical Robotics Channel Strategy](/field-notes/2026-05-05-olympus-endorobotics-distribution-channel-strategy) [May 4 Surgical Robotics Moves to ASCs as the Site-of-Care Shift Resets the Platform Spec](/field-notes/2026-05-04-surgical-robotics-asc-site-shift-platform-spec) [May 3 Stealth AXiS Goes Global as Integrated Platforms Redraw Surgical Robotics Competition](/field-notes/2026-05-03-stealth-axis-ce-mark-platform-stack-surgical-robotics) [May 2 Stryker’s Q1 Cyber Recovery Reframes Resilience as a Medtech Strategic Asset](/field-notes/2026-05-02-stryker-cyber-recovery-medtech-resilience) [May 1 CMR’s Versius Plus Gynecology Filing Anchors a Week of Indication Expansion](/field-notes/2026-05-01-cmr-versius-plus-gynecology-510k-indication-expansion) [April 2026 10 entries](/field-notes/archive/2026-04) [Apr 30 Adagio’s VT Pivotal Anchors a Week of Specialist Interventional Wins](/field-notes/2026-04-30-adagio-vclas-vt-pivotal-specialist-interventional-wins) [Apr 28 CMS and FDA Launch RAPID Pathway as Medicare Coverage Compresses to Months for Breakthrough Devices](/field-notes/2026-04-28-cms-fda-rapid-medicare-coverage-breakthrough-devices) [Apr 27 J&J Wins CE Mark for an Ottava-Compatible Stapler Before the Robot Ships](/field-notes/2026-04-27-jj-ottava-ethicon-stapler-ce-mark-platform-moat) [Apr 26 Boston Scientific Cuts 2026 Guidance as Watchman and EP Categories Restructure](/field-notes/2026-04-26-boston-scientific-2026-guidance-cut-watchman-ep-warning) [Apr 25 SS Innovations Showcases Next-Generation Surgical Robotics Platforms at SMRSC 2026](/field-notes/2026-04-25-next-gen-surgical-robotics-platforms-smrsc-2026) [Apr 24 Intuitive Surgical Raises 2026 Outlook as Robotic Surgery Expands Into General Surgery and Gynecology](/field-notes/2026-04-24-robotic-surgery-procedure-growth-q1-2026-market-expansion) [Apr 23 AI-Integrated Surgical Navigation Reaches Neurosurgery as Medtronic Clears Stealth AXiS for Cranial and ENT](/field-notes/2026-04-23-ai-surgical-navigation-neurosurgery-cranial-ent) [Apr 22 Robotic Surgery Expands Into Endovascular Practice as Microbot Launches Liberty at SIR](/field-notes/2026-04-22-endovascular-robotics-commercial-launch-sir-2026) [Apr 21 CMR Surgical’s US Launch and the Cost of Getting Market Entry Wrong](/field-notes/2026-04-21-cmr-surgical-us-launch-execution-discipline) [Apr 20 The Surgical Robotics Market Is No Longer a One-Platform Decision](/field-notes/2026-04-20-surgical-robotics-competition-intensifies-2026) --- # Founders Who Finish — the field manual for founders done with startup theater | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/founders-who-finish [← All Field Notes](/field-notes) The Book Most founders don’t fail. They stall. They start with an idea they love, get buried under tasks and second-guessing, and the spark fades until the company quietly drifts. _Founders Who Finish_ is the book I wrote for the founder who is sick of that pattern and done performing the startup version of success. It is not a textbook and it is not hustle porn. It is a direct guide to making better decisions and building a company that does not need you awake at 2 a.m. to survive. ## Why I wrote it I spent a decade in Silicon Valley and three decades working in and around the Bay Area, and I know the game. I have built companies that launched, scaled, and thrived, and I have walked away from ones that were eating the part of me I most wanted to keep. The lesson that runs through all of it is simple. Finishing is rare. Almost every founder says they will finish. Almost none of them do, because somewhere along the way they confuse motion with progress and start optimizing for how the work looks instead of what it produces. The book exists to break that habit. I am not here to cheerlead. I am here to help you decide what you are actually building, who it helps, and what happens if you stop. That answer is your lighthouse, and you will need it when the fog rolls in, because it will. The founders who finish are the ones who keep that light in view while everyone around them is busy looking busy. ## How the book is built _Founders Who Finish_ runs in three parts. The first gets your head straight, the second builds the machine, and the third is about actually finishing the thing rather than letting it stall in sight of the line. Part One Get your head straight Finish what you started, kill the fantasy before it kills the company, nail the right problem, and build products that make money instead of products that win applause. The early chapters are about the founder, because the biggest thing holding most companies back is the person running them. Part Two Build the machine The three numbers that actually tell you whether the business survives, why founders are builders and not bottlenecks, how to fund the right way or not fund at all, how to hire like you mean it, and how to sell like a grown-up. This is the operating core: the systems that let a company run without heroics. Part Three Finish the damn thing Build the business to survive without you, finish strong, and understand that your exit starts now, not on the day you decide to sell. A company you cannot leave is not finished. It is a job you gave yourself with worse hours. ## The idea underneath all of it There is a difference between performing commitment and producing results, and once you see it you cannot unsee it. The networking coffee that goes nowhere, the deck you keep restyling instead of fixing, the standing meeting nobody has the nerve to kill: that is theater, and the only person it fools is the founder doing it. Real businesses reward outcomes. The test I keep coming back to is brutal and useful. If you stopped doing a thing tomorrow, would anything actually break? If the honest answer is no, it was theater. Pets.com raised three hundred million dollars, ran a Super Bowl ad, and went from IPO to liquidation in two hundred and sixty-eight days. Mailchimp was two people in Atlanta who never took venture money and sold to Intuit for twelve billion dollars in cash. In 1999 one of them looked far more impressive. Only one of them still exists. The founders who finish are not the ones who perform the best story. They are the ones who build the best business. That is the whole argument of the book, told through the scars I picked up learning it the hard way. Founders Who Finish Get the book The full field manual for founders who are done stalling and ready to finish what they start. Available now. [Get the Book](https://davesaunders.net/book) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) Want this judgment applied to your company? [Work with Dave →](/work-with-me) Read the latest [Field Notes →](/field-notes) Subscribe to [The Build newsletter →](/field-notes/the-build) --- # Field Notes — Industry intelligence for hard-tech founders | Base Reality Group Source: https://www.baserealitygroup.com/field-notes Industry intelligence # Field _Notes._ Industry intelligence for founders building medical devices, surgical robotics, diagnostics, defense, and other regulated hardware. What I’m reading, what it means, and what it changes for the companies I work with. [ October 5, 2026 Amprius Rents Its Battery Factories. Every New Plant Has to Be Qualified Again. Amprius designs silicon-anode drone cells and rents the factories that build them, now adding CAMINUS in Indianapolis with volume planned for early 2027. Renting moves the capital off the books, but each plant still has to be qualified format by format, and the company should own the records that make the next plant faster. Read more →](/field-notes/2026-10-05-amprius-rents-battery-factories-qualifies-every-plant)[ October 4, 2026 The FDA Approved a Heart Valve That Grows With Children. Proving It Saves Surgeries Takes a Decade. The FDA approved Edwards' Autus valve, which goes into toddlers at about 13 millimeters and can be balloon-expanded toward 22 as they grow. The six-month data shows the valve works, and the promise of fewer surgeries runs on a ten-year follow-up. Read more →](/field-notes/2026-10-04-fda-valve-grows-with-children-proof-takes-a-decade)[ October 3, 2026 Agility Robotics Built Its Name on Legs. Its Digit 5 Launch Showed Wheels. Agility Robotics built its name on legs, then showed a wheeled rendering near the end of its Digit 5 launch video. Its filings show $1.8 million of 2025 revenue and $300 million in orders from a single customer, which leaves the next form factor open while it is still cheap to change. Read more →](/field-notes/2026-10-03-agility-built-on-legs-now-sketching-wheels)[ October 2, 2026 Integra’s Flooded Plant Shows What Inventory, a Second Source and Insurance Can’t Cover Integra LifeSciences cut its 2026 guidance after a July flood at its Cincinnati plant, which won't return to full manufacturing until the second quarter of 2027. Inventory, alternate sources for certain products and insurance each protect something different, and none of them keeps product reaching the customer. A founder who makes a product in one place can count the weeks before customers have to switch. Read more →](/field-notes/2026-10-02-integra-flood-what-inventory-second-source-insurance-cover)[ October 1, 2026 Terumo Is Buying Arsenal Medical Before Its Pivotal Trial Has Data. Medtronic Paid for the Proof. Terumo Neuro agreed to buy Arsenal Medical while NeoCast, Arsenal's liquid embolic for chronic subdural hematoma, is still investigational and its pivotal trial is enrolling. The trial has to show NeoCast is no worse than Medtronic's Onyx, and Medtronic had already run the trial showing the procedure beats surgery alone. Arsenal tested its material in two uses, and the one that drew a buyer was the one another company had proved. Read more →](/field-notes/2026-10-01-terumo-buys-arsenal-medical-before-pivotal-trial-data)[ September 30, 2026 The FDA Would Credential Generative AI Like a Clinician. You Still Answer for a Model You Don’t Own. The FDA's device center floated a competency-based review for generative AI devices, modeled on how clinicians are credentialed, with more monitoring after launch. Its proposed Foundation Model Master File is voluntary, and the device maker still answers for a model it didn't build. Founders putting a chatbot or agent in front of clinicians can settle version control in the vendor contract first. Read more →](/field-notes/2026-09-30-fda-genai-devices-credentialed-model-you-dont-own)[September 29, 2026 Microbot’s FDA-Cleared Robot Now Meets the Hospital’s Value Analysis Committee Earlier Microbot Medical's LIBERTY robot is FDA-cleared and on full market release, and the company now brings hospital value analysis committees into each sale earlier, citing a four-to-six-month review. The committee weighs per-case cost, which a startup can start answering in its design review. Read more →](/field-notes/2026-09-29-microbot-robot-cleared-hospital-committee-comes-first) [Browse all Field Notes →](/field-notes/archive) Free courses, when they ship Field Notes is daily. The [Operator’s Guides](/guides) are the deep end: full free courses on standards like ISO 14971. Leave your email and each new guide lands in your inbox the day it ships. Nothing else. Email address Send me each guide Done. You’ll get the next guide when it ships. --- # The Build — a monthly print newsletter for founders | Base Reality Group Source: https://www.baserealitygroup.com/field-notes/the-build [← All Field Notes](/field-notes) The Build _The Build_ is a monthly newsletter for founders and entrepreneurs. It is about making money, building systems, and getting to where you want to go. It is not a medtech newsletter and it is not a daily inbox blast. It is a physical issue, printed and mailed to you once a month, written for anyone running a real company who is tired of advice that sounds good and changes nothing. ## Why it comes in the mail The format is the point. _The Build_ arrives as a printed issue you can hold, mark up, and keep on the desk instead of one more thing buried in an inbox you already cannot keep up with. Reading on paper is a different kind of attention than skimming a screen between notifications. An issue you can annotate and return to does work that a deleted email never gets the chance to do. Once a month, one thing worth sitting with, away from the feed. ## What is inside an issue Each issue leads with a story from the real work of building companies and then turns it into something you can use. It is direct, first person, and honest about the parts most business writing skips. A few examples from recent issues: Issue No. 1 The CEO Who Called Me an Embarrassment The difference between performing commitment and producing results. A CEO once called me an embarrassment for leaving to pick up my kids, while never once questioning my output. That is startup theater, and the issue lays out how to spot it in your own week, including the contrast between Pets.com and Mailchimp: one looked impressive in 1999, only one still exists. Issue No. 2 The Growth Numbers The numbers that actually tell you whether your company survives, and why most founders are watching the wrong dashboard. Early-stage founders obsess over vanity metrics that make good screenshots. Growth-stage founders track the basics and ignore unit economics. The issue is about building the habit of looking at the numbers that matter before the bank account makes the decision for you. Issues also carry a quotes page, a reader question, and a short opinion piece. The throughline is the same as everything else I write. Build the business, not the performance of one. Pricing $14.95 for the first 30-day trial, then $79 per month. The Build Start your trial Get the next issue in your mailbox. Founders and entrepreneurs who want signal instead of noise, one month at a time. [Start the Trial](https://davesaunders.net/free-trial) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He publishes _The Build_ and wrote _Founders Who Finish_. [More about Dave →](/about) Want this judgment applied to your company? [Work with Dave →](/work-with-me) Read the latest [Field Notes →](/field-notes) Read about [Founders Who Finish →](/field-notes/founders-who-finish) --- # The Fractional Chief Product Officer Engagement — Terms | Base Reality Group Source: https://www.baserealitygroup.com/fractional-cpo The Fractional Chief Product Officer Engagement # Senior product leadership, _inside your company._ Six months minimum. One to two days a week. I join your team as your fractional head of product and run the product function until the structure belongs to your team, not to me. This page is the whole deal: the format, the fee, the guarantee, and what happens after you apply. If we end up on a call, it will match what you read here. [Apply to Work With Dave](/contact?interest=fractional) The terms ## One page. Everything on it. Six-Month Fractional CPO Engagement Dave Saunders · Base Reality Group The role I join your team as **fractional Chief Product Officer** and run the product function: roadmap, priorities, product and market definition, risk discipline, and the commercial story that carries your next round. Your team does the work with me and learns the method as we go. When the engagement ends, the structure belongs to them. Cadence **One to two days a week**, about ten hours, inside your company. On your standups, in your design reviews, in front of your board when it helps. Fee **$10,000 to $25,000 per month.** Stage and scope set the number, and we fix it before we start. Commitment **Six months minimum.** Product structure isn't built in a sprint, and I don't sell engagements shorter than the work. Capacity I run **a small number of engagements at a time**, by design. The work needs full attention, and the advice only stays honest when I'm not overextended. What we cover Five topics, sequenced for where your product is: **product and market definition**, the **productization roadmap**, **quality culture and regulatory pathway**, **verification, validation, and design discipline**, and the **go-to-market and investor narrative**. The full breakdown is on [Work With Me](/work-with-me). The guarantee By the end of six months, your company has a clear, actionable plan to reach your next real gate. For medical devices, that means a defined path through FDA and to market. For other regulated or safety-critical hardware, the equivalent certification and market path. For everything else, first revenue and the next funding milestone. If you don't have that plan, I keep working with you at no additional cost until you do. One condition: your team shows up and does the work. Who it's for ## Funded founders whose product works, but whose path to revenue isn't proven yet. You're building a real product: software, SaaS, connected devices, medical devices, robotics, or any product where quality and reliability decide whether it sells. My deepest credential is medical devices and surgical robotics; most of what I've shipped is software. You have a working prototype, a funded company, and a team. Nobody owns the product full time, and you can feel it: the roadmap shifts with the loudest voice in the room, and the structure underneath is thinner than anyone wants to admit. Two boundaries worth naming. I'm not a regulatory consultant; the pathway is part of the work, not the whole of it. And I'm not a fundraising advisor; I don't raise money for companies or take fees on a raise. Not there yet? Start with the [$500 working session](/work-with-me#consultation), or with the book and newsletter at [davesaunders.net](https://davesaunders.net). I want your business to succeed either way. The process ## What happens after you apply. 01 **You apply.** Seven questions about your company, your product, and what success looks like in six months. It takes about ten minutes. 02 **We talk.** If the application looks like a fit, I set up a call. It costs you nothing, and I'm not selling on it. We work through where the product actually stands. 03 **You get a scoped proposal.** The fee, the cadence, and the first ninety days, in writing. If it's not a fit, I say so and point you somewhere useful instead. [Apply to Work With Dave](/contact?interest=fractional) [Compare both ways to work with me](/work-with-me) --- # Operator's Guides: Free Courses on the Standards Hard Tech Runs On | Base Reality Group Source: https://www.baserealitygroup.com/guides ![Papercraft workbench with paper calipers, a rolled blueprint, screws, and a gear](/brand-assets/opt/papercraft/operators-workbench.webp) Thirty years of getting hard technology out of the lab leaves you fluent in a stack of standards most people only ever skim. These guides teach that stack in full: real course depth, one worked device example per series, a written article beside every video. The first series covers ISO 14971 risk management. Quality systems, electrical safety, and carbon accounting are next in the queue. [The Operator's Guide to ISO 14971 Twelve modules. One worked device. A complete ISO 14971:2019 risk management file, built from scratch in front of you. 12 modules · 6.5 hours · free](/guides/iso-14971) In the queue ISO 13485 quality management systems. IEC 60601 electrical safety for medical devices. The GHG Protocol and what Scope 1, 2, and 3 actually ask of a hardware company. Each one gets the same treatment: a full course, taught free. Get the next guide ISO 14971 is the first Operator’s Guide. ISO 13485, IEC 60601, and the GHG Protocol are in the queue. Leave your email and each new guide lands in your inbox the day it ships. Nothing else, no drip sequence. Email address Send me each guide Done. You’ll get the next guide when it ships. The standard is free. So is this course. Building the product is the work. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) --- # ISO 14971 Explained: A Free 12-Module Medical Device Risk Management Course | Base Reality Group Source: https://www.baserealitygroup.com/guides/iso-14971 ![Papercraft diorama of the WearPump insulin pump surrounded by a risk matrix, shield, checklists, and calipers](/brand-assets/opt/papercraft/series-iso-14971.webp) ISO 14971:2019 is the risk management standard every medical device company answers to. Most training either reads the clauses at you or hides the useful part behind a consulting engagement. This course does neither. Each module works the standard on one running device, WearPump, a wearable insulin pump, so every concept lands on something real: its hazards, its risk matrix, its control decisions, its file. I teach it because I have had to live it. I spent eight years taking a surgical robotics platform from prototype through FDA De Novo authorization, and the risk file was load-bearing the whole way. Watch the modules in order, or jump to the clause you are fighting with today. The [full playlist is on YouTube](https://www.youtube.com/playlist?list=PLbwRXUCI5bdI); every module page here pairs the video with a written version. [01 What Is ISO 14971? Why Medical Device Risk Management Exists One standard. One device. Twelve modules. This is the foundation every medical device risk management program starts from. 28:12](/guides/iso-14971/what-is-iso-14971) [02 Hazard, Hazardous Situation, and Harm: ISO 14971 Clause 3 Definitions General safety training teaches a two-step model. ISO 14971 uses three terms, and the one in the middle decides whether your risk file survives its first audit. 45:00](/guides/iso-14971/hazard-hazardous-situation-harm) [03 ISO 14971 Clause 4: The Risk Management Plan and System Most risk management starts at the hazard table. The standard says it starts well before that, with a plan, a governance structure, and an acceptability line you draw before you have measured anything. 42:03](/guides/iso-14971/risk-management-plan-clause-4) [04 Intended Use and Device Characteristics: ISO 14971 Risk Analysis Part 1 Before you can identify a single hazard, you have to answer three questions precisely. Almost every course on medical device risk skips straight past them to the analysis technique. 32:45](/guides/iso-14971/intended-use-and-device-characteristics) [05 Hazard Identification with FMEA and FTA: ISO 14971 Risk Analysis Part 2 Fault tree analysis finds a category of hazard that failure modes and effects analysis is structurally built to miss. Here's how to run both on the same device, and where they converge. 35:08](/guides/iso-14971/hazard-identification-fmea-fta) [06 Risk Estimation with Severity and Probability: ISO 14971 Risk Analysis Part 3 RPN and ISO 14971 risk look like the same number. They are not. Here's why the difference matters, and how to score every hazard on a device honestly. 31:47](/guides/iso-14971/risk-estimation-severity-probability) [07 Risk Evaluation and the Acceptability Line: ISO 14971 Clause 6 Risk evaluation is where a color on the matrix becomes a verdict. Here's how to draw the acceptability line, demonstrate ALARP, and be honest about the gray zone most companies fudge. 32:07](/guides/iso-14971/risk-evaluation-acceptability) [08 The Risk Control Hierarchy: ISO 14971 Clause 7 ISO 14971 gives a device its own three-tier risk-control hierarchy, in a mandatory order. Here's how to work it on a real device, why you usually can't lead with a warning label, and the two steps almost every risk file skips. 32:55](/guides/iso-14971/risk-control-hierarchy) [09 Residual Risk and Benefit-Risk Analysis in ISO 14971 ISO 14971 rarely lets you drive a risk to zero. Here's how you re-score what's left, build a benefit-risk case an auditor will accept, and tell the limit of your engineering apart from a risk actually worth shipping. 28:31](/guides/iso-14971/residual-risk-benefit-risk) [10 Overall Residual Risk Evaluation: ISO 14971 Clause 8 Clause 8 is the step most risk files skip: judging the whole device, not each risk. Here's how residuals cluster on one harm, why you weigh the device against the therapy it replaces, and the four-part statement a reviewer can actually check. 25:16](/guides/iso-14971/overall-residual-risk) [11 The Risk Management Report: ISO 14971 Clause 9 The file is the work; the report is the signed judgment about it, and a regulator wants the judgment. Here's what Clause 9 requires, the seven elements of the report, the one acceptance sentence that decides your submission, and who has to sign it. 30:27](/guides/iso-14971/risk-management-report) [12 Post-Market Surveillance: ISO 14971 Clause 10 Your device launches, and the real data starts arriving. Here's what post-market surveillance actually requires, the five sources worth watching, the trigger that turns watching into acting, and how twelve complaints in one month become a dated change in the risk file. 29:01](/guides/iso-14971/post-market-surveillance) Get the next guide ISO 14971 is the first Operator’s Guide. ISO 13485, IEC 60601, and the GHG Protocol are in the queue. Leave your email and each new guide lands in your inbox the day it ships. Nothing else, no drip sequence. Email address Send me each guide Done. You’ll get the next guide when it ships. The standard is free. So is this course. Building the product is the work. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) --- # Hazard, Hazardous Situation, and Harm: ISO 14971 Clause 3 Definitions | Base Reality Group Source: https://www.baserealitygroup.com/guides/iso-14971/hazard-hazardous-situation-harm [← All ISO 14971 modules](/guides/iso-14971) Module 02 45:00 video + article Watch this module Module 2 of the video course. The article below covers the same ground in written form, so you can watch, read, or both. \# The Word That Sits Between Hazard and Harm: ISO 14971 Clause 3, Explained In the spring of 1986, a radiation therapy machine in Tyler, Texas delivered a dose of radiation that was somewhere between four hundred and several thousand times what the patient was prescribed. The machine, a Therac-25, had been cleared for clinical use. Trained staff were operating it. Everything about the setup looked correct. The treatment began, and then it didn't. The investigation took multiple accidents before anyone understood what they were looking at. The cause was a software race condition: two processes in the control system could interfere with each other if an operator entered commands inside a very specific timing window. When that happened, the machine fired its high-energy electron beam directly, without the scanning magnet that would have spread it safely across the treatment field, while showing a minor error code on the screen. Here is the part that matters for everything below. That race condition had been in the code for some time before anyone was hurt. Devices were deployed at multiple facilities. Staff operated them. Nothing happened, because the flaw on its own was not enough to cause harm. It needed something else. It needed an experienced operator, someone who had cleared error messages so many times that the sequence was muscle memory, fast enough to land inside the timing window. The very thing that made that person good at the job was the thing that activated the hazard. A hazard. A sequence of events. A hazardous situation. And then harm. That chain is the foundation of ISO 14971 Clause 3, and it is the vocabulary that separates a risk file that survives an audit from one that does not. ## The two-step trap Most engineers who have been through occupational health and safety training learned a model with two terms. Hazard: anything that can cause harm. Risk: how likely that harm is, combined with how bad it would be. Two steps, P times S. For slips, chemical exposure, and machine guarding, that model works. It has been refined over decades and it is adequate for what it was built to do. Medical devices break it. When your device can deliver the wrong drug concentration, generate a fault current, or fail during a sterile procedure, the two-step model does not give you enough resolution to document what you actually know. You know more than "hazard exists, therefore P times S." You know _how_ the hazard becomes a harm. You have a path. The two-step model has nowhere to put the path. I'm Dave Saunders. I've spent more than thirty years in medical devices and regulatory affairs, a good share of it in audit rooms watching risk files get evaluated. The gap between how general safety vocabulary works and what ISO 14971 actually requires is real, it shows up in real risk files, and it is correctable once you see it clearly. ISO 14971 closes it by adding a third term between hazard and harm: the hazardous situation. The rest of this piece is about what that term does, and why the standard needed it. ## Clause 3: thirty-one terms, one architecture Clause 3 is the terms-and-definitions clause. It defines 31 terms precisely enough that the rest of the standard can use them without ambiguity. They cluster into four groups. The chain: hazard, hazardous situation, harm, risk, probability, severity. The process: risk management, analysis, estimation, evaluation, control, monitoring. The context: intended use, reasonably foreseeable misuse, use error, life cycle. And the foundations: safety, state of the art, benefit, objective evidence, verification. You don't need to memorize all 31. You need to understand the eight or nine that drive everything downstream, and you need to know where to look for the rest. The chain is where we go deep, because those terms are the structure your risk analysis is built on. Misunderstand them and your probability values measure the wrong thing. ## Hazard: a property, not an event ISO 14971 defines a hazard, in Clause 3.4, as "a potential source of harm." Five words, and the word _potential_ is doing real work. A hazard hasn't caused harm yet. It is a property of the device, or of the energy, material, or process associated with it, that has the capacity to cause harm under the right circumstances. The high-pressure spring in a delivery mechanism is a hazard. The current through a circuit board is a hazard. The sharp edge of a scalpel is a hazard. None of them has hurt anyone. That framing matters when you write the file: when you record "hazard: high-pressure reservoir," you are making a claim about the device's physical state, not yet about who might be hurt or how likely it is. The hazard column is where you inventory the potential energy in the device. ISO/TR 24971, the technical report that supports the standard, gives you hazard categories as a starting checklist: energy hazards, biological, chemical, and use-related. The categories are not a closed taxonomy. They are a prompt to make sure you haven't missed a whole class. And the standard asks you to analyze both normal and fault conditions, because hazardous situations can arise even when the device works exactly as designed. For our running example, WearPump, a wearable subcutaneous insulin pump for Type 1 diabetes, four hazards anchor the analysis: the pressurized insulin reservoir, holding up to 3 mL under tension; the concentrated insulin itself; the electrical energy in the battery, motor, and Bluetooth subsystem; and the skin-adhesive and infusion interface. Four hazards. Not four risks. Not four harms. Four potential sources of harm that we now trace through the chain. ## Hazardous situation: where probability lives Clause 3.5 defines a hazardous situation as "a circumstance in which people, property, or the environment is exposed to one or more hazards." The load-bearing word is _circumstance_. A specific set of conditions, a specific hazard, and a specific person or thing exposed to it. The hazard is a property of the device. The hazardous situation is what happens when a person comes into contact with that hazard through a specific sequence of events. This is the structural answer to the question the whole module opens with. Probability does not attach cleanly to a hazard. A pressurized reservoir is either under tension or it isn't. Probability attaches to a _sequence of events_: the chain of circumstances that brings the hazard and a person together in a way that could cause harm. The hazardous situation is where that chain ends, and that is where your P value belongs. Go back to the Therac-25. The race condition is the hazard. The hazardous situation is not "race condition exists in the code." It is the specific circumstance: an experienced operator clears the malfunction message during the timing window, the machine fires the high-energy beam without the scanning magnet, and the patient is in the treatment position receiving it. That is the exposure. That is the probability you care about, not whether the code has a flaw, but whether that sequence occurs during a treatment session. A single hazard can lead to several hazardous situations through different sequences, and each deserves its own entry, because each has a different probability, a different population, and a different severity. When an auditor asks how you arrived at a P value, the sequence of events is the answer. If the path isn't documented, the number is an assertion, not an estimate. ## Harm, and the spectrum of severity Clause 3.3 defines harm as "injury or damage to the health of people, or damage to property or the environment." Three things are worth noting. It is not limited to physical injury; the standard includes psychological harm. The 2019 revision made damage to property and the environment explicit. And harm reaches people, plural, not only the primary patient: caregivers, clinicians, and bystanders can all be exposed through the same hazardous situation. Severity, Clause 3.27, is a measure of the possible consequences, and it is not a binary. For WearPump, an unintended insulin delivery ranges from a mild glucose dip the patient corrects with a snack, all the way to severe hypoglycemia with unconsciousness, seizure, and a potentially fatal outcome. You document the full range, tied to the specific hazardous situation. Severity divorced from the situation is just a guess. There is a time dimension, too: some harms hit in seconds, others surface over 24 to 48 hours, and that latency changes whether anyone can intervene before the harm becomes serious. ## The chain, and what Clause 5.4 actually requires Put the four together and you have one continuous narrative. Here is what the device carries (hazard), here is how that thing reaches a person (sequence of events), here is the moment harm becomes possible (hazardous situation), and here is what happens (harm). Four concepts, four columns in your risk analysis table. This is not just a teaching frame. Clause 5.4 of the standard requires you to "identify the sequences of events that could result in a hazardous situation." Not identify hazards and estimate risk. Identify the sequences of events. The chain is the clause's requirement expressed as a data model, and your risk analysis is the populated instance of it. When an auditor reviews your file, they run structural checks: does every hazardous situation trace to a named hazard, does every harm trace to a named hazardous situation, is the sequence of events written out specifically enough that someone else could evaluate the probability. Those are not judgment calls. A "no" is a finding. In Module 5 we'll populate the chain with two complementary tools. FMEA starts at the device's failure modes and traces forward toward hazardous situations. Fault tree analysis starts at an undesired harm and traces backward, through logical gates, to the combinations of events that could produce it. Forward from failures, backward from harms; together they cover the chain from both ends. ## Risk: the probability of _what?_ Clause 3.18 defines risk as "the combination of the probability of occurrence of harm and the severity of that harm." P times S. Severity is the tractable half, a defined scale we'll build for WearPump in Module 3. Probability is the subtle half, and the real question is not "what is the probability" but "probability of _what?_" If you answer "probability that this component fails," you have attached your P to the wrong place in the chain. The standard says probability of occurrence of _harm_, which means the probability of the whole chain running to completion. For WearPump's reservoir hazard, the wrong P is the probability that an occlusion occurs. The right P is the probability that an occlusion occurs _and_ pressure builds to failure _and_ a drug bolus is delivered _and_ the patient experiences clinically significant hypoglycemia. An occlusion is relatively common; an occlusion that runs all the way to harm is much less so. One useful way to hold this is two components: P1, the probability that the hazardous situation occurs, and P2, the probability that, once it does, harm results. Your file's P is the product of both. Miss P1, which is common, and you are only measuring the danger of the exposure, not the likelihood of getting into it. ISO/TR 24971 has an annex on probability estimation that lists the evidence you can draw on: published failure-rate data for comparable components, field data from similar devices, human factors results from usability testing, fault tree or Markov analysis, and structured expert judgment. For a new device without field history, your estimates will be semi-quantitative, and the standard accepts that, as long as you document the rationale, tie it to the specific hazardous situation, and update it as real-world data arrives. ## The rest of the vocabulary Safety, Clause 3.26, is "freedom from unacceptable risk." Not zero risk, not the absence of hazards. You cannot remove insulin concentration from an insulin pump; safety means the risks that flow from that hazard, through their documented situations and paths, fall within your acceptability criteria. Risk management itself is four sequential tasks, analysis, evaluation, control, and monitoring, and the course follows that arc across the remaining modules. Two terms people use interchangeably are precise here: risk analysis produces the estimates, and risk assessment means analysis plus evaluation. There is always residual risk; the question is whether it is acceptable and outweighed by clinical benefit. The context terms set the boundaries. Reasonably foreseeable misuse is use that results from readily predictable human behavior, like wearing WearPump past its 72-hour limit while traveling. A use error is an honest mistake with no intent to deviate, such as programming a dose one decimal place too high. Intended use sets the outer edge of the whole hazard analysis; when it is vague, the boundary is vague, and your hazard identification has no principled place to stop. ## WearPump's first four risk entries Here is the deliverable, the format every future entry follows. > **Entry 1. Pressurized reservoir.** Hazard: mechanical energy stored under tension. Sequence: an occlusion develops, the pump keeps delivering, pressure builds, the cannula fitting fails. Hazardous situation: a sudden unintended subcutaneous bolus. Harm: hypoglycemia, from a mild dip to loss of consciousness. **Entry 2. Concentrated insulin.** Hazard: pharmacological energy. Sequence: air enters the tubing during a cartridge change, the set isn't fully primed, an air gap precedes the next dose. Hazardous situation: the patient receives no insulin while believing delivery is normal. Harm: hyperglycemia or diabetic ketoacidosis, developing over hours. **Entry 3. Electrical energy.** Hazard: battery and motor drive current. Sequence: a component fault creates a current path to the housing, a person contacts the surface while the device is active. Hazardous situation: an unintended electrical shock. Harm: a localized burn, arrhythmia, or ventricular fibrillation. **Entry 4. Skin interface.** Hazard: adhesive and subcutaneous cannula. Sequence: extended wear in a high-moisture environment degrades the adhesive while the cannula irritates the site. Hazardous situation: prolonged irritation at the skin boundary. Harm: contact dermatitis, skin breakdown, or infection over 24 to 48 hours. Four hazards, each with a documented sequence, a named hazardous situation, and a harm with a severity range. In Module 5 we'll run a systematic identification and expand this to a dozen or more, but every entry will follow the same chain, because the chain is what Clause 5.4 requires. ## The short version A hazard is a property of your device. A hazardous situation is the circumstance in which a person is exposed to it, reached through a documented sequence of events. Harm is the consequence, at some severity. Risk is the probability of that harm times its severity, and the probability belongs to the sequence, not to the hazard. Get that one distinction right and the rest of the risk file becomes logical instead of procedural. Get it wrong, and an auditor will find the broken P column, every time. If this was useful, Module 3 builds the risk management plan, where the acceptability criteria and the severity scale get defined. More on product strategy, roadmaps, and fractional CPO work is at [baserealitygroup.com](https://baserealitygroup.com), and if you want the written version of this work in your inbox, the newsletter is _The Build_, at [davesaunders.net](https://davesaunders.net). [← Module 1What Is ISO 14971? Why Medical Device Risk Management Exists](/guides/iso-14971/what-is-iso-14971) [Module 3 →ISO 14971 Clause 4: The Risk Management Plan and System](/guides/iso-14971/risk-management-plan-clause-4) Get the next guide ISO 14971 is the first Operator’s Guide. ISO 13485, IEC 60601, and the GHG Protocol are in the queue. Leave your email and each new guide lands in your inbox the day it ships. Nothing else, no drip sequence. Email address Send me each guide Done. You’ll get the next guide when it ships. The standard is free. So is this course. Building the product is the work. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) --- # Hazard Identification with FMEA and FTA: ISO 14971 Risk Analysis Part 2 | Base Reality Group Source: https://www.baserealitygroup.com/guides/iso-14971/hazard-identification-fmea-fta [← All ISO 14971 modules](/guides/iso-14971) Module 05 35:08 video + article Watch this module Module 5 of the video course. The article below covers the same ground in written form, so you can watch, read, or both. \# FMEA Alone Will Not Satisfy ISO 14971, and Most Teams Don't Know Why Back in Module 2, I told you about the Therac-25, a radiation therapy machine that delivered doses hundreds of times what was prescribed. The cause was a software race condition that only fired when a specific operator typed a specific sequence fast enough to land inside a specific timing window. You know the story. What I never showed you is the diagram: the actual formal picture an investigator would draw to explain why it happened. Here's what that diagram makes visible that the story alone doesn't. The harm didn't come from one thing failing. It came from two things being true at the same moment: an experienced operator typing fast enough to fall inside the timing window, and a race condition sitting quietly in code that had shipped and run without incident for months. Neither of those, on its own, is a broken part. The operator typing fast isn't a failure. The code existing isn't a failure. It was the combination, arriving together, that mattered. Take away either one and nothing happens. I don't want you to file this away as one strange story about one old machine. Harm from a combination of ordinary conditions, rather than a single broken part, is why aviation investigators build fault trees for a plane that flew exactly as designed into terrain, and why nuclear engineers build them for a reactor where every gauge read normal right up until it wasn't. It's a recognized category in every high-consequence field that takes hazard identification seriously. Medical devices are behind on this. Fault tree analysis has been standard in aviation and nuclear safety for decades; medical device hazard identification, in practice, still leans overwhelmingly on one tool. ## The thesis No failure modes and effects analysis, run by anyone, no matter how careful, would have found the Therac-25 combination. Not because the engineers doing it would have been sloppy, but because of how the technique itself works. FMEA is structurally the wrong shape of tool for a hazard that comes from two things that both work exactly as intended, arriving together. If your hazard identification stops at FMEA, there is a whole category of risk in your device you have never actually looked for. ISO 14971 Clause 5.4 requires you to identify hazards and the sequences of events that could lead to a hazardous situation. It doesn't name a specific technique, but "systematic" implies more than one. Most engineers only really know one: failure modes and effects analysis, the one taught in every quality certification and Six Sigma course. Search for FMEA tutorials and you'll find some of the most-watched risk content online, genuinely well made. Search for fault tree analysis applied to a medical device, and you'll find almost nothing. I'm Dave Saunders. I've spent more than 30 years commercializing technology, almost 20 of them in medical devices and regulatory affairs, including several surgical robots, a good part of it in audit rooms. The finding I've seen most often isn't a bad probability estimate. It's a hazard analysis that only ever ran one technique, and missed an entire class of risk because nobody went looking for it a second way. This is Module 5 of a twelve-module masterclass building a complete ISO 14971 risk management file from scratch, using one running device: WearPump, a wearable subcutaneous insulin pump for Type 1 diabetes. Module 4 gave WearPump its three inputs: an intended use statement, a foreseeable-misuse list, and a safety-characteristics table. This module turns those three inputs into hazards, using two techniques that run in opposite directions. ## Two directions FMEA runs forward, from the part: what could fail? Fault tree analysis runs backward, from the harm: what combination of conditions causes it? Both belong in a systematic hazard identification. Neither replaces the other. FMEA has a simple spine: failure mode (how could this part fail), local effect (what happens right where it happens), and downstream effect (what happens all the way to the patient, which is where a hazardous situation lives). It comes in two flavors: design FMEA, for the device itself, and process FMEA, for how it gets made. Each failure mode gets scored on severity, occurrence, and detection, and every score needs a documented reason. A shrug is a finding waiting to happen. Run it as a team exercise, not solo: pull in engineering, clinical, and manufacturing perspectives, and walk the safety-characteristics table row by row. ## Running FMEA on WearPump Failure mode Local → downstream effect Hazardous situation Harm Motor driver stuck "on" Motor keeps turning past a stop command, plunger keeps advancing Continuous, uncommanded insulin delivery Severe, fast-onset hypoglycemia Dosing algorithm counter corrupts Algorithm under-counts, re-delivers an already-given dose Unintended double bolus Hypoglycemia Reservoir seal degrades Insulin seeps past the seal, not delivered through the cannula Silent underdelivery Hyperglycemia (delayed) Bluetooth radio firmware locks up Pump stops sending status, app shows a frozen reading Delayed recognition of a problem Depends on concurrent conditions That's a real, useful table, and it's fair to say what FMEA is good at: forward, part-by-part, it will find most of a device's single-point-failure risk. But look at the shape of the question it asks: how could this one thing fail, while everything else is assumed to be working normally? That's not a flaw in how carefully someone runs the analysis. It's built into the structure of the technique, and it means FMEA can't reach two categories of hazard. The first: hazards that arise when nothing fails at all, when every part does exactly what it was designed to do (we named these normal-condition hazards back in Module 2). The second is the one that got the Therac-25 patient: hazards that require two or more things to be true at the same time, where none of them, alone, is a failure. FMEA looks at one failure mode against a normal backdrop. It was never built to ask what happens when two normal things collide. ## Fault tree analysis, working backward Fault tree analysis starts from a specific harm, the top event, and works backward through logic gates to the basic events that could cause it. An OR gate means any one branch alone is enough; an AND gate means every branch underneath it has to be true at the same time. The AND gate is the one that reaches what FMEA cannot, because it can trace a harm back to a combination of ordinary conditions, none of them a failure. Picking which harms get a tree is a judgment call: start from the worst outcomes in your severity table, and build a tree for each one worth the effort. A basic event you trace back to is a minimal cut set, the smallest combination sufficient to trigger the top event on its own. Here's the Therac-25's actual tree. Top event: patient receives a massive radiation overdose. Underneath it, a single AND gate with two basic events: a race condition present in the code, dormant on most days, AND an operator entering the sequence fast enough to hit the timing window. Neither branch alone reaches the top event. Only their combination does, and that's exactly what makes it invisible to a technique that only ever asks "what could fail?" ## Running FTA on WearPump WearPump's tree starts from the same top event as its worst FMEA finding: severe, undetected hypoglycemia from an unintended bolus. Underneath sits an OR gate with three branches. Branch one is the motor driver failure FMEA already found. Branch two is the dosing algorithm double-count FMEA also found. Branch three is where fault tree analysis earns its keep: an AND gate requiring an occlusion in the tubing (foreseeable, not a failure), the alarm set to silent or vibrate (a normal setting), and the patient asleep or otherwise unable to notice it. Three ordinary conditions form a minimal cut set, and none of them is a broken part. The arithmetic explains why a combination like this is rare even though every condition inside it is common: for an AND gate you multiply the branch probabilities, you don't add them. An occlusion on roughly one wear cycle in 20, an alarm left silent on roughly one cycle in 10, a patient asleep for roughly a third of a cycle, multiplied together, lands around one in 600, not one in 20. Walk back through how FMEA works and ask which of those three would ever show up as a failure mode: none of them. An occlusion isn't a pump failure. A silent alarm is a feature working as designed. A sleeping patient isn't a failure at all. Only an AND gate, traced backward from the harm, was ever going to surface that branch. The same backward tracing turns up a second AND-gate entry (extended wear plus water exposure degrading the adhesive seal) and a third, OR-gate entry on the connectivity path (a DIY loop command, or a mid-command Bluetooth drop). ## Where both techniques converge Assembled, WearPump's hazard traceability matrix now has 11 entries: the four original ones from Module 2, four more FMEA found today (motor driver, dosing algorithm, reservoir seal, Bluetooth radio), and three more fault tree analysis found working backward (the silenced-alarm combination, the moisture combination, the connectivity path). Finish the pass, running FMEA across every remaining part and fault tree analysis backward from every harm category in the characteristics table, and the real count lands at a dozen or more, exactly what Module 3 promised, not because the device is unusually dangerous, but because that's what a device has once you stop looking with only one tool. Both the FMEA table and every fault tree belong in the risk management file as records, not scratch work you discard once the matrix is populated. When an auditor asks how you arrived at a hazardous situation, you hand them the diagram; you don't reconstruct your thinking from memory. ## The practical rule Run FMEA thoroughly, part by part, function by function. It's the right tool for single-point failures and will find most of a device's risk. Then run fault tree analysis backward from every serious harm you can name, specifically hunting for the combinations FMEA's one-failure-at-a-time structure cannot reach. Neither replaces the other. Running only one leaves a specific, predictable, nameable gap in your hazard analysis, whether or not anyone ever notices it. Every entry in that expanded matrix now has a hazard, a sequence of events, a hazardous situation, and a harm. What it doesn't have yet is a real number. Module 6 takes this table and assigns actual probability and severity to every row, against the five-by-five matrix built in Module 3, and finds out which hazardous situations demand a control. If this was useful, Module 6 is next. More on product strategy, roadmaps, and fractional CPO work at [baserealitygroup.com](https://baserealitygroup.com), and I write more broadly for founders too, in a newsletter called The Build, at [davesaunders.net](https://davesaunders.net). [← Module 4Intended Use and Device Characteristics: ISO 14971 Risk Analysis Part 1](/guides/iso-14971/intended-use-and-device-characteristics) [Module 6 →Risk Estimation with Severity and Probability: ISO 14971 Risk Analysis Part 3](/guides/iso-14971/risk-estimation-severity-probability) Get the next guide ISO 14971 is the first Operator’s Guide. ISO 13485, IEC 60601, and the GHG Protocol are in the queue. Leave your email and each new guide lands in your inbox the day it ships. Nothing else, no drip sequence. Email address Send me each guide Done. You’ll get the next guide when it ships. The standard is free. So is this course. Building the product is the work. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) --- # Intended Use and Device Characteristics: ISO 14971 Risk Analysis Part 1 | Base Reality Group Source: https://www.baserealitygroup.com/guides/iso-14971/intended-use-and-device-characteristics [← All ISO 14971 modules](/guides/iso-14971) Module 04 32:45 video + article Watch this module Module 4 of the video course. The article below covers the same ground in written form, so you can watch, read, or both. \# Your Intended Use Statement Is a Risk Control: ISO 14971 Risk Analysis, Part 1 In December 2014, a woman named Dana Lewis went to sleep wearing an insulin pump that was quietly making its own decisions about how much insulin to give her. Not because the manufacturer built it that way. She and a collaborator had reverse-engineered the pump, written their own software, and wired it to a glucose sensor, so that for the first time the system could adjust her insulin automatically through the night. She wasn't waiting for industry to build it. That became the rallying cry: We Are Not Waiting. Lewis and Scott Leibrand open-sourced the project in early 2015 and called it OpenAPS, the open artificial pancreas system. People with Type 1 diabetes around the world started building their own. By mid-2016 there were around 100 people running these home-built systems. By 2020, more than 1,700. Today it's thousands, all trusting their overnight insulin to software no company sold them and no regulator approved. To understand why they did it, you have to understand what they were escaping. Managing Type 1 diabetes by hand means being the algorithm: checking glucose, dosing, correcting, all day and, worse, all night, waking at 2 and 4 a.m. to keep blood sugar from drifting somewhere dangerous. What these people built themselves was a full night's sleep. But every one of those systems ran on commercial pumps built for a patient in the loop, pressing the buttons. Handing dosing to an outside algorithm was completely outside the intended use the manufacturers had written down. The patients weren't breaking the pumps. They were using them in a way nobody who made them ever put on the label. The FDA had to reckon with it. Its 2019 warning called the systems unauthorized and unreviewed, capable of the wrong insulin dose and real harm. The warning was reasonable, and it changed almost nothing. The systems worked well for the people using them, and commercial closed-loop pumps were still years behind. I'm not telling this story to take a side on DIY looping. I'm telling it because it's the clearest example I know of a simple, uncomfortable truth: the real-world use of your device will not stay inside the intended use statement you wrote. People will use it in ways that are predictable, even if they aren't approved. If your risk analysis only considered the device used exactly as intended, your analysis has a hole in it the size of the looping community. ## The three questions everyone skips Before you can identify a single hazard, you have to answer three questions, precisely. What is this device actually for, and for whom? How will people predictably use it that you didn't intend? And what is it about the physical nature of the device, the energy in it, the materials, the way it touches a person, that could ever cause harm? Get those three wrong, and every hazard you find afterward is aimed at the wrong device. That's ISO 14971, Clause 5. Most people skip all three. They're told to start the risk analysis, they open a failure-modes spreadsheet, and they start listing things that could go wrong. It feels productive, because the rows are filling up. But a failure mode of what, exactly? Used how, by whom, where? If you can't answer that, you're not analyzing your device. You're analyzing a vague idea of it, and the gaps in that idea become the hazards nobody catches. Almost every course on medical device risk teaches exactly the opposite order. FMEA tutorials and risk-matrix explainers are everywhere; all of them assume the device and its use are already nailed down. They almost never are. The scoping step, the part that decides what you're even analyzing, gets treated as something you scribble in a header box and move past. I'm Dave Saunders. I've spent more than 30 years commercializing technology, almost 20 of them in medical devices and regulatory affairs, including several surgical robots, a good part of it in audit rooms. The most expensive findings I've seen, the ones that send a company back to redo months of work, rarely come from a math error. They come from a hazard never identified, because the device's use and characteristics were never written down precisely enough to surface it. ## Intended use is a risk control, not a description Here's the claim people find hardest to believe: your intended use statement, that dry paragraph at the front of your documentation, is a risk control. Not a minor one. It's the cheapest and most powerful risk control you'll ever have, and you get to write it before you've spent a dollar on engineering. Intended use, or intended purpose in European regulation, is the objective intent of the manufacturer about what the device is for, as shown in your specifications, instructions, labeling, and even your marketing. Your advertising is part of your intended use. Imply the device does something in a sales page, and you've just expanded your intended use, and with it, everything you're now responsible for analyzing. Here's how a sentence becomes a safety device. The intended use statement draws a boundary around the hazard space you're responsible for. Everything inside it, every use you intend, you must analyze and control. Make that boundary narrow and specific, and you have a tractable problem. Make it broad and vague, and you've told an auditor, and a court, that you're responsible for analyzing everything inside that enormous boundary. This is where people get it exactly backwards. They write the statement broad on purpose, to keep options open, to not limit the market. What that actually does is enroll them in analyzing, controlling, and defending every use their words now cover. Two versions of the same claim make the difference concrete. Version one: "for the management of glucose-related conditions." Version two: "for the subcutaneous delivery of rapid-acting insulin for Type 1 diabetes, in patients six years and older." The first sounds bigger, and it's a trap: it puts Type 2 diabetes, gestational diabetes, and hospital glucose control all inside the fence. The second is smaller, sharper, and it's freedom, because everything it leaves out is a hazard analysis you don't have to do and a claim you don't have to defend. I've sat in a review where a company's intended use statement said their device was for use "in the home and in clinical environments." Sounds harmless. But "clinical environments" meant they now had to consider operating rooms, with their electrosurgery and specific electrical safety requirements, and intensive care, with its tangle of other devices. They didn't mean to claim any of that. One vague phrase added months of analysis they never intended to take on. One more consequence catches people off guard: intended use helps determine your device's regulatory classification, and with it, how hard your road to market is. Claim a use that puts you in a higher-risk class, and you've signed up for more clinical evidence, more scrutiny, more time. The statement isn't only a risk control. It's a business decision. ## What goes into a real statement A complete intended use statement answers a specific set of questions: the medical indication (what condition it diagnoses, treats, or monitors), the patient population (who it's for, including age and condition limits), the intended user (a trained clinician or a layperson at home), the environment of use, the part of the body it contacts and how, and the duration of that contact. Each element changes the hazards you have to consider. A device for trained clinicians can assume a level of competence a home-use device cannot. A device used for an hour has different risks than one worn for days. WearPump's, written for real: > **WearPump: Intended Use** For the subcutaneous delivery of rapid-acting insulin, for the management of Type 1 diabetes, in patients aged six years and older, used by the patient or a trained caregiver, in home and everyday environments, applied to the skin and delivering through a subcutaneous cannula, for a wear period of up to 72 hours. Every clause is a deliberate fence. Rapid-acting insulin, not any drug, so WearPump isn't on the hook for analyzing other infusions. Age six and older excludes the youngest children and the very different risks of dosing tiny bodies; reopening that market means reopening the analysis on purpose. Home and everyday environments means considering water and travel, but not the operating room. Each fence is a hazard chosen or explicitly left out. Once that boundary exists, it does something else: it tells you where to look for misuse. Anything a person does outside the intended use is either reasonably foreseeable misuse, which must be analyzed, or genuinely unforeseeable misuse, documented as beyond scope. You can't sort uses into those two bins until you've drawn the line. ## Misuse is predictable, not exotic Reasonably foreseeable misuse isn't an edge case you check a box for and forget. It's the richest source of real-world hazards you have, and it's almost entirely predictable if you're honest about how people actually behave, not how you wish they would. Tired, rushed, optimistic, improvising, and occasionally building an artificial pancreas in a bedroom because they got tired of waiting. The definition, from Clause 3.15: use of the device in a way not intended, resulting from readily predictable human behavior. The load-bearing word is _reasonably_. You're not required to anticipate every bizarre thing a human could conceivably do. You're required to anticipate what ordinary people, under real conditions, predictably will do, and the standard does require it. Worth separating two things that get blurred: misuse is using the device in a way not intended; use error is trying to use it correctly and still getting a result nobody wanted, because the design let the user slip. Entering the wrong dose because the screen is confusing is a use error, not misuse. Both belong in the analysis; use errors also get dedicated treatment under the usability standard, IEC 62366. Predictable misuse comes from a few familiar places: stretching a consumable to save hassle, skipping a step under time pressure, using a device on someone it wasn't meant for, working around an annoying limitation, or, like the looping community, extending a device past its purpose because it solves a real problem better than the approved option does. The moment a device has a wireless interface and an app, the question is unavoidable: what happens when someone connects something else to it? A third-party app, a home automation script. You may never approve it. But if it's predictable, and for a connected insulin pump it is, it belongs in the analysis. Five foreseeable misuse scenarios for WearPump, none of them a device failure: 1. 1\. A patient extends wear past 72 hours while traveling, accepting a degraded adhesive and higher infection risk. 2. 2\. A caregiver gives a correction dose without checking current glucose first, stacking insulin on insulin. 3. 3\. A parent lets a child swim past the splash rating. 4. 4\. A patient pairs WearPump with an unofficial app or a DIY loop. 5. 5\. A patient keeps using a single infusion set past its recommended life because changing it is inconvenient. Every one is a human being behaving the way human beings predictably behave, and every one becomes an input to the hazard analysis. ## The checklist almost nobody uses The third input is the most overlooked, and the most useful: the physical and functional facts about the device that could, under some circumstance, matter for safety. Not the hazards yet, just the characteristics. ISO 14971's companion guidance document, the technical report numbered 24971, has an annex that's essentially a checklist: roughly three dozen questions, each designed to surface a relevant characteristic. It's the closest thing the standard has to a hazard-finding machine, sitting in a document most engineers never open. Brainstorming from memory only finds the hazards you already worry about; you'll never find the class you've never thought about, precisely because you've never thought about it. A structured list, written by people who've seen devices fail in every imaginable way, forces categories you'd have sailed right past. The questions group into categories: energy (electrical, thermal, mechanical, radiated), materials and substances, biological and chemical, the patient, the user, the environment, data and connectivity, and end of life. Sample questions: Does the device deliver energy to the patient? Does it deliver substances? Are materials in contact with tissue? Is it susceptible to temperature, humidity, or electromagnetic fields? Is it controlled by software? Does it connect to other devices? Each yes isn't a hazard by itself. Each yes is a place to go look for one. WearPump's inventory: Category Characteristics Energy Pressurized reservoir, battery and motor, Bluetooth emissions, the plunger drive Substances and materials A potent drug, reservoir and fluid-path materials, an adhesive and cannula in sustained tissue contact Patient and user Type 1 diabetes, ages 6 to 80-plus, wide range of dexterity and health literacy, self-operated or by a caregiver, often distracted or in low light Environment Shower, pool, air travel, a hot car, everywhere a life happens Connectivity and software Bluetooth pairing, a companion app, a dosing algorithm, a foreseeable third-party or DIY connection End of life Home disposal with a used cannula, a battery, and drug residue This is an inventory, not a hazard analysis. Confusing the two produces a worse version of both. The hazards come from the inventory, in the next module. ## One characteristic, walked through Take a single line: WearPump stores insulin under pressure. On its own, that's just a fact. Now ask the hazard question: under what circumstance could that stored pressure cause harm? If an occlusion builds and the pressure has nowhere to go, a sudden release could push an unintended bolus into the patient. A neutral characteristic, plus a circumstance, becomes a hazardous situation. That move repeats for every line in the table. ## Why the order matters Three inputs define the scoped problem the hazard analysis actually solves: the intended use statement (the boundary), the foreseeable misuse list (how people predictably step outside it), and the safety-characteristics table (what about the device could ever cause harm). The order isn't arbitrary. Scope first, then identify, then estimate. The looping community is a hazard a manufacturer only catches by asking not just how the device is intended to be used, but how it will actually be used. Skip the scoping and jump to the FMEA, and the result is a rigorous analysis of a device that doesn't quite exist, while the real one gets used in ways nobody wrote down. If there's one thing to take from this: intended use is a decision, not a description. Choose it, write it narrowly, defend every word. Then look honestly at how real people will step outside it, and inventory everything about the device that could ever matter for safety, using the checklist instead of memory. Do that, and the hazard analysis in the next module has something solid to stand on. Module 5 runs a systematic hazard identification on this scoped foundation, using FMEA to trace forward from failures and fault tree analysis to trace backward from harms, turning these three inputs into WearPump's actual hazard table. If this was useful, Module 5 is next. More on product strategy, roadmaps, and fractional CPO work at [baserealitygroup.com](https://baserealitygroup.com), and the written version of this work goes out in a newsletter called The Build, at [davesaunders.net](https://davesaunders.net). [← Module 3ISO 14971 Clause 4: The Risk Management Plan and System](/guides/iso-14971/risk-management-plan-clause-4) [Module 5 →Hazard Identification with FMEA and FTA: ISO 14971 Risk Analysis Part 2](/guides/iso-14971/hazard-identification-fmea-fta) Get the next guide ISO 14971 is the first Operator’s Guide. ISO 13485, IEC 60601, and the GHG Protocol are in the queue. Leave your email and each new guide lands in your inbox the day it ships. Nothing else, no drip sequence. Email address Send me each guide Done. You’ll get the next guide when it ships. The standard is free. So is this course. Building the product is the work. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) --- # Overall Residual Risk Evaluation: ISO 14971 Clause 8 | Base Reality Group Source: https://www.baserealitygroup.com/guides/iso-14971/overall-residual-risk [← All ISO 14971 modules](/guides/iso-14971) Module 10 25:16 video + article Watch this module Module 10 of the video course. The article below covers the same ground in written form, so you can watch, read, or both. \# Every Risk Was Acceptable. The Whole Device Might Not Be. (ISO 14971 Clause 8) Picture your risk file with every row green. Every hazard found, every one controlled, every residual signed off as acceptable, every box ticked. You're done. Except ISO 14971 has one more question, and it's the one that catches the most careful teams, because it isn't about any single risk on the list. It's this: are all of them, together, on the same patient, every day, for years, still acceptable? Because ten green risks can add up to a device you shouldn't ship. Here's how that actually goes wrong. Imagine a device with twenty residual risks, every one individually acceptable, every one signed off. Then someone finally lays them out on one table and notices that eight of the twenty land hardest on exactly the same kind of patient: the elderly one, managing the device in low light, with poor eyesight, at the end of a long day. Each risk was fine on its own. Stacked on that one person, they're a systematic problem nobody saw, because nobody looked at the whole at once. Individual risk evaluation answers one question at a time: is this hazard, on its own, acceptable? It cannot answer the question about the whole device, and those are genuinely different questions with different answers. This is the step where a file that looks finished turns out not to be, and it gets skipped more than almost any other, because it feels like paperwork after the real work is done. It isn't paperwork. It's the real work, one more time, at a higher altitude. This is Module 10 of The Operator's Guide to ISO 14971. I'm Dave Saunders. I've spent more than 30 years commercializing technology, almost 20 of them in medical devices and regulatory work, including several surgical robots. Across this series we've built the WearPump risk file from nothing: eleven hazardous situations, identified, estimated, evaluated, and controlled, each one closed individually in Module 9. Now we stop looking at rows and look at the shape of the whole file. ## Clause 8, and the door that isn't there Clause 8 is the overall residual risk evaluation. After every individual risk is controlled and evaluated, you gather all of that residual risk together and make one judgment: considering the whole picture, and considering the medical benefit the device gives the patient, is the overall residual risk acceptable? If yes, you document the conclusion with your reasoning and the file closes on it. If no, you have exactly two options. Add more control, or abandon the device concept. Sit on that for a second, because it's the part with teeth. There's no third door. There's no option where the overall residual risk comes back unacceptable and you ship anyway because the launch date is set, or the investors are waiting, or the market window is closing. The standard gives you no commercial override on this decision. It's one of the few places where a standard about paperwork quietly has real power over a business. And it has to be its own evaluation, because the mistake is subtle. Teams finish the individual evaluations, see a file full of acceptable verdicts, and assume the overall evaluation is automatically satisfied. Every part passed, so the whole must pass. That's not how complex systems fail. A device can be a collection of individually reasonable risks that, stacked on one patient over years, add up to something you wouldn't put your name to. Clause 8 exists precisely to catch what the row-by-row view is structurally blind to. ## What "overall" actually means: three dimensions The word _overall_ is carrying three different kinds of weight. The first is **cumulative burden**. A patient wearing the WearPump isn't exposed to one risk at a time. They aren't carrying the occlusion risk on Monday and the connectivity risk on Tuesday. They're carrying all of it, at once, every hour the device is on their body. Every failure mode in the file is live, simultaneously, on one person. The overall view has to reflect that total load, not the eleven separate slices you evaluated it in. The second is **clustering by harm**, and this is the one that earns its keep. When you lay every residual risk on a single table instead of reading them one at a time, patterns jump out that were invisible at the row level. Risks that looked unrelated turn out to be different roads to the exact same place. On the WearPump, that's not hypothetical. Look at where our residuals actually lead. The motor driver fault over-delivers insulin. The silenced-alarm situation dumps a backed-up dose all at once when an occlusion finally lets go. The connectivity path could push an unintended command. The dosing algorithm could stack a double bolus. Four separate hazards, from four different corners of the device, hardware, firmware, wireless, and software, and every one of them ends in the same place: a severe low. Severe hypoglycemia, the dangerous drop that can put a person on the floor, or worse. Each of those four was individually acceptable. We closed each one honestly, on its own terms. But row by row, you never once see that they all drain into the same catastrophic outcome, and that matters for a hard, quantitative reason. The real-world probability of a severe low from this device isn't any one of those four risks. It's all four added together. Four small streams feeding one river, and the river is the number that actually reaches the patient. Individual evaluation hides that sum. Clause 8 is what makes you compute it. The third dimension is **exposure over time**. The WearPump is a chronic-use device. A patient wears it around the clock, changes it every three days, and does that for years, maybe decades. Every three days is roughly 120 wear cycles a year. Over a decade, that's more than a thousand exposures to every one of these risks. A per-cycle probability that looks reassuringly tiny gets multiplied by a thousand, and reassuringly tiny can become near-certain across a life. Put those together and overall residual risk isn't a headcount of how many yellow cells you have left. It's a structured look at the whole weight of the device resting on one real person over the entire life of their therapy. You can see why a stack of individually green rows doesn't answer it. ## Judge it against what? _Acceptable_ is a meaningless word in a vacuum. You need something to compare that total picture to, and if you pick the wrong comparison, the entire evaluation is worthless no matter how rigorous the rest looks. Here's the rule that fixes it. You never judge a device against zero risk, because zero is not a choice anyone actually has. A Type 1 diabetic doesn't get to pick a risk-free life. The only comparison that means anything is against the state of the art: what would this patient be doing right now, today, if your device didn't exist? For the WearPump, the answer is multiple daily injections, and that therapy is not safe. It carries its own serious, well-documented risks, and those risks, not zero, are your benchmark. So characterize the comparator in hard clinical terms. On multiple daily injections, severe hypoglycemia, the same dangerous low our residuals cluster on, is common. In the DCCT, the landmark trial of intensive insulin therapy published in the _New England Journal of Medicine_ in 1993, it ran around 60 events per 100 patient-years. Read that slowly: 60 severe lows, per hundred patients, per year, on the standard of care. That's not a rare edge case. It's the routine, documented danger of the therapy this device is trying to replace, and it's the number every part of your evaluation gets measured against. Now characterize the benefit in the very same currency, because a benefit measured in different units than the risk isn't a comparison, it's a dodge. Pump therapy versus injections, in the pooled trials, cut severe hypoglycemia by a wide margin. Pickup and Sutton, in _Diabetic Medicine_ in 2008, found injections carried close to three times the rate in the randomized trials, and more than four times across all the studies together. Same harm, same units, one therapy with three to four times more of it than the other. So run it. The WearPump's residual risks cluster on severe hypoglycemia, the honest weak point of the device. And the very same device drives severe hypoglycemia down by three to four times against the injections it replaces. In the one currency that matters most, the aggregate residual of our device sits far below the comparator. That's the overall residual risk argument, and it's strong precisely because it's measured in the same harm on both sides of the scale. To see the shape of the calculation and not just the vibe of it, put an illustrative number on it. Suppose the four risks that cluster on severe lows, taken together and translated from their residual probabilities into an event rate, land the device somewhere around 15 to 20 severe lows per 100 patient-years. That's an estimate built from our own file, not a measurement, and it's labeled as such. Set it next to the comparator: 60 on injections, 15 to 20 on the pump. The device isn't risk-free, and it's also plainly, quantitatively safer on the exact harm we were most worried about. Don't let a strong number make you sloppy, though, because there's one category where the device adds risk injections simply don't carry: the skin. The adhesive site, the cannula, the chance of a local infection or a reaction where the device meets the body for three days at a time. Injection patients wear nothing between doses; ours wear the device continuously. We own an entire class of harm the comparator doesn't have. The honest evaluation names that out loud, estimates it, and notes that these are lower-severity harms, mostly local and treatable, traded against the far more dangerous lows we're preventing. One discipline on those numbers, because this is exactly where files overreach and get caught. You don't need one perfect, precise harm rate for your device. Nobody has that before launch, and pretending you do is a tell. What you need is a defensible direction, grounded in the probability estimates already in your risk file and in the published literature on pumps like yours. When you put a device number on the table, mark it as the estimate it is. The comparator figures are sourced and citable; your device figures are reasoned projections. Keep that line visible and no reviewer can accuse you of grading your own homework. ## The statement, and who signs it All of that resolves into a single required output, and it isn't optional: the overall residual risk statement. This is not a cover letter and not a marketing line. It's a formal document in your risk file that captures your team's evaluated judgment, backed by evidence, in a form a stranger can audit. It has four parts, the same shape as the benefit-risk argument from the last module, scaled up to the whole device. 1. 1\. **The aggregate residual risk summary.** What the residual picture actually is, the clustering, the estimated rates, honestly stated. 2. 2\. **The clinical benefit evidence, cited, not asserted.** Every benefit claim carries a reference. 3. 3\. **The comparator risk profile.** The injection data, documented as rigorously as your own device. 4. 4\. **The conclusion.** An explicit, affirmative statement that the overall residual risk is acceptable in light of the benefit. Or a statement that it isn't, and what you're doing about it. No hedging in either direction. Written out, the conclusion reads like a verdict a team is standing behind, not a hope they're floating: _The overall residual risk of the WearPump, after all controls, has been evaluated against the benefit of continuous insulin infusion in adults with Type 1 diabetes. The comparator is multiple daily injection therapy, which carries a severe hypoglycemia rate of roughly 60 events per 100 patient-years. The device's aggregate residual harm sits substantially below that. The overall residual risk is acceptable._ Every clause in that paragraph traces to a document or a citation, and that's the entire point of writing it that way. This judgment doesn't belong to one person. The technical report that supports the standard is explicit: it's not a call an engineer makes alone in a spreadsheet at eleven at night before a deadline. The overall residual risk statement gets signed by regulatory, by clinical, and by quality. Three disciplines, three kinds of expertise, looking at the same aggregate picture and agreeing to it. If your acceptance statement carries one signature from one function, a reviewer notices immediately, and they're right to, because the whole idea is that no single perspective can see the whole risk. One more phrase lives inside Clause 8: _state of the art_. It doesn't mean the theoretical best device that could ever exist. It means the current, generally accepted standard of care as it's actually practiced today, by real patients who miss doses and forget things and live messy lives. You compare the WearPump against real-world injections, with their real-world 60-per-hundred rate, not against some idealized therapy that eliminates all risk. There's a useful gut check for it: would an informed, reasonable patient, one who genuinely understood both the benefits and the residual risks, choose this device over the alternative? It's not a legal test or a formula. It's a way of standing in the patient's shoes and asking whether the trade you're documenting is one a clear-eyed person would actually take. If the honest answer is yes, your evaluation has a spine. If you find yourself hoping they wouldn't look too closely, that's worth listening to. ## Acceptable is not silent Say you get to yes. The statement is signed, the file is closing. You're still not done, and this is the piece teams treat as a formality and regulators treat as a hard requirement. Clause 8, read together with the standard's rules on information for safety, requires that significant residual risks are communicated to the person using the device. Deciding the overall risk is acceptable means the device is allowed to exist. It does not mean the patient gets to be uninformed about the risks they take on when they put it on their body. Which residuals count as significant? The standard doesn't hand you a bright line, but the working test is clear enough: a residual is significant if it's something the patient could act on, or something that would change their decision to use the device at all. If knowing about it lets them catch a problem sooner, prepare for it, or choose more carefully, it belongs in the instructions. If it's truly invisible to them and nothing they do changes it, it may not. When in doubt, disclose. For the WearPump, the instructions carry a specific set: a warning about occlusion and dangerous lows, with guidance to check blood glucose more closely after a site change; a precaution about skin sensitivity, with a hard contraindication for anyone with a known reaction to the silicone adhesive; a firm three-day wear limit regardless of how much is left in the reservoir; what to do if the phone loses its connection mid-dose; and the specific fault codes that mean stop and call support. Every one traces back to a named hazard in the file, and that tracing runs both directions, which is exactly what an auditor checks. Every significant residual points to a section of the instructions, and every safety line in the instructions points back to a risk. If a significant residual has no home in the manual, that's a finding. If the manual warns about something nowhere in your risk file, that's a different and more embarrassing finding, because now your own two documents contradict each other in front of the reviewer. ## Five ways it fails an audit These are the exact patterns that generate findings and stall submissions. 1. 1\. **The aggregate step is simply never done.** Twenty acceptable rows, and the file just ends. No overall evaluation, no benchmark, no acceptance statement. The team assumed green rows summed to a green device, and the single most important judgment in the file was never made, only presumed. This is the most common one. 2. 2\. **Benefit is claimed with nothing behind it.** The statement says the device provides significant clinical benefit, and there's no trial, no registry, no citation near it. An assertion is not evidence; the reviewer will ask for the study, and if you can't produce it, the statement collapses. 3. 3\. **The comparator is cherry-picked.** The team quietly reaches for best-case injection numbers from a tightly controlled trial instead of the messier real-world rate. A flattering comparator makes your advantage look bigger than it is, and the distortion gets found. 4. 4\. **The significant residuals never reach the instructions.** Disclosure gets treated as the labeling team's problem, not the risk team's, and the traceability matrix ends up empty. 5. 5\. **One function signs off alone.** The engineer certifies the whole thing with no clinical and no quality judgment beside it, and the cross-functional requirement quietly didn't happen. All five are avoidable, and all five come from treating Clause 8 as the administrative tail of a process you're ready to be finished with. ## The recap Individual evaluation answers one risk at a time; Clause 8 asks whether the whole device, on one patient, over years, is acceptable. "Overall" carries three dimensions: cumulative burden (all of it at once), clustering (many risks draining into one harm), and exposure (a thousand cycles across a life). You judge the aggregate not against zero but against the state of the art, in the same units, which for the WearPump means severe hypoglycemia at roughly 60 per 100 patient-years on injections versus a reasoned 15 to 20 on the pump. You write it up in four parts, cite the benefit, document the comparator, conclude without hedging, and get three functions to sign. Then you disclose the significant residuals in the instructions, traceable both ways. Most risk files get read the way they were written, one row at a time. Clause 8 asks you to read yours all at once. So the question worth sitting with: if you laid every residual in your file on one table today, what harm would three or four of your risks quietly turn out to share, and who is carrying the heaviest version of that combined load? Module 11 turns this evaluation into the one document that has to carry it: the risk management report, the thing a reviewer opens first, and the place where a brilliant analysis goes to die if you write it badly. \--- If this way of thinking through a whole device is useful to you, I write more of it in my newsletter, _The Build_, at [davesaunders.net](https://davesaunders.net), and the rest of my work lives at [baserealitygroup.com](https://baserealitygroup.com). Subscribe so the next module finds you when it lands. [← Module 9Residual Risk and Benefit-Risk Analysis in ISO 14971](/guides/iso-14971/residual-risk-benefit-risk) [Module 11 →The Risk Management Report: ISO 14971 Clause 9](/guides/iso-14971/risk-management-report) Get the next guide ISO 14971 is the first Operator’s Guide. ISO 13485, IEC 60601, and the GHG Protocol are in the queue. Leave your email and each new guide lands in your inbox the day it ships. Nothing else, no drip sequence. Email address Send me each guide Done. You’ll get the next guide when it ships. The standard is free. So is this course. Building the product is the work. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) --- # Post-Market Surveillance: ISO 14971 Clause 10 | Base Reality Group Source: https://www.baserealitygroup.com/guides/iso-14971/post-market-surveillance [← All ISO 14971 modules](/guides/iso-14971) Module 12 29:01 video + article Watch this module Module 12 of the video course. The article below covers the same ground in written form, so you can watch, read, or both. \# The Risk File Doesn't Close When You Ship (ISO 14971 Clause 10) Your pump launched four months ago. Sales are good, nobody is panicking, the first big safety report is still months away. Then the monthly complaint count lands on someone's desk. Twelve occlusion reports. Not across the life of the product. In a single month. So here is the question that whole month now turns on. Is that a problem? Your pre-market file said occlusion was rare, and acceptable. Your notified body agreed. But twelve in thirty days is a number your file never saw. You have a signal. What you do not have yet is a verdict. The distance between those two words is where this whole module lives. A signal is just a number that looks wrong. A verdict is what you get after you have found the denominator, chased the root cause, and asked honestly whether the risk you accepted before launch is still the risk you actually have. The dangerous gap is the time between the signal arriving and someone connecting it to the risk file. That gap is where devices hurt people, and closing it is the entire job of the last clause in the standard. This is Module 12 of The Operator's Guide to ISO 14971, the finale. I'm Dave Saunders. I've spent more than 30 years commercializing technology, almost 20 of them in medical devices and regulatory work, including several surgical robots. Across this series we've built the WearPump risk file from an empty page: eleven hazardous situations, each identified, estimated, controlled, and closed. We judged the whole device acceptable against injection therapy, wrote the report, got it signed, and shipped. By every pre-market measure, the risk work is done. And that is exactly the moment the standard says you are most at risk of stopping too early, because the device just met its first real patients, and they are about to teach you things no analysis ever could. ## What Clause 10 actually asks for Clause 10 is titled production and post-production activities, and you could read that title and assume it is about manufacturing quality. It is not, or not only. The core of it is three verbs. You establish a system to collect information generated after the device is released, about your device and about similar ones. You review that information against your risk estimates. And when it changes the picture, you feed it back into the risk process. Collect, review, feed back. The clause wants all three. I want to be blunt about the failure here, because it is everywhere. You cannot satisfy Clause 10 by saying you have a complaint system. A complaint system that collects reports and files them is one third of the requirement, and it is the easy third. You also need a review step that measures those reports against the numbers in your risk file, and a feedback path that actually changes the file when the numbers move. Collect without review is a filing cabinet. Review without feedback is a report nobody acts on. The clause requires the whole loop, closed. ## Surveillance is the input, not the judgment > Post-market surveillance is not the same thing as risk management. Surveillance is the stream of real-world information coming in the door. Risk management is what you do with it: the analysis, the judgment, the decision to act. That distinction clears up a lot of confusion. Your surveillance system feeds your risk process. Teams that treat the two as one thing, or worse, hand surveillance to a quality group that never talks to the risk team, end up with a pile of data and a risk file that never learns from it. One piece of the incoming stream is a legal line, not a judgment call. Some of what comes in the door is not just risk data, it is a reportable event: a death, a serious injury, a malfunction that could cause one if it recurred. When one of those lands, a clock starts, and the regulator has to be told on a defined timeline whether or not you have finished investigating. In the United States that duty lives in the medical device reporting rule. In Europe it is the vigilance system. Your surveillance has to catch those the instant they arrive and route them two ways at once, to the regulator and into the risk file. ## The five sources, and why the mix is the point So what do you collect for a device like the WearPump? Five sources, and each one sees something the others miss. 1. 1\. **Complaints and adverse-event reports.** Every complaint that arrives, by phone, email, or through the app support channel, gets captured in a controlled system, and some events trigger mandatory reporting by law. The shift in thinking is this: a complaint about occlusion is not a customer-service ticket. It is a data point about the probability of a hazardous situation you already estimated on paper, before launch. It is the real world checking your math. 2. 2\. **The companion app.** The WearPump talks to a phone, and that phone is a surveillance instrument. Every logged connection failure, unexpected alarm, or dosing exception is data, and with the right consent and privacy controls, it gives you a population-level signal far earlier than complaints can. A patient with a Bluetooth dropout usually does not file anything. But the app knows it happened, and if it is happening to thousands of patients, the app can tell you before anyone picks up the phone. 3. 3\. **Field service and returns.** When a unit comes back, why did it come back, and when a technician opens it, what do they actually find? Returned devices are physical evidence, and root-cause work on them can confirm or flatly contradict what the complaint descriptions say. A patient tells you the device failed; the returned unit tells you whether it failed the way they thought or a different way entirely. 4. 4\. **Published literature and similar devices.** Competitors' incident databases, adverse events reported against pumps like yours, peer-reviewed papers on wearable insulin delivery. The technical report that accompanies the standard is explicit that similar devices are in scope. If another wearable pump is seeing unexpected occlusion rates, that is relevant to your file, because your patients live in the same physical world theirs do. 5. 5\. **Structured user feedback.** Surveys, calls logged by your clinical support team, reports from the doctors and nurses who prescribe the device. This is where usability problems surface first: the application technique that is harder than you thought, the training gap, the placement mistake people keep making. Those show up in feedback long before they escalate into complaints or harm. The point of five sources is not five separate reports gathering dust. It is triangulation. A complaint says the device failed. The returned unit tells you how. The app telemetry tells you how often, across the whole population. The literature tells you whether the rest of the field is seeing the same thing. The user feedback tells you whether it is really a training gap wearing a hardware costume. No single source is trustworthy on its own. It is where they agree, and disagree, that the real signal lives. ## Cadence, and the trigger A surveillance system without a rhythm is a pile of data with good intentions. So you give it one. Weekly, someone runs a simple count of complaints by category, and those categories map straight to the hazards in your file: occlusion, adhesion failure, connection loss. That is your early-warning read, no analysis, just the count. Monthly, you move from raw counts to rates, complaints per use, because a count with no denominator is meaningless. Quarterly, you run a full review across all five sources. Once a year, you write the formal safety summary. And then the one thing that turns all that watching into action: a trigger, defined as a number, not a feeling. Not "if the rate seems high." Not "if the team feels concerned." A specific number, written down before you ever need it. For the WearPump, say it plainly: if the observed rate for any hazard exceeds twice its pre-market estimate for two consecutive monitoring periods, escalation is mandatory. The instant that condition is met, you stop the routine monthly trending and convene a real risk review. The reason it has to be a number on paper is simple. When someone new is watching the dashboard at the end of a long week, the trigger cannot live in one experienced person's gut. It has to be on the wall. ## The signal: twelve occlusions Come back to those twelve occlusions, because we now have exactly that situation, and we are going to work it all the way through. First question the team asks, before anyone panics: what is the denominator? Twelve complaints out of how much use? You pull the distribution data. Say there are around five hundred WearPumps active in the field. Each patient runs a fresh wear cycle every three days, so in a thirty-day month that is roughly ten uses per patient, and five hundred patients gives you about five thousand uses that month. Twelve events against five thousand uses is a rate of about one in four hundred, call it two to three events per thousand uses. The denominator is where teams quietly go wrong, because it is harder than the count. How many devices are actually on patients right now, not sitting in a warehouse? How many wear cycles did each of them really run this month? Someone who started three weeks ago has fewer uses than someone who started at launch. Get the denominator too small, and a manageable rate looks like a crisis. Get it too big, and a real crisis hides in the noise. The number twelve is the easy part. The five thousand underneath it is the part that takes real work to trust. Now compare that to what your file predicted. Before launch, you estimated the occlusion pathway, the same one you first mapped in the fault tree back in Module 5, as remote. On your scale, remote meant fewer than one harmful event in ten thousand uses. Your observed rate is roughly one in four hundred. That is more than twenty times higher than you estimated, and it is not the first month either. Month three was already running elevated. Two consecutive periods above the line. The trigger is not close. It is blown. Before the review starts, the team sorts the twelve by what actually happened to the patient. Most caught it, saw their glucose climbing, changed the site, and corrected before it got dangerous. A few swapped the device out with no real harm. But one patient ended up hospitalized with diabetic ketoacidosis, a serious adverse event, and that one is already moving through mandatory reporting. So the picture is twelve events, one serious injury, a rate more than twenty times the estimate, two months running. That one hospitalization also starts a second clock. Telling the regulator, on their timeline, that a serious injury happened is a compliance duty. Figuring out whether your risk estimate was wrong is a risk-management duty. The same twelve complaints trigger both, and a mature system runs them side by side instead of letting the investigation hold up the report. ## The afternoon the pattern fell out Twelve tells you the frequency. It does not tell you the mechanism, and you cannot fix a mechanism you have not found. So your quality engineer pulls all twelve records and builds a simple matrix: wear duration at the time of the event, cannula lot number, whether the patient had in-person training. Within an afternoon, a pattern falls out. Every one of the twelve happened after forty-eight hours of wear. Nine of the twelve share a single cannula lot, one where the tubing wall sat on the thin end of the tolerance. And several of the patients had no hands-on training, just the written instructions in the box. Pull that cannula thread, because it is instructive. The suspect lot was not out of specification. The tubing wall was inside the tolerance you set, just sitting at the thin end of it. Which means the problem is not a broken part. It is a specification drawn a little too loose for how the device is actually used, late in a wear cycle, under real body movement. That is a harder finding, and a more honest one, because it sends you back not just to the supplier but to your own drawing, to ask whether the tolerance itself was wrong. Now map that to the risk file, and notice what it is and what it is not. This is not a new hazard. You are not discovering something you never imagined. It is the occlusion pathway you already had, the one that fed the fault tree, now telling you that your pre-market probability estimate was optimistic. The real world runs occlusions more often than your bench analysis assumed, especially late in the wear cycle, with a marginal cannula lot, in the hands of an untrained user. The hazard was right. The number was wrong. ## The updated estimate, and the acceptance you reopen The observed rate pushes the occlusion probability up a full band, from remote to occasional. Severity does not move: the harm is still under-delivery driving blood sugar up, serious, dangerous if it runs to ketoacidosis, exactly as you scored it. So the risk moves, driven by probability, and you write it down properly. Old estimate, new estimate, the real-world data that drove the change, the date, and a signature. You do not quietly edit a number. You record a change. And you do not treat one month as the whole story. You plot the rate across months. Month three ran a little hot. Month four crossed the line. That climbing trend is more convincing than any single data point, because it rules out a one-off bad week. When you take this to a regulator, or to your own management, the story is not twelve complaints in isolation. It is a rate that was already rising, that you caught on the second period and acted on before it climbed a third. That is a system working, and it is a very different story to tell than one alarming number. That new number does two things immediately. First, the acceptance you granted this risk before launch was based on the old probability, so that acceptance is no longer valid, and its rationale has to be reopened and re-examined. You cannot lean on a judgment built from a number you now know was wrong. Second, you have to ask whether more risk reduction is now reasonably practicable. And the root cause has already answered that: a cannula lot problem, a wear-time pattern, a training gap. Those are not theoretical. They are things you can act on this month. ## The decision, and the time-bound call Three options land on the table. One, hold the suspect cannula lot and open a formal investigation with the supplier. Two, make hands-on training mandatory before any patient gets a device, in person or a verified remote session, no exceptions. Three, cut the labeled wear time from seventy-two hours down to forty-eight, since every single event happened past the forty-eight-hour mark. The first two are cheap, fast, and clearly worth doing, so you commit to both inside sixty days. The third one is harder, and you do not decide it today. Dropping to forty-eight hours would directly address the timing, but it also takes something real from the patient. That extra day of convenience is part of why the pump beats injections in the first place. So you make a time-bound call instead: implement the lot hold and the training now, watch the occlusion rate over the next ninety days, and if fixing the lot and the training brings it back down, you may not need the labeling change at all. That is benefit-risk analysis in the post-market world, and it is the whole course coming back around. You weighed a residual risk, the occlusions, against a benefit to the patient, that extra day of wear, and made a call you could defend, with a date to revisit it. It is the exact logic from Module 9, except now the numbers are real, measured in the field, instead of estimated at a desk. The process does not change when the device ships. The evidence just gets better, and the stakes get more concrete. ## Closing the loop, the step teams skip Here is where teams lose the thread: documenting the loop closure. The analysis means nothing if it does not land back in the file. So the occlusion entry gets its revised estimate, dated, referencing the surveillance report that drove it. The two new controls get their own records, the lot hold and the training requirement, each with an owner and a verification method. The surveillance review itself gets filed as a dated, signed record. The corrective action in your quality system gets cross-linked to the risk file, and the risk file gets cross-linked back. And a ninety-day date goes on the calendar to check whether any of it worked. That cross-link is not bureaucratic decoration. Picture the auditor a year from now, opening the occlusion entry. If they see the old estimate, the original acceptance, and nothing else, they ask the obvious question: you had a serious adverse event and a complaint spike, where is that in the risk file? If your answer is "it is over in the corrective-action system," the follow-up is "and the two are supposed to be linked, so where is the link?" The link is what proves you did not just handle the complaint, you let it change your understanding of the device. Miss it, and a good response looks like a cover-up. All of this rolls up, once a year, into a single formal summary. In Europe it is called a periodic safety update report, and for a device like this one it is a requirement, not a courtesy. It gathers everything the year taught you, the complaint rates, the signals, the changes you made, like our occlusion revision, and it states in writing whether the benefit-risk balance you certified at launch still holds. It is the living file, compressed into one document a regulator reads to decide whether they still trust your device on the market. ## The living file That is really the whole idea of Clause 10: the living risk file. A file last touched on the day the device was approved is a file that stopped learning the moment the device met a patient, and that is the worst possible time to stop. So every surveillance review ends one of two ways, an update or a documented finding of no change required, and both get recorded. Even a quiet month where nothing crossed a threshold gets written down, because the absence of a signal is itself evidence that your system is running and watching. When an update does happen, it does not erase what came before. The file is additive. The new version contains the old one, and the revision history shows every change, every date, every author, every reason. So when that auditor reads the current version, they can trace the entire path, from your pre-market estimate, to the real-world signal, to how fast you acted, to what you decided and why. That traceability is not just a requirement. It is your proof, in writing, that you were paying attention the whole time. Think about how a regulator opens your file after something has gone badly wrong in the field. They are not asking whether the device was designed well, they already know something failed. They are asking four things. Did you know this was possible? Did you estimate it honestly? Did you control it? Did you respond when the data warned you? A living file, with a dated trail from estimate to signal to action, answers all four. A file frozen on launch day answers none of them, and that silence is its own kind of answer. There is also a reason to do this well that has nothing to do with fear of an auditor. The team that catches its own occlusion signal in month four and acts has better real-world data by month twelve than any pre-market study could have handed them. They ship a better cannula next year, because the field told them exactly where to look. The team that waits to be told, by a regulator or a lawyer, learns the same lesson years later, at ten times the cost. Surveillance is not a tax on a finished product. It is the fastest, truest feedback you will ever get on how your device really behaves. ## The complete file, and one honest admission So here it is, one last time: the complete WearPump risk file, the thing we have been building brick by brick since Module 2. Four original hazards from the first pass. Four more from the failure-mode analysis. Three from the fault trees. Eleven in all, each estimated, each controlled through the hierarchy, each residual either accepted outright or carried on a benefit-risk argument. The whole device judged acceptable against injections. The report written and signed. And now one entry freshly reopened by twelve occlusions in the real world. That is not a finished document. It is a living one. I want to be straight with you about something, now that the whole file is in front of us. If you have been taking careful notes across these twelve modules, you may have caught a place or two where we described a hazard a little differently from one module to the next, or sharpened a number as our thinking got clearer. That is not a slip I am hoping you missed. That is the process, running in front of you. Our understanding of the silenced-alarm situation got sharper the second time we looked at it. That is exactly what a living file does. When you find one of those discrepancies in your own file, you make a real decision about it, and this is a genuinely useful distinction to carry. Was it just how you described the thing, a loose wording, a stale reference? Then you correct the record under document control and move on. Or was it a real gap in the analysis, a risk you underestimated, a control that was not doing what you claimed? Then it is not an edit. It is a finding, and it goes through the improvement cycle, the same corrective-action loop we just ran on the occlusion signal. Knowing which of the two you are looking at is a real skill, and now you have it. So the honest answer to "when is a risk file done" is that it is not, not as long as the device is on the market. It closes for a design freeze and opens again for a change. It absorbs a field signal, revises, and re-closes. It is a document that breathes for the entire life of the product, and the day it stops breathing is the day you have stopped listening to the people wearing your device. ## Your turn So here is the question I want to leave you holding. If your device, the real one you are working on right now, received twelve complaints of the same failure mode in a single month, how long would it take your team to know? Not to hear about one of them, but to know that all twelve had arrived, that they shared a pattern, and that the pattern needed to reach someone who could hold it up against the estimate in the risk file. Less than a week? A few weeks? Or are you genuinely not sure? Your answer tells you exactly where your surveillance system needs the most work. And that is the series. You have done something here that most people in this field, even experienced ones, have never actually done. You built a complete risk management file, from an empty page to a living document, on one real device, following every clause of the standard in order. You know why the process exists, how to run it, and how to keep it alive after launch. That is not a small thing to carry into your next project, and the patients on the other end of that work are the reason it matters. \--- If this series has been useful to you, I write more of this thinking in my newsletter, _The Build_, at [davesaunders.net](https://davesaunders.net), and the rest of my work is at [baserealitygroup.com](https://baserealitygroup.com). Subscribe there so the next series finds you when it starts. Now go keep that risk file alive. [← Module 11The Risk Management Report: ISO 14971 Clause 9](/guides/iso-14971/risk-management-report) Get the next guide ISO 14971 is the first Operator’s Guide. ISO 13485, IEC 60601, and the GHG Protocol are in the queue. Leave your email and each new guide lands in your inbox the day it ships. Nothing else, no drip sequence. Email address Send me each guide Done. You’ll get the next guide when it ships. The standard is free. So is this course. Building the product is the work. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) --- # Residual Risk and Benefit-Risk Analysis in ISO 14971 | Base Reality Group Source: https://www.baserealitygroup.com/guides/iso-14971/residual-risk-benefit-risk [← All ISO 14971 modules](/guides/iso-14971) Module 09 28:31 video + article Watch this module Module 9 of the video course. The article below covers the same ground in written form, so you can watch, read, or both. \# Residual Risk, Done Honestly: When "We Reduced It" Isn't the Same as "It's Acceptable" Most risk-management tutorials end with a tidy matrix full of green cells. Real ones don't. When you control a hazard honestly, you usually don't get to zero. You get to smaller. The risk is lower than it was, and it's still sitting there on the page. So you end up staring at a hazard that responded to everything you threw at it and still didn't clear your line, and the question that stops teams cold is a simple one: now what? That is the whole job of this module. When a risk won't go to zero, how do you decide it's acceptable, and how do you prove that to someone who wasn't in the room? ISO 14971 has a specific process for exactly this moment, and it runs straight into one of the most consequential arguments you'll ever write in a risk file. This is Module 9 of The Operator's Guide to ISO 14971. I'm Dave Saunders. I've spent more than 30 years commercializing technology, almost 20 of them in medical devices and regulatory work, including several surgical robots. The series builds a full risk management file on one running device: WearPump, a wearable subcutaneous insulin pump for Type 1 diabetes. Four clauses close that file, and we take them in order: 7.3, residual risk evaluation; 7.4, benefit-risk analysis; 7.5, the new hazards your controls create; and 7.6, the completeness check. ## Where WearPump stands Coming out of Module 8, the WearPump file is not a clean sheet of green. The motor driver fault, the one that can over-deliver insulin, came down out of the elevated band into something we can live with. The connectivity hazard moved off red. The yellows came down too. And along the way we created a brand-new hazard of our own, alarm fatigue, by reaching for a louder alarm. That whole picture is what we close out here. ## Clause 7.3: re-scoring what's left Residual risk is whatever is left after your controls are in place. Clause 7.3 says you evaluate it, one risk at a time, against the criteria you already wrote down. Not new criteria you invent now that you can see the answer. The exact acceptability line you set back in your risk management plan in Module 3, before you knew how any of this would land. You set the bar first, then you measure against it. That order is the entire point. Here's what re-estimation actually does. After a control goes in, you score the risk again. Both numbers can move, but they don't move the same way. A control almost always changes probability, and it rarely touches severity, for a clean reason: a control interrupts the chain of events between the hazard and the harm. It makes the harm less likely. That's a probability effect, every time. Severity is different. If the control fails and the patient is exposed anyway, the harm is the same size it always was. An overdose is an overdose whether or not there was an alarm that failed to stop it. So the default rule is blunt. You may lower probability when you can explain why. You don't get to lower severity because you feel better about the risk now. The harm is the harm. ## When severity legitimately drops There is one real exception, and WearPump used it twice, so it's worth getting exactly right. Severity can drop when the control changes the injury itself, not the odds of it. The motor driver fault started as high as severity goes, because a full electronic runaway could deliver a fatal amount of insulin. Then we added two hard caps: a firmware dose ceiling, and a 3 mL reservoir that physically limits how much can ever come out. Walk the worst case again and it's no longer a lost life. It's a serious overdose the body can survive with treatment. The design changed the size of the harm, so severity genuinely comes down, from an S5 to an S3. The connectivity hazard did the same thing a different way. It came in at severity 4, a critical harm, because a bad wireless command could drive a dangerous dose. The fix wasn't a warning about pairing. It was a design decision: the pump is completely safe on its own, with the phone switched off and thrown in a drawer. Once the phone can never command anything unsafe, the worst case of a connectivity failure is no longer a dangerous dose. Severity falls to an S2. That's the line you have to hold in the documentation. If severity drops, you show the mechanism that made it drop: the dose ceiling, the reservoir geometry, the safe-standalone design. The specific thing that shrank the injury. What you can never write is that you lowered a number because the team was more comfortable. An auditor will challenge any severity drop that doesn't come with a mechanism attached, and they'll be right to. Probability gets the same discipline in reverse. You don't just write a smaller number. You name the control, state how it cuts the odds, and point at the evidence. Not "the sensor helps." Something concrete: the occlusion sensor caught the blockage in 95 out of 100 bench tests, so probability drops a full band, and here is the report. The bad version is "we added a control and re-estimated lower." The good version has a number and a report attached. ## The mistake that fails more files than any other It hides in the gap between two sentences that sound identical and are not. > "We can't reduce this any further" is an engineering fact, a statement about the limits of what you can practicably build. "This risk is acceptable" is a judgment about whether what's left is worth it. One is about your hardware. The other is about the patient. Teams weld those two together and it costs them. Reaching the end of your engineering does not automatically make the leftover risk okay. So for a stubborn residual you owe two things, not one. First, the demonstration that you hit the limit of practicable control: you tried the design changes, you tried the guards, and here's why you stopped where you stopped. Second, the argument that the risk that remains is outweighed by something. Write only the first and you've described an engineering dead end and called it a safety case. It isn't one. And the second argument has a clause of its own. ## Clause 7.4: benefit-risk is not an escape hatch When a residual risk won't clear your criteria and you genuinely can't reduce it further, Clause 7.4 gives you one more move: you may weigh whether the benefit of the device outweighs that specific risk. This clause has a bad reputation it half deserves. Done lazily, it's the escape hatch teams reach for when they don't want to engineer something out. Done properly, it's the most evidence-heavy paragraph in the entire file. Two rules govern it. First, the benefit has to be clinical. A better outcome for the patient: longer life, fewer emergencies, a real improvement in how they live. What does not count, and the standard is explicit, is anything on your side of the table: not the schedule, not the money you already spent, not what a redesign would cost this quarter. Benefit means benefit to the person wearing the device. Second, you never weigh a device against zero risk, because zero isn't on the menu for anyone. You weigh it against what the patient would actually be doing instead. For a Type 1 diabetic, the alternative isn't a risk-free life. It's multiple daily injections, 4 to 6 needles a day, for the rest of their life, and that therapy carries its own serious risks. That's your comparator, and you have to name it. ## The residual that stays catastrophic Now the hazard that survives everything. It's the silenced-alarm situation: insulin backs up behind an occlusion, the alarm that should catch it has been silenced, and the patient is asleep. When the blockage finally gives way, the backed-up dose goes in all at once. A severe low, at three in the morning, with no alarm to wake anyone. We drove the probability way down. But the severity hasn't moved, because if that chain completes, a person can die in their sleep. Rare, yes. Catastrophic, still. And we're honestly at the limit on it. We already made the smart move in Module 8, an alarm that fires only when it truly matters so it doesn't get tuned out. We can't make an undetected occlusion during sleep impossible without taking away function the patient needs. So probability is as low as we can practicably get it, and severity is inherent. That's exactly the situation Clause 7.4 was written for. We stop trying to shrink it, and we start weighing it. ## The evidence, and how much it can carry Here's the clinical evidence we've held back since Module 7. Pickup and Sutton, a study in _Diabetic Medicine_ in 2008, pooled the trials comparing pump therapy against injections in Type 1 diabetes. Severe hypoglycemia, the dangerous lows that put people on the floor, happened far less often on a pump: in the randomized trials, injections carried close to three times the rate; across all the studies together, more than four times. The device attacks the exact harm we're worried about, and it attacks it hard. A benefit-risk case is only as strong as the evidence underneath it, so it helps to know the hierarchy. Randomized controlled trials and the meta-analyses that pool them sit at the top. Registry data and real-world evidence sit below that, useful but noisier. A single expert saying "it seems fine" sits at the very bottom. For an insulin pump you're lucky, because the literature comparing pumps to injections is deep and decades old, so you get to stand on the strongest rung. Put a floor under how common the harm is, too. In the DCCT, the landmark trial published in the _New England Journal of Medicine_ in 1993, severe hypoglycemia on injections ran around 60 events per 100 patient-years. That's not a freak occurrence. It's a routine danger of the therapy this pump replaces. There's a quality-of-life piece as well: injections mean 4 to 6 needles a day, somewhere between 1,460 and 2,190 a year, while the pump replaces that with one cannula change every three days, roughly 120 a year. For a person managing a lifelong disease, that difference shows up in whether they stick with the therapy at all. Now assemble it for the silenced-alarm risk. Yes, there's a rare path where the device fails to catch a dangerous low and someone is harmed. But the same device, used as intended, prevents far more of those exact dangerous lows than the injections it replaces. The benefit and the risk are measured in the same currency, severe hypoglycemia, and the benefit is the larger number by a wide margin. The residual is real, and the benefit outweighs it. You don't say that with confidence. You say it with citations. A real benefit-risk argument has four parts, every time: 1. 1\. **Name the residual risk,** its score, and the fact that you couldn't reduce it further. 2. 2\. **Name the comparator therapy** the patient would use instead. 3. 3\. **Cite the clinical evidence,** by author and year, with the actual numbers. 4. 4\. **Draw an explicit conclusion.** No "it appears that," no "we believe." The conclusion reads like a verdict, not a wish: _The residual risk of an undetected occlusion during sleep is accepted. It cannot be reduced further without removing clinical function the patient depends on, and the device delivers a documented reduction in severe hypoglycemia, relative to multiple daily injection therapy, that outweighs it._ That's a sentence a reviewer can check. They can pull the study, confirm the comparator, and run the logic. And that is exactly the checklist a notified body reviewer runs: did you name the risk, name the comparator, cite real clinical data, and reach a conclusion. Four boxes. If one is missing, it reads as an opinion dressed up as an analysis, and it comes back. One more thing: accepting a residual doesn't mean hiding it. A significant risk you've decided to live with gets disclosed to the person living with it, in the instructions for use. Accepted is not the same as buried. ## Clause 7.5: the loop nobody runs Clause 7.5 is the one I see skipped more than any other in the whole standard: the risks your own controls create. Every control you add changes the device, and a changed device can fail in a brand-new way it couldn't before. A safety feature is never free. We don't have to hunt for an example, because we walked straight into one. On the silenced-alarm hazard, the obvious instinct was to make the alarm louder, harder to silence, add a backup tone. All reasonable. All aimed at safety. And every one of those moves quietly makes the device more dangerous. It's called alarm fatigue, and it's not a theory. The Joint Commission, which accredits hospitals across the United States, reviewed alarm-related events over a three-and-a-half-year window: of 98, 80 ended in a patient death, and by their estimate somewhere between 85% and 99% of alarm signals need no clinical action at all. So people stop trusting the alarms and tune them out. A louder alarm doesn't add safety. It adds noise, and the noise gets someone killed. Read that back against what we were about to do. We were going to protect the patient with a more aggressive alarm, and a more aggressive alarm is itself a documented cause of patient death. The control became the hazard. The real fix wasn't more alarm, it was a smarter alarm, one that fires only when it genuinely matters so it keeps its meaning. And here's what 7.5 actually demands: that new hazard, alarm fatigue, doesn't just get noticed and waved at. It goes back into the process as its own line, with its own estimate, its own control, its own residual. You added a control, you asked what it just broke, you found something, and you run it through the whole loop again. That's not the process failing. That's the process working as designed. Once you see it, it's everywhere: we also considered dropping the adhesive patch for a mechanical strap and clip, which would have solved the skin reaction cleanly, and introduced a new pressure-injury hazard and a device more likely to pop off mid-exercise. Every control is a trade, and 7.5 is where you're forced to count the cost, not just the benefit. ## Clause 7.6: the completeness check The last clause is the gate before you're allowed to freeze the design. Before this design goes anywhere, you confirm that every hazardous situation you ever identified has been dealt with. No gaps. No hazard you wrote down in Module 5 and quietly lost track of. You run it as a cross-check: take the hazard list from your analysis and walk it against your risk file, one row at a time. Every hazard needs four things, and you're hunting for three kinds of gap. **Every hazard needs:** * An estimate. * A control, or an explicit note that says "none required, and here's why." * A re-estimated residual. * A documented resolution. **You're hunting for:** * A hazard with no row at all. * A blank control column (blank doesn't mean acceptable, it means unfinished). * A control with no verification linked to prove it was ever real. Run it on WearPump's eleven hazards. Seven we scored and worked on camera across these modules: the motor driver, the connectivity path, the silenced alarm, the reservoir seal, the Bluetooth lockup, the moisture combination, and the adhesive site. The other four we named but didn't fully walk, the pressurized reservoir, the concentrated insulin, the electrical energy, and the original adhesive and cannula hazard, each still gets its row, documented at its level, with the reason it sits where it sits. Eleven identified. Eleven addressed. And now the honest part the tidy tutorials skip. This file does not end all green. The silenced-alarm risk is accepted, not eliminated. The motor driver still carries a residual, smaller than before, but not zero. That's correct. A risk file where every single cell is green is usually a file that lied somewhere, or a device so trivial it never needed this process. Real risk management ends with residuals you've chosen, on purpose, with your eyes open. So be precise about what "closed" means: not risk-free, but every hazard addressed, every acceptance documented, every benefit-risk argument on paper with its citations, every control backed by verification. Complete for this version of the design, not forever. A change tomorrow reopens it. And closing is a formal action, not a vibe: you capture the date, name the person with authority to approve it, and they sign the completeness check. ## The question the standard won't answer You can do all of this perfectly, the re-estimation, the benefit-risk, the completeness check, every citation in place, and there's still a question underneath it the standard won't answer for you. If the person wearing this pump were someone you love, would the residual risk you just documented as acceptable still feel like enough? Not on paper. In your gut, at two in the morning. That question doesn't have a clean answer, and it isn't supposed to. The standard gives you a disciplined, evidence-based way to make the decision. It doesn't make the decision for you. Because the decision still belongs to people: engineers, a clinical lead, a quality manager, a regulatory lead, who have to look at each other and say, out loud, that they believe this device is safer than the alternative and they're ready to defend that to anyone who asks. That's what "acceptable" actually means past the matrix cell. It's a claim real people put their names to, and the framework is how you make that claim honestly instead of hopefully. ## The recap When you control a risk honestly, you usually reach smaller, not zero. Clause 7.3 makes you re-score what's left against the line you set first: controls move probability, and only move severity when the design changed the injury itself, with the mechanism shown. Clause 7.4 handles the residuals that won't clear, by weighing clinical benefit against the risk, in the same currency, against the real comparator, cited by author and year, with an explicit conclusion. Clause 7.5 sends the new hazards your controls create back through the loop. Clause 7.6 confirms every hazard is addressed before you freeze the design. The file closes with residuals you chose, documented and signed, not with a wall of green. Module 10 asks the question a row-by-row review can't: are all of these residual risks, together, on one patient, still acceptable? That's Clause 8, the overall residual risk, and it catches what the line-by-line view misses. \--- If this way of working through a device is useful to you, I write more about challenges like this in my newsletter, _The Build_, at [davesaunders.net](https://davesaunders.net), and you can find the rest of my work at [baserealitygroup.com](https://baserealitygroup.com). Subscribe so the next module in this series finds you when it lands. [← Module 8The Risk Control Hierarchy: ISO 14971 Clause 7](/guides/iso-14971/risk-control-hierarchy) [Module 10 →Overall Residual Risk Evaluation: ISO 14971 Clause 8](/guides/iso-14971/overall-residual-risk) Get the next guide ISO 14971 is the first Operator’s Guide. ISO 13485, IEC 60601, and the GHG Protocol are in the queue. Leave your email and each new guide lands in your inbox the day it ships. Nothing else, no drip sequence. Email address Send me each guide Done. You’ll get the next guide when it ships. The standard is free. So is this course. Building the product is the work. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) --- # The Risk Control Hierarchy: ISO 14971 Clause 7 | Base Reality Group Source: https://www.baserealitygroup.com/guides/iso-14971/risk-control-hierarchy [← All ISO 14971 modules](/guides/iso-14971) Module 08 32:55 video + article Watch this module Module 8 of the video course. The article below covers the same ground in written form, so you can watch, read, or both. \# The Hierarchy of Controls You Know Is the Wrong One for a Medical Device If you've ever seen a hierarchy of controls, there's a good chance you saw the wrong one. The version almost everybody knows, elimination, substitution, engineering controls, on down to protective equipment, was built to protect workers on a factory floor. It wasn't built for a medical device. And if you carry it over to your device, it quietly points you at the weakest control there is. Here's the difference that matters. A workplace hierarchy is about protecting a person from a dangerous environment around them. You take the hazard out of the room, or you put the worker in a respirator. A medical device isn't the room. It's the thing touching the patient. So ISO 14971, the standard that governs it, hands you a different hierarchy, with three tiers, in a specific order you're not allowed to shuffle. Those three tiers are inherently safe design, then protective measures, then information for safety. That last one, information for safety, is the polite name for a warning: a label, a line in the manual, a training class. It's the weakest thing on the list. Yet under a deadline, it's almost always the first thing a team reaches for, because it's the cheapest and the fastest. Hold onto that, because it turns out to be the whole game. This is Module 8 of The Operator's Guide to ISO 14971. In the last module, every one of WearPump's eleven hazards got scored, plotted on the matrix, and given a verdict. The connectivity hazard came back red, unacceptable, blocked from shipping. The motor driver fault landed in the elevated band, because it can over-deliver insulin. A cluster of yellows each owed a real reduction effort. That list is a work order. Here, we work it. I'm Dave Saunders. I've spent more than 30 years commercializing technology, almost 20 of them in medical devices and regulatory work, including several surgical robots. ## Why the order is fixed Why three tiers, and why in an order you can't rearrange? Because the tiers aren't equally reliable, and the standard knows it. Read from the top down, the order is a ranking of how much you're trusting a human being to do the right thing. At the top, you trust no one. At the bottom, you're trusting a tired clinician at two in the morning to remember one line from a manual she read once. I learned this on the other side of the medical world, in security. On connected devices and surgical robots, I watched team after team treat security as something you bolt on right before you ship. It never held. Security you design in at the architecture stage is cheap and solid; security you staple on after the design freezes is expensive, and it leaks. Safety works the same way. The protection you build into the design is the protection that holds. ## Tier 1: inherently safe design Inherently safe design means you remove the hazard, or shrink the risk, by changing the design itself. Not by warning about it. Not by alarming on it. By making the dangerous thing unable to happen, or unable to happen as badly. The risk that isn't there is the only one you never have to manage or explain to an auditor again. The standard calls this tier inherently safe design _and manufacture_, so it also covers how the thing is built, not just how it's drawn. A control at this tier doesn't depend on anyone noticing anything, reading anything, or reacting in time. It just holds, silently, every second. There's no human in the loop to get tired, distracted, or overruled by a busy shift. On WearPump, the worst hazard is the motor pushing too much insulin into the patient. The Tier 1 move isn't a warning that says _watch for over-delivery_. It's a hard limit, written into the firmware, that physically cannot command more than a set dose in a set window of time. The software can be told to hurt someone, and it refuses. There's an even purer version. The reservoir holds 3 mL and not a drop more, because the housing geometry won't accept more. Even in a total runaway, the most the device can deliver is what fits in that chamber. You didn't write a procedure against overdose; you made a large overdose mechanically impossible. The tubing connectors are keyed too, so a half-asleep user can't cross-connect them: the parts refuse to mate the wrong way. Same principle every time. Take the human decision out, and pour the safety into the shape of the thing. Tier 1 is also the hardest and most expensive to reach, especially late. Three weeks from a launch you've already announced, "change the design" is a brutal sentence to hear, which is exactly why teams talk themselves out of it and drift toward the cheap stuff. The standard is built to stop that drift. So the first question on every hazard is the same: can I design this out? ## Tier 2: protective measures When you genuinely can't design the hazard away, you add something that actively guards against it. The word that matters is _active_. A protective measure does something. It detects, blocks, stops, or forces an extra step. It isn't a note hoping you'll behave. It's a mechanism that intervenes whether you're paying attention or not. On WearPump, the cleanest example is the occlusion alarm. If the line blocks and pressure climbs, the device doesn't just beep and hope. It sounds, and it stops the pump automatically. The alarm informs the human, but the automatic stop is the real protection, because it acts even when there's no human to hear it. That pairing, detect and then act, is what makes it Tier 2 and not Tier 3. Others are layered through the device: an air-in-line detector that catches a bubble before it reaches the patient, a two-step confirmation before any large bolus so one accidental tap can't dump a dose. But protective measures sit a step below design for a reason. They can fail. An alarm needs power, a working sensor, a speaker, and a person within earshot. A confirmation step can be worn down by a user who taps through it out of habit. The protection is real, but it has moving parts and a human somewhere in the chain. ## Tier 3: information for safety Tier 3 is labels, the instructions for use, warnings, and training. It's the last resort, and the standard is blunt about that. On WearPump it looks like a line in the instructions that says _don't wear the device longer than three days without a clinician check_. It looks like a low-battery alert well before the battery dies, and a training session for the nurse setting it up the first time. All of it is information, handed to a person, in the hope they read it, remember it, and act on it. That hope is exactly why it's the weakest tier. Every other tier does the work for you. Information for safety asks the human to do the work, under stress, in the moment. A warning is only ever as strong as that clinician's worst night, and you don't get to pick her night. To be fair to Tier 3, good labeling saves people. The problem is never that warnings are useless. It's treating a warning as the _fix_ when it's really just the floor, the thing you're left with after the real work. And this is the pull that makes it go wrong: a design change is expensive and slow, an alarm is a project of its own, but a warning label is a sentence you can write this afternoon and close before the meeting is over. So when the schedule is tight, and it's always tight, gravity drags every team to the bottom. ## The rule nobody honors > You don't get to skip to a warning label if a design change was reasonably practicable. Read it slowly, because that one sentence is the entire point. If you could have designed the hazard out, or guarded it with a real protective measure, and it was practicable to do so, then a warning is not an acceptable answer. You work top down, and you have to be able to show why you stopped where you stopped. Two things follow, and teams get both wrong. First, the order is mandatory: Tier 1 before Tier 2 before Tier 3, on every hazard. Second, and this is the one people miss in the other direction, you're allowed to combine tiers. For a serious hazard you should. What you can't do is leapfrog up the cheap way and pretend the higher tiers weren't available. This is what a good auditor hunts for. They'll find a warning in your file and ask one quiet, devastating question: _did you consider designing this out first? Show me._ A real analysis, with options weighed and rejected for documented reasons, is fine. A shrug means that warning just became the most expensive sentence in your submission. The files pulled apart in a courtroom years later are full of warnings added exactly where a design change was the honest answer, and nobody wanted to pay for it that quarter. The line worth keeping: real risk control is a design problem before it's a labeling problem. The best control is the hazard you engineered out of existence. The worst is a warning you reached for because it was Tuesday and the release was Friday. ## Working WearPump, tier by tier Start with the ugliest hazard, the motor driver fault that can over-deliver insulin. Severity as high as it goes. So we don't pick a control, we stack everything practicable. Tier 1 first: the firmware dose ceiling and the 3 mL reservoir cap, neither of which depends on anyone noticing anything. Then Tier 2 on top: an occlusion sensor and an independent watchdog circuit that can stop the motor if the main processor misbehaves. Then, and only then, Tier 3: a clear note in the instructions on what a fault alert means. Design, then guard, then inform. We didn't choose between the tiers, we layered them, so each one catches the failure of the layer above it. That's defense in depth, and it starts at the top. Now the red one, connectivity. The tempting fix is pure Tier 2: encrypt the link, add authentication, done. That's real, and we keep it. But stop there and you've missed the Tier 1 question: why does a safety function depend on the phone at all? Ask it honestly and the design changes. The pump has to be safe completely on its own, running its own alarms and enforcing its own limits with the phone switched off. The app becomes a convenience, not a safety control. Two teams can ship the exact same encryption. The one that also made the pump safe standalone has a defensible file; the one that leaned on encryption as the whole answer left a serious hazard resting on a wireless link, and an auditor will find it. The yellows go faster once you see the pattern. The skin reaction at the adhesive site: switching to a silicone adhesive was a Tier 1 move, a design change, with an app reminder to rotate the site as a Tier 3 touch behind it. But one yellow hides a trap, and it's the doorway to the part everyone skips. ## Every control can create a new risk This is what separates a mature risk file from a naive one: every control you add can create a brand-new hazard of its own. A safety feature is never free. ISO 14971 has a clause for this, 7.5, and it's the one I see skipped most. Every control you introduce must be reviewed for the new hazards it brings with it. You don't get to add a control, tick the box, and walk away. You turn around and ask: what did that fix just break? Take the silenced-alarm hazard. Your instinct is to add more alarm: louder, harder to silence, a backup alert. All reasonable. All Tier 2. And every one of them can make your device more dangerous, not less. Picture where the device actually lives, a busy ward or a home with a dozen other devices all beeping. Make your alarm more aggressive, multiply it across every device in the room, and what you get isn't safety. It's noise. And there's a name for what that noise does to people. It's called alarm fatigue, and it isn't a theory. The Joint Commission, which accredits hospitals across the United States, reviewed alarm-related events over a three-and-a-half-year window. There were 98 of them. 80 ended in a patient death, not from the original problem, but from the alarm that was supposed to catch it, and got tuned out. The reason is almost mathematical: on a single hospital unit, devices can throw off hundreds of alarm signals a day, and the Joint Commission estimated that 85% to 99% of them need no action at all. So people do the only rational thing they can when nearly every alarm is a false one. They stop trusting the alarms. They turn them down, silence them, and eventually stop hearing them. So the occlusion alarm you added as a Tier 2 control, the good one, is also a contribution to that wall of noise. Push it too hard and your safety feature becomes one more alarm a clinician learns to ignore. The control didn't just have a side effect. Under the wrong conditions, the control became the hazard. And the answer to alarm fatigue isn't another warning telling nurses to take the alarms seriously. That's Tier 3 answering a problem Tier 3 created. The real answer is to design the alarm to fire only when it truly matters, so it stays credible. That's a Tier 1 move on your Tier 2 control. You go back up the hierarchy. There's no free move here. Mature risk management isn't a hunt for controls to pile on. It's a running argument about whether each control truly makes the patient safer, all things counted. ## Existence is not effectiveness You've chosen your controls and checked them for new risks. You're not done. A control on a slide is not a control on a patient. Clause 7.2 gives you not one obligation but three, and they aren't the same: implement the control, verify you implemented it, and verify it's effective. The first is obvious: you decided on a firmware dose ceiling, so somebody writes that firmware. The next two get skipped. Verifying _implementation_ means confirming the control actually reached the finished device, not just the plan. I've seen a dose limit that lived proudly in the requirements document and was never written into the shipped firmware, because everyone assumed someone else had done it. On paper the hazard was controlled; on the bench the pump would happily over-deliver. Verifying _effectiveness_ is the one with teeth. Not a document check: a bench test, a usability study, a validation with real users, proof the control genuinely reduces the risk the way you claimed. Go back to the alarm. On the bench you can verify it fires, every time. Implementation confirmed. But is it effective? Put it in a real ward with real background noise and watch whether a real nurse notices it and acts in time. That's a different test, and plenty of alarms pass the first and fail the second. Existence is not effectiveness. A control that's present but doesn't work isn't a safety feature. It's a comforting story you tell yourself, right up until the day it matters. All of it goes in the file: what you implemented, the evidence it reached the device, the evidence it works. ## The completeness check, and what's left One last step, and skipping it is how good teams still get hurt. Clause 7.6 is the completeness check: go back through every hazardous situation you ever identified and confirm each one has been addressed. Not most. Every one. It's a deliberately boring sweep, and it exists to catch the hazard that fell through the cracks between two busy engineers. Because the hazard that hurts someone is almost never the one you fought hardest over. It's the one nobody remembered to control. Put WearPump back on the matrix. The motor driver fault, once elevated, now sits behind a stack of controls, and its residual risk has dropped into a band we can live with. The connectivity hazard, once red and blocking, has moved off red, not because we encrypted it, but because the pump no longer needs the phone to keep the patient safe. The adhesive reaction is softened by the silicone. But not everything landed on green. Some hazards still carry a residual risk, smaller than before, real reduction behind it, but not zero. That leftover isn't a failure. It's the true output of risk control, and pretending it away is how files become fiction. That residual, the risk that remains after you've done everything the hierarchy asks, isn't the end of the story. It's the beginning of the next one, where you decide whether the total risk left across the whole device is acceptable given everything the device does for the patient. That's overall residual risk, and it's the next module. ## The takeaway For every hazard on that grid, the file now tells a clean, defensible story: here was the hazard, here's how we tried to design it out, here's how we guarded what we couldn't remove, here's the warning we were left with, and here's our proof it works. That story is what risk control really is. Everything else is paperwork pretending to be safety. Hold onto the order. Design it out. Then guard it. Then, and only then, warn about it. Reach for the warning first and you haven't controlled the risk, you've just documented that you noticed it. The hierarchy isn't bureaucracy. It's the difference between a device that's safe and a device that merely has a thick binder claiming it is. More on product strategy, roadmaps, and fractional CPO work at [baserealitygroup.com](https://baserealitygroup.com). I also write more broadly for founders in a newsletter called The Build, at [davesaunders.net](https://davesaunders.net). [← Module 7Risk Evaluation and the Acceptability Line: ISO 14971 Clause 6](/guides/iso-14971/risk-evaluation-acceptability) [Module 9 →Residual Risk and Benefit-Risk Analysis in ISO 14971](/guides/iso-14971/residual-risk-benefit-risk) Get the next guide ISO 14971 is the first Operator’s Guide. ISO 13485, IEC 60601, and the GHG Protocol are in the queue. Leave your email and each new guide lands in your inbox the day it ships. Nothing else, no drip sequence. Email address Send me each guide Done. You’ll get the next guide when it ships. The standard is free. So is this course. Building the product is the work. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) --- # Risk Estimation with Severity and Probability: ISO 14971 Risk Analysis Part 3 | Base Reality Group Source: https://www.baserealitygroup.com/guides/iso-14971/risk-estimation-severity-probability [← All ISO 14971 modules](/guides/iso-14971) Module 06 31:47 video + article Watch this module Module 6 of the video course. The article below covers the same ground in written form, so you can watch, read, or both. \# Your FMEA Score Is Not Your ISO 14971 Risk Someone finished running an FMEA on their device. Severity, occurrence, detection, all scored, every row filled in, a risk priority number calculated for each one and ready to drop straight into the risk management file. Then they went looking at ISO 14971's own definition of risk, and they left a comment on a training video asking a question nobody had answered: why aren't we considering the detection measures in risk evaluation? That is not a small mix-up. It is the difference between two entirely different numbers ending up in your risk management file, one of them wrong for the standard you are trying to satisfy, sitting there looking exactly as official as the right one would have. And the person asking is not alone. Failure modes and effects analysis is the technique almost everyone learns first, so almost everyone arrives at this exact fork: an RPN in one hand, an empty risk column in the other, and a guess about which number goes where. I'm Dave Saunders. I've spent more than 30 years commercializing technology, almost 20 of them in medical devices and regulatory affairs, including several surgical robots. This is Module 6 of The Operator's Guide to ISO 14971, a twelve-module series building a complete risk management file from scratch using one running device: WearPump, a wearable subcutaneous insulin pump for Type 1 diabetes. Module 5 identified eleven hazards. This module gives every one of them a number. ## Two ingredients, not three Start with the definition the standard itself uses. ISO 14971 defines risk as the combination of the probability of occurrence of harm and the severity of that harm. Two ingredients: probability and severity. That is the entire recipe. You already know how the probability piece is built. From Module 2, and again in Module 5: P1 is the probability the hazardous situation occurs, P2 is the probability that situation actually turns into harm, and P1 times P2 gives you the probability of occurrence of harm. Now put that next to what FMEA gives you. A risk priority number is severity times occurrence times detection. Three factors, not two. Occurrence is really just another way of estimating probability, but detection, how likely you are to catch the failure before it reaches anyone, has no equivalent anywhere in ISO 14971's definition of risk. So when you calculate an RPN and drop that same number straight into your hazard traceability matrix, you are smuggling a detection score into a place the standard never asked for one. The numbers can look similar. They are not measuring the same thing, and ranking your hazards by one instead of the other can change which ones look urgent enough to fix first. Here is the actual answer to that comment. Detection tells you how good your ability to catch a problem is. It says nothing about how likely the harm is, or how bad it would be once it happens. Fold detection into the risk number and a hazard you are simply bad at detecting looks artificially dangerous, while one you are excellent at catching looks artificially safe, even when the actual harm, if it occurs, is identical. Picture two hazards with the same severity and the same probability of harm, one with an obvious warning sign and one that fails silently. By RPN, the silent one looks catastrophically worse. By ISO 14971's definition they are correctly identical, because the danger to the patient never changed. None of this makes detection worthless. It belongs in risk control, later, as something you improve to reduce residual risk. Keep it in FMEA. Keep it out of the number you write in the risk column. ## Score it as if you built nothing A foundational rule trips people up constantly: you estimate this risk assuming nothing has been done to reduce it yet. No packaging, no protective circuit, no warning label, no software interlock. You score the device as if none of your engineering existed. That does not mean you assume chaos. The device is still used in its intended environment, by its intended user, the way Module 4 defined them. Only the deliberate safety engineering disappears. It feels backwards the first time. Why imagine your own safeguards away? Because if you score the risk after your controls are already baked in, you can no longer tell whether those controls are doing anything. You have hidden the exact thing you are trying to measure. This is why the standard wants two estimates for the same hazard: an initial risk, scored with nothing in place, and a residual risk, scored again once your controls exist. Compare the two and you can see what your engineering bought you. A second mistake sits right next to the first, and I see it almost as often. When a team adds a control, a warning label or an alarm, they often lower the severity score to reflect it. That is usually the wrong axis to move. A warning label does not make the harm any less severe if it occurs. What it does, if it works, is make the harm less likely, because a warned user behaves differently. Severity describes the worst reasonable outcome of the harm itself; a control that changes behavior belongs in probability. Move the wrong number and your residual risk looks better than it really is, which is exactly what an auditor is trained to notice. There is one genuine exception. If a control physically changes what harm is even possible, not just how likely someone is to encounter it, severity can legitimately move. A hard mechanical cap on the maximum dose deliverable per hour does not just make an overdose less likely; it shrinks the worst-case overdose itself. That is a real severity change. ## Where the numbers come from All of this assumes you have a number to plug in for probability. The standard's guidance gives you a hierarchy, in order of how much to trust each source. At the top: published standards, articles, and data on similar devices already in use. Somebody has probably already measured how often the failure you are worried about actually happens, and an auditor will recognize the citation. Next: your own product's field history, if you have one, from complaints, post-market surveillance, and service records. Next: testing, simulation, and analysis you run yourself, including fault tree analysis from Module 5. Last: expert judgment, pulling your team into a room and asking honestly how often this happens. If your estimate came from that last category, you are not doing anything wrong. For a genuinely new device with no track record, expert judgment is often all you have, and most first-generation devices are estimated this way. The limit is simple: the worse the potential harm, the less comfortable you should be resting on a guess alone. A catastrophic outcome deserves real data or real testing before you write the number down. A rating should also mean something specific. A probability rating of 3 on a five-point scale is not a feeling; it maps to a frequency band you defined ahead of time. A workable scale might read: 1 is under 1 in 100,000 uses, 2 is 1 in 100,000 to 1 in 10,000, 3 is 1 in 10,000 to 1 in 1,000, 4 is 1 in 1,000 to 1 in 100, and 5 is over 1 in 100. Severity gets the same treatment, described as outcomes a patient's family would recognize: negligible, minor, serious, critical, catastrophic. Five points is a good default. Fewer than three and you cannot tell risks apart; more than about ten and the distinctions get too fine to defend. Whatever you choose, the whole team uses the same scale the same way every time, agrees in advance whether probability means per use or per year, and writes down the reasoning behind every score. A rating with no reason is worthless to the next engineer and to an auditor. A rating with a one-sentence reason and a data source is a defensible estimate. The difference costs you one sentence. ## Scoring WearPump, live WearPump's matrix already had four columns from Module 5: hazard, sequence of events, hazardous situation, harm. Now it gains three more: probability, severity, and the risk rating those two combine into. On the five-by-five grid, the coloring is severity-weighted on purpose. The top-right corner, high severity and high probability together, is deep red. The bottom-left is green. A severity of 5 stays in an elevated band even at a low probability, because a catastrophic outcome, even a rare one, is never allowed to hide in green. Take the motor driver stuck on, delivering insulin continuously. Failing stuck-on is rare, a single hardware fault, so P1 is a 2. With no controls yet, the harm follows almost inevitably, so P2 is a 4; multiplied and rounded, the probability of harm is a 2. Severity is fast-onset, life-threatening hypoglycemia, a 5. Probability 2, severity 5 lands in the amber band. Now the hazard fault tree analysis found and FMEA structurally could not: an occlusion, a silenced alarm, and a sleeping patient, all at once. Multiply one wear cycle in twenty by one in ten by a third and you land near 1 in 600, the lowest probability rating, a 1. Same severity, still a 5. Probability 1, severity 5 sits one column to the left of the motor driver, in yellow instead of amber, despite being every bit as severe if it happens. Two hazards, identical worst case, very different probability, purely because one needs three ordinary things to line up and the other needs one part to fail. That is why you need both techniques feeding this matrix. The rest fill in the range. The moisture combination, extended wear plus water degrading the seal, scores probability 2, severity 3, comfortably green, a reminder that not every combination fault tree analysis finds turns out to be dramatic. The connectivity path, an OR gate where either a DIY loop command or a mid-command Bluetooth drop is enough, scores probability 3, severity 4, and lands in red. The reservoir seal, a gradual wear-out, scores 3 and 3, dead center, where most of a device's risk really lives. And the Bluetooth radio, whose frozen reading looks exactly like a working one, also scores 3 and 3, for a completely different reason: a conditional harm that depends on what a distracted person does next. That last one is exactly the hazard RPN would have ranked far too high, purely for being hard to detect. ## What the grid does, and does not, decide Once all eleven are scored and plotted, the value of a matrix like this is plain: eleven rows of dense reasoning, compressed into something you can read across a conference table in about four seconds. Anyone on the team, or in an audit, can see where the pressure is. But notice what the grid does not tell you. It does not say which of these are acceptable to ship with, or which need a control before you can move forward. A color is a description of where a risk sits. It is not yet a decision about what to do with it. That separation is deliberate, and it is one of the standard's better design choices: estimation stays objective, and evaluation, the judgment call of what your organization is willing to accept, is kept as its own step so the math never quietly gets contaminated by it. Back in Module 5 the promise was that WearPump's hazard table would reach a dozen or more situations once both techniques finished running. Eleven are scored here, with more of the device still to analyze the same way. The promise holds. If you take one thing from this, take this: your FMEA's risk priority number and your ISO 14971 risk are not rivals, and they are not the same number wearing two names. They are two different questions. Answer both, on the same hazard, and you know something neither one could tell you alone. Module 7 applies WearPump's actual acceptability criteria to each scored hazard and decides which ones require a control before this device ever reaches a patient. More on product strategy, roadmaps, and fractional CPO work at [baserealitygroup.com](https://baserealitygroup.com), and I write more broadly for founders too, in a newsletter called The Build, at [davesaunders.net](https://davesaunders.net). [← Module 5Hazard Identification with FMEA and FTA: ISO 14971 Risk Analysis Part 2](/guides/iso-14971/hazard-identification-fmea-fta) [Module 7 →Risk Evaluation and the Acceptability Line: ISO 14971 Clause 6](/guides/iso-14971/risk-evaluation-acceptability) Get the next guide ISO 14971 is the first Operator’s Guide. ISO 13485, IEC 60601, and the GHG Protocol are in the queue. Leave your email and each new guide lands in your inbox the day it ships. Nothing else, no drip sequence. Email address Send me each guide Done. You’ll get the next guide when it ships. The standard is free. So is this course. Building the product is the work. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) --- # Risk Evaluation and the Acceptability Line: ISO 14971 Clause 6 | Base Reality Group Source: https://www.baserealitygroup.com/guides/iso-14971/risk-evaluation-acceptability [← All ISO 14971 modules](/guides/iso-14971) Module 07 32:07 video + article Watch this module Module 7 of the video course. The article below covers the same ground in written form, so you can watch, read, or both. \# In ISO 14971, "Acceptable Risk" Is a Decision You Make Before You Analyze ISO 14971 defines a safe device in five words: freedom from unacceptable risk. Not freedom from risk. Freedom from unacceptable risk. Read it carefully and that one word, unacceptable, hands you the hardest question in the whole standard, because the standard never tells you what unacceptable means. It can't. What's unacceptable for a wearable insulin pump is different from what's unacceptable for a tongue depressor. Somebody has to decide where the line sits, and when they decide it turns out to matter as much as where. Back in Module 6, every one of WearPump's eleven hazards got a probability, a severity, and a place on the five-by-five matrix. Every hazard got a color. But a color doesn't tell you what to do about it. Green, yellow, and red describe where a risk sits; they aren't yet decisions. Risk evaluation, ISO 14971 Clause 6, is the step that turns a color into a verdict. Each hazard walks in with a position on the matrix and walks out with a decision: acceptable, ship it. Or tolerable, but only if you've done something specific first. Or unacceptable, and this device does not go out the door until you fix it. ## The trap: you don't draw the line now Here's the trap I've watched sink more risk files than any other. You do not get to decide where the acceptability line sits right now, looking at your results. You had to have drawn it earlier, back in the risk management plan from Module 3, before you had a single number in hand. Draw the line after you see the results, and you've built a circular argument an auditor takes apart in about a minute. The reason the order matters is human. Finish your analysis, and suppose one nasty hazard lands just inside the unacceptable zone. If you're allowed to redraw the line now, under deadline pressure, the line moves. Just a little. Just enough. And now your criteria aren't a safety judgment, they're a rationalization of whatever results you happened to get. Reviewers have seen that move a hundred times, and they look for it specifically. The single most common version, which I've watched happen at more than one young company, is setting the risk matrix after seeing the analysis. It feels efficient. It's fatal. I'm Dave Saunders. I've spent more than 30 years commercializing technology, almost 20 of them in medical devices and regulatory affairs, including several surgical robots, a good part of it in audit rooms watching these decisions get defended, or fall apart. Clause 6 itself is short, almost anticlimactically so: compare each hazard's estimated risk against the criteria defined in your plan; if it's acceptable it becomes a residual risk, and if it's not, you go to risk control. The entire weight of the clause rests on four words, defined in your plan. The clause has no opinion of its own. It just says: compare against the criteria you already wrote down. Those criteria come from a hierarchy. Top management sets a policy, a written commitment to how the company approaches acceptable risk, and that policy has to rest on real anchors: regulations, international standards, the generally acknowledged state of the art, and known stakeholder concerns. The policy drives the criteria for a specific device, written into the plan. And the criteria drive the evaluation, applied to each hazard. Policy, then criteria, then evaluation, all three consistent and traceable. The policy also has to name which reduction principle you're committing to, ALARP, ALARA, or SFAIRP, because the choice changes how hard you're obligated to push in the tolerable zone. Name it on purpose, before you need it. ## Three zones, not two The criteria carve the matrix into three zones, not two, and the third one is where evaluation stops being paperwork. On a five-by-five, it's a coloring rule: every one of the twenty-five cells is pre-assigned to a zone, so when a hazard lands on a cell, its verdict is already written. The acceptable zone, usually the green cells, needs no further reduction. You record the result and carry the risk forward as a residual risk. Even here, though, acceptable is not the same as ignore it. An empty row in a file doesn't read as "this was fine." It reads as "nobody looked." The unacceptable zone, the red cells, stops the line. You cannot ship, you go to risk control, and you re-evaluate whatever residual is left. Some companies draw only these two zones, a hard line with fix above and ship below. It's simpler, and it's quietly worse, because it lets a genuinely uncomfortable risk fall just under the line and get waved through with no obligation to reduce it. The middle zone, the yellow band, is the tolerable zone, and tolerable is not acceptable. It means the risk may be allowed to remain, but only if you have first reduced it as far as reasonably practicable. That's ALARP, and it's one of the most misunderstood ideas in the field. You keep reducing until the cost and effort of reducing further becomes grossly disproportionate to the safety you'd gain. The part people get exactly backwards: ALARP is not a synonym for "good enough." The default assumption is that more reduction is required. Stopping is the exception, and stopping requires justification. You don't prove you should keep going; you prove you were allowed to stop. And "grossly" is doing a lot of work, a deliberately high bar, because the whole point is to push manufacturers past the cheap, comfortable stopping points. There's no clean dollars-per-risk formula for it, and anyone selling you one is overselling. It's a judgment, and the answer to that discomfort isn't a fake formula, it's documentation good enough that your judgment is legible to a stranger. Four natural-sounding justifications are simply not valid ALARP arguments. It would be too expensive, with no evidence of gross disproportion. We already meet the regulatory minimum, when regulations are floors, not ceilings. Other devices have the same risk, when the state of the art can still require you to do better. And the probability is low, when severity still matters and a rare catastrophic outcome is a serious concern. ## The table an auditor actually asks for So what does count? Demonstrating ALARP comes down to four things you do and four things you write down, because ALARP that lives only in your head is worth nothing. Document every option you considered, not just the ones you used. Implement every measure whose cost isn't grossly disproportionate to the benefit, all of them. Document why you rejected the ones you didn't. And compare what's left against the state of the art. That third step, the rejections, is where the real defensibility lives; anybody can list the controls they added, but the honest record of the ones you considered and chose not to add, with reasons a stranger could evaluate, is the evidence you actually thought about it. Put the four together and you get a table. Here's a real one, for WearPump's skin reaction at the adhesive site, which scored probable on probability and moderate on severity, landing it in the yellow zone. Risk control option Cost Risk reduction Decision Switch to silicone adhesive \+ $0.80 / unit P: 4 to 3 Implemented Add app skin-check reminder \+ $15k one-time P: 3 to 2 Implemented Add real-time skin sensor \+ $45 / unit, 6 mo P: 2 to 2 (marginal) Rejected, grossly disproportionate Drop adhesive for mechanical clip redesign, 18 mo, new hazard net negative Rejected, introduces a greater risk Two implemented, two rejected with documented reasons. The residual, after the silicone and the reminder, sits at probability three, severity two: green, acceptable. That table is a demonstrated ALARP argument. Not a shrug, not "we did our best," but a specific, auditable record. Compare it to what most files contain for a yellow hazard: a single line, "risk reduced, acceptable." No options, no costs, no rejections. That's not ALARP. That's the word ALARP written next to a decision nobody can check. ## The floor that moves That fourth step, state of the art, deserves its own moment, because it quietly raises the bar on everyone. State of the art, in ISO 14971, is what is currently and generally accepted as good practice, not the most advanced thing anyone has built. And it sets the floor for what counts as acceptable. If a safer design already exists and is in common use across your field, your design has to match it or justify, in writing, why it doesn't. Which means a competitor can raise your bar: if another pump solves a hazard you're still living with, their solution can redefine what reasonably practicable means for you. Their good idea just became your new minimum. This is exactly why meeting the regulatory minimum isn't a valid argument; standards are floors, and they lag the state of the art because codification takes years. You find the state of the art in consensus standards like 60601 and 62366, in clinical practice guidelines, in the safety features standard on comparable pumps, in FDA design guidance, and in the incident databases. That last one is the most humbling: the FDA's public MAUDE record is a searchable history of exactly how pumps like yours have hurt people in the real world. If you haven't read what's already gone wrong in your category, you can't credibly claim your analysis reflects the state of the art. And because it moves, this isn't a one-time check; the 2019 standard requires you to keep monitoring it. A device built to the state of the art of ten years ago can quietly fall out of acceptability without a single thing about it changing. The world moved. The device didn't. There's one honest path out of the yellow zone for a risk you can't reduce any further: benefit-risk analysis, Clause 7.4. It asks whether the clinical benefits, for this patient, outweigh the residual risk that's left. If they do, the risk may be judged acceptable on those grounds, documented with clinical evidence. But it's not a rubber stamp and not a way around ALARP; you get there after reducing as far as reasonably practicable, not instead of it, and if the benefit doesn't outweigh the risk, the device can't ship. ## Every hazard, a verdict With the machinery in hand, WearPump's eleven hazards each get a verdict. The connectivity hazard, probability three and severity four, is red: unacceptable, blocked until controlled. The motor driver stuck on, probability two and severity five, lands in the elevated amber band, because a severity-five hazard never sits quietly; it needs control and a documented ALARP argument on the residual. The yellow cluster, the silenced-alarm combination, the reservoir seal, the Bluetooth lockup, each needs the full ALARP table we just built for the adhesive site. The acceptable ones, the moisture combination and the adhesive site after its ALARP work, get recorded and carried forward. I promised the honest part, so here it is. The yellow zone is where real companies fudge. With a tolerable-zone hazard it's tempting to write "acceptable" and move on, because technically it's in a band the plan allows, and at the end of a long project with a deadline looming the pressure is enormous. That's the move that turns a risk file from a safety document into a liability waiting to be discovered. Because a yellow hazard you waved through, with no ALARP table behind it, is fine right up until someone is harmed and a lawyer or regulator pulls the file. Then the question is simple and devastating: you called this tolerable, so show me what you did to reduce it. If the answer is a blank space, tolerable becomes negligent, in front of exactly the audience you least want to explain it to. If you take one thing from this, take this. Acceptable risk is not a discovery you make at the end, looking at your numbers. It's a line you commit to at the beginning, before the numbers exist, and then honor hazard by hazard, even when it's inconvenient. And tolerable is not a synonym for acceptable. It's a promise that you reduced the risk as far as you reasonably could, and can prove it. Module 8 takes every hazard we just marked red or amber, and every unfinished yellow, and engineers it down, using the three-tier risk control hierarchy in the order the standard requires. More on product strategy, roadmaps, and fractional CPO work at [baserealitygroup.com](https://baserealitygroup.com), and I write more broadly for founders too, in a newsletter called The Build, at [davesaunders.net](https://davesaunders.net). [← Module 6Risk Estimation with Severity and Probability: ISO 14971 Risk Analysis Part 3](/guides/iso-14971/risk-estimation-severity-probability) [Module 8 →The Risk Control Hierarchy: ISO 14971 Clause 7](/guides/iso-14971/risk-control-hierarchy) Get the next guide ISO 14971 is the first Operator’s Guide. ISO 13485, IEC 60601, and the GHG Protocol are in the queue. Leave your email and each new guide lands in your inbox the day it ships. Nothing else, no drip sequence. Email address Send me each guide Done. You’ll get the next guide when it ships. The standard is free. So is this course. Building the product is the work. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) --- # ISO 14971 Clause 4: The Risk Management Plan and System | Base Reality Group Source: https://www.baserealitygroup.com/guides/iso-14971/risk-management-plan-clause-4 [← All ISO 14971 modules](/guides/iso-14971) Module 03 42:03 video + article Watch this module Module 3 of the video course. The article below covers the same ground in written form, so you can watch, read, or both. \# Set the Line First: Why ISO 14971 Clause 4 Comes Before Any Risk Analysis Around 80,000 people were walking around with a small piece of metal inside their heart that could snap without warning. It was a heart valve, and in most respects a good one. It kept tens of thousands of patients alive who otherwise would not have been. But one part of it, the strut that held the disc in place, could fracture. And when it fractured, the valve failed in an instant, and roughly 2 out of every 3 people it happened to did not survive. This was the Björk-Shiley convexo-concave valve, and the fractures started showing up in the early 1980s. Not in a lab. In people. About 619 of those valves eventually broke that way. The company had data on the fracture risk. Changes they had made to the design, and to how fast they built the valve, had actually made that little strut weaker. And the valve stayed on the market, year after year, while the reports came in. The point of the story is not that some engineers were careless. The point is about a decision. Somewhere inside that company, someone decided the fracture risk was acceptable, that the valve was good enough to keep selling. And the thing worth noticing is who made that call, and what they made it against. As far as anyone can tell, it was made commercially, and it was made against no acceptability criteria that had been written down in advance. There was no system designed to catch that decision and hold it honest. The risk got weighed, quietly, by the wrong people, against a bar that moved to wherever it needed to be. The FDA withdrew approval of the convexo-concave valve in 1986. Pfizer, which owned it by then, eventually settled with patients and their families for somewhere around $215 million. But the lawsuits are not the lesson. People had already died. The lesson is everything that should have happened before the first valve was ever implanted, and didn't. ## The failure was the system, not the math Here is the thesis. The failure was not a missing calculation. It was a missing system. Risk management does not start the moment you open a spreadsheet and fill in your first hazard. It starts well before that, with the thing that decides who is allowed to call a risk acceptable, on what criteria, set down in advance, before anyone has seen a single result. That thing has a name in ISO 14971. It is Clause 4. Most people do not start there. The day they are told to do risk management, they open a hazard analysis template and start typing: hazard, hazardous situation, harm, a number for probability, a number for severity. The rows fill up, and it feels like progress. But if you started there, you started in the middle. You skipped the part that was supposed to tell you what those numbers even mean, and what counts as a pass. Almost every explanation of ISO 14971 you will find does the same thing, mentioning the plan in a sentence and a half on the way to the analysis table, because the table is the part that looks like the work. I'm Dave Saunders. I have spent more than 30 years in medical devices and regulatory affairs, a good share of it in audit rooms watching risk files get pulled apart. The single most reliable tell that a file is in trouble is a risk management plan whose date is later than the analysis it supposedly governed. It means the line was drawn after the shooting. ## What Clause 4 actually requires Clause 4 is the general requirements for risk management, and it is short. But it carries four things your whole risk file rests on. A risk management process, established, documented, and maintained. A risk management plan for the specific device. People who are competent to do the work, with records to prove it. And a risk management file that stays alive for the life of the product. Two distinctions clear up most of the confusion. The first is process versus plan. The process is the general way your company does risk management across every product, your standard operating procedure, the rules of the road. The plan is the specific application of that process to one device. Process is the company. Plan is the product. You need both. The second is the policy. Sitting above the process, Clause 4 asks top management to define and document a policy for establishing criteria for risk acceptability. Not the criteria themselves yet, but the policy for how the criteria get set. And notice who that lands on. Not the engineer with the spreadsheet. Top management. We will come back to why. If you ever feel like you are inventing all of this from a blank page, the companion technical report, ISO/TR 24971, exists to help with the how; the standard tells you what, the report shows you how. ## The plan, element by element The plan is where most of the real decisions of Clause 4 live, and the standard gives you a list of what it has to cover. Miss an element and that is a gap an auditor is trained to find. * **Scope.** What device, and which parts of its life. Scope has to run the whole life cycle: manufacturing, shipping, storage, setup, maintenance, and disposal. Each phase can create its own hazards. * **Responsibilities.** Who does what, and who signs off on what. This is the element that quietly prevents the heart valve story. If the person who approves a residual risk as acceptable is not the same person carrying the revenue target, you have built a small wall between the judgment and the pressure. * **Review.** When risk management gets reviewed, by whom, at which milestones. Design freeze, design transfer, before launch, then a defined cadence after that. * **Verification.** When you put a risk control in place, the plan commits, up front, to proving the control works and that it actually made it into the shipping product. * **Production and post-production.** How you will collect and review information from the field after launch. This is the loop that keeps the file alive. And the plan is not frozen once written. If the device changes, the plan changes, and that update gets recorded. ## Set the line before you measure There is one more thing the plan has to include, and it is the one everything turns on: the criteria for risk acceptability. The actual bar. And here is the rule I would tattoo on a new risk manager if they would let me. You set that bar before you measure. Before you estimate a single probability, before you know where your risks will land, you decide, with no skin in the result, what level of risk your company is willing to stand behind. Then you measure, and you live with what the measurement tells you. The reason is human, not technical. Run the analysis first, then draw the line, and you will draw it in the wrong place. Not because you are dishonest, but because the line will drift to wherever makes your device pass. Your worst risk comes in a little high, and the bar quietly moves up to meet it. That is the backfill, and it is exactly what happened to that heart valve. The risk did not get evaluated against a fixed standard. The standard bent around the risk. So what does a real criterion look like? Specific enough that two reasonable people, handed the same probability and severity, reach the same accept-or-reject decision. "Residual risks shall be acceptable as judged by the team" is a mirror; whatever the team judges becomes the definition of acceptable. "Any life-threatening harm with a probability above remote is unacceptable and must be reduced" is a line, and it does not care how inconvenient it is today. Two more constraints shape the line. The first is state of the art, which does not mean the most advanced technology that exists; it means generally accepted good practice for your kind of device, now. If comparable pumps all have an occlusion alarm, you do not get to call a missing one acceptable because your math came out low. The second is ALARP, "as low as reasonably practicable," the most abused idea in the field. Real ALARP means you drove the risk down as far as was practical, documented what you tried, and can show why stopping there was reasonable. It is a verb you prove, not a label you claim. And none of this is only an ISO preference: the FDA expects a documented process, the EU Medical Device Regulation makes it law, and a notified body reviewing your file will look at the plan, and yes, at its dates. ## The 5x5 matrix, built before the analysis How do you write that line so a team can use it on dozens of risks? You build a matrix: probability on one axis, severity on the other, every cell colored to say acceptable, tolerable but watch it, or unacceptable. Here is WearPump's, our running example, a wearable insulin pump for Type 1 diabetes. > **WearPump risk acceptability matrix** **Severity, five levels:** Negligible (no clinical effect) · Minor (the patient self-corrects) · Serious (reversible, needs intervention) · Critical (permanent or life-threatening, survived) · Catastrophic (death). Each level gets a written definition, or the team will slot the same harm differently on different days. **Probability, five levels:** Improbable · Remote · Occasional · Probable · Frequent. For a new device with no field data, these are educated judgments at first. That is allowed; hiding the reasoning behind them is not. **Three zones across the 25 cells:** Green, broadly acceptable. Amber, tolerable only if driven as low as reasonably practicable and justified. Red, unacceptable, not shippable until reduced. For WearPump, anything that can kill is red unless it is improbable, and even then it lands in amber and gets scrutinized. Anything critical is at best amber until it is remote or rarer. Walk one cell. WearPump delivering a sudden unintended dose of insulin is critical, because severe hypoglycemia can be life-threatening. Say the early probability estimate lands at remote. Critical and remote falls on the red-amber border, and the plan says that border is treated as unacceptable until proven otherwise. So before WearPump ships, that risk must be controlled down. The matrix made the decision, and nobody had to argue about it in the moment. The point is the timing. We built that entire matrix, both scales, all three zones, every boundary, without estimating a single real WearPump risk. When the analysis comes, in a later module, we are not deciding what passes. We already decided. The matrix is a contract you signed with yourself before you could be tempted, and when the pressure comes, a launch date, a competitor, a board meeting, it is the one thing in the room that does not care about any of that. ## The people, and why management owns the line A plan and a matrix are only as good as the team applying them. Risk management sits across engineering, software, medicine, human factors, manufacturing, and regulatory, and no one person holds all of it, so the competence requirement is really a requirement to assemble a team that, between them, covers the device. You want them in the room at once, not in a relay, because the worst hazards hide in the handoffs: the software assumes the hardware will catch the over-pressure, the hardware team assumes the software will, and the patient took the device off and put it in a drawer. Which answers the question I left open. Why does the standard hand acceptability to top management instead of the experts? Not because executives understand the device better; they do not. Setting the acceptable level of risk is not a technical question. It is a question of how much risk the company is willing to own, on behalf of patients, in exchange for the good the device does and the business it brings. That is a values-and-accountability decision, and it has to sit with the people who carry the legal and moral weight for the whole organization. The engineers estimate the risk. Management decides what the company will accept. Keeping those two jobs in different hands is the entire design, and it is precisely the separation the heart valve never had. ## The file that outlives the meeting The last of the four is the risk management file, and it is the most misunderstood object in the standard. People picture a thick binder assembled the week before submission. That picture is wrong, and it causes real problems. The file is the traceable record of everything the process produces, often a master index that points to the plan, the analysis, the verification records, the reviews, the clinical evaluation, and the post-market data. Pull any residual risk and you can walk the chain back to the hazard that started it. And it is alive: you maintain it every time the field teaches you something, a complaint that reveals a missed hazard, data showing a feared risk barely happens, a design change that introduces a new one. When an auditor shows up, the file is what they read. If a decision is not traceable in it, then as far as the audit is concerned, the decision was not made. ## What WearPump has now By the end of this, WearPump has the two artifacts Clause 4 exists to produce before any analysis is allowed to count. A risk management plan: scope across the full life cycle, responsibilities with sign-off held outside the commercial line, three review gates plus a quarterly cadence, verification attached to every control, and named post-market sources with named owners. And the 5x5 matrix it points to, dated and signed, sitting at the front of the file. That sets up everything that follows. The next module begins risk analysis proper, with WearPump's intended use and safety characteristics, now that there is finally a system to hold it. The order matters. We built the courtroom before we tried the case. So if you take one thing from this, take the order. Policy, from top management. Then the plan. Then the criteria and the matrix, set cold. And only then the analysis, measured against all of it. Risk management is not a math problem you solve at a table. It is a system of judgment you build before the table, so that when the hard calls come, they are made by the right people, against a bar set in advance, recorded where it cannot be quietly moved. If this was useful, Module 4 is where the analysis begins. There is more on product strategy, roadmaps, and fractional CPO work at [baserealitygroup.com](https://baserealitygroup.com), and the written version of this work goes out in a newsletter called The Build, at [davesaunders.net](https://davesaunders.net). [← Module 2Hazard, Hazardous Situation, and Harm: ISO 14971 Clause 3 Definitions](/guides/iso-14971/hazard-hazardous-situation-harm) [Module 4 →Intended Use and Device Characteristics: ISO 14971 Risk Analysis Part 1](/guides/iso-14971/intended-use-and-device-characteristics) Get the next guide ISO 14971 is the first Operator’s Guide. ISO 13485, IEC 60601, and the GHG Protocol are in the queue. Leave your email and each new guide lands in your inbox the day it ships. Nothing else, no drip sequence. Email address Send me each guide Done. You’ll get the next guide when it ships. The standard is free. So is this course. Building the product is the work. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) --- # The Risk Management Report: ISO 14971 Clause 9 | Base Reality Group Source: https://www.baserealitygroup.com/guides/iso-14971/risk-management-report [← All ISO 14971 modules](/guides/iso-14971) Module 11 30:27 video + article Watch this module Module 11 of the video course. The article below covers the same ground in written form, so you can watch, read, or both. \# Your Risk File Is Not Your Risk Management Report (ISO 14971 Clause 9) The device is done. The design is locked, manufacturing is validated, the clinical team is standing by, and somebody files the regulatory submission on a Friday afternoon with a long exhale of relief. Three months later the response comes back. Not about the clinical data. Not about the manufacturing. One line: where is your risk management report? The team stares at it, because they were sure they'd already sent it. They sent the risk file. Those aren't the same document, and that mistake just cost them four months. That's not a rare story. It plays out more than anyone in this industry likes to admit. The company assumed the risk file, the hazard analysis, the tables, the FMEA, all that work, was the report. It wasn't. The file is the work. The report is the judgment about the work, signed, and a regulator wants the judgment. And the gap is expensive out of all proportion to the fix, because a missing report doesn't get repaired with an afternoon of writing. When a regulator loses confidence in your risk documentation, they don't re-read one page. They re-open the whole file, because if the summary was missing, what else got skipped? So four months isn't the cost of writing a report. It's the cost of rebuilding trust in everything the report was supposed to stand behind. The document is cheap. The trust it signals is not. This is Module 11 of The Operator's Guide to ISO 14971. I'm Dave Saunders. I've spent more than 30 years commercializing technology, almost 20 of them in medical devices and regulatory work, including several surgical robots. Across this series we've built the WearPump risk file from nothing: eleven hazardous situations, identified, estimated, evaluated, controlled, and closed. Module 9 resolved every individual residual; Module 10 judged the whole device acceptable against the therapy it replaces. Every piece of analysis is done. There's exactly one step left before this device goes anywhere near a market: the formal review, and the report that records it. ## Clause 9: the review, then the report Clause 9 has two halves. 9.1 is the review. 9.2 is the report. The review, before you release the device, requires that a competent person or group confirms three specific things: that the risk management plan was actually carried out the way it was written; that the overall residual risk is acceptable by the criteria you set; and that appropriate methods are in place to collect and review information once the device is out in the world. That third one catches people, so hold onto it. The review is a real activity, not a signature at the end. Someone competent, or a group, works back through the whole risk process and confirms it holds together. In a small company that's a cross-functional meeting with a checklist and minutes; in a large one it's a formal design review with an agenda and a scribe. Either way it produces a record of who looked, what they checked, and what they found. That record is your evidence the review actually happened and wasn't just assumed after the fact. The third confirmation, the post-production information, trips teams up because it feels like it belongs to a different department. But field data, complaints, the whole surveillance system, those are risk management inputs, not just quality-system chores. Clause 9.1 requires you to confirm the machinery for gathering that data is already in place before you ship. Not planned. Not in draft. Operational, on the day you sign. We build that machinery in the final module, but the report has to point at it and say: it exists, it runs, here it is. Then 9.2, the report. The standard uses the word _shall_: a risk management report shall be produced. It has to be traceable back to the plan, it has to contain a clear statement about whether the overall residual risk is acceptable, and it has to be signed. It's a permanent part of the risk management file for the life of the device. ## The distinction whole submissions rise or fall on > The risk file is the evidence. The risk management report is the judgment about that evidence. You need both, and one doesn't stand in for the other. The clinical side works the same way, which makes it a useful mirror. A clinical trial produces data, piles of it. The clinical evaluation report is the document that interprets that data and draws a conclusion about benefit and risk. Nobody hands a regulator the raw trial database and calls it the evaluation. Your risk file is the raw work. Your risk management report is the document that reads all of it and concludes, in plain language, this device is ready to release, and here's why. Different document, different job, both required, every time. ## The seven required elements The report has seven required elements, and a missing element is the most common finding there is. Walk each one and put it on the pump. 1. 1\. **Scope and device identification.** The report states plainly what device it covers: the WearPump, its intended use, the patient population, the model, any variants in or out of scope. For the WearPump, that's the pump, the firmware, the companion app, the adhesive patch, the fill kit, and the cannula, and it explicitly excludes the insulin itself, which is a drug governed by a different framework. A reviewer should never have to guess what's covered. 2. 2\. **Traceability to the plan.** The report names the risk management plan from Module 3 by document number and revision, and confirms the work was done the way the plan said. If you changed the plan mid-stream, you say so, and why. 3. 3\. **Confirmation the plan was executed.** You state, formally, that every activity the plan required actually happened, and you list them, each with its output document cited by number and revision. You don't make a reviewer infer that the work is complete from the fact that some documents exist. You assert it, line by line, and back each line with a reference. 4. 4\. **The overall residual risk acceptance statement.** The single most important line in the document. It states that the overall residual risk is acceptable by your defined criteria, and points at the evidence that makes it true. (More on this sentence below, because how you write it matters more than almost anything else in the document.) 5. 5\. **A reference to the benefit-risk evaluation.** Any residual you accepted by weighing it against benefit, like the silenced-alarm situation, gets referenced here with its conclusion. You don't repeat the analysis; you point to it. 6. 6\. **Confirmation the post-production systems are in place.** The surveillance plan, the complaint-handling procedure, the follow-up plan, all active and named by document number. _Active_ is the word teams quietly fudge. It's not enough that a plan exists in a folder. The procedure is running, the plan has a named owner, the field-data review has a date on the calendar. A reviewer can ask to see the last complaint that went through the system, and the answer can't be "we haven't switched it on yet." 7. 7\. **A clear release recommendation.** The report ends with a clear call: recommended for release, or not. No neutral, no maybe. If it's conditional, the conditions are spelled out with measurable completion criteria ("recommended for release once the final biocompatibility report closes"). A conditional recommendation is perfectly legitimate as long as the conditions are specific and tickable. What isn't legitimate is a soft maybe with no exit criteria, because that leaves everyone guessing whether the gate is open. Element three, the completion confirmation, isn't prose, it's a table, and it's worth picturing because a reviewer reads it in about ten seconds. Four columns: the activity, whether the plan required it, whether it's complete, and the exact output document, by number and revision, that proves it. Row after row, every activity the plan named, every one marked done, every one pointing at its evidence. That table is often the first thing an investigator scans, because it tells them instantly whether this was a real process or a box-checking exercise dressed up after the fact. ## Who signs, and what each one certifies The standard doesn't name a job title. It says the review is done by a competent person or group, and your quality system defines what competent means. In practice, a mature risk management report is signed by a cross-functional panel, and the structure that holds up under scrutiny is four signatures with four different scopes: * **The risk management lead** signs that the process was run the way the plan said, every required activity is complete, and the file is accurate and current. * **The systems engineering lead** signs for technical accuracy: the hazard analysis and the residual estimates are sound engineering judgments. * **The clinical lead** signs that the benefit-risk picture is a real clinical picture, the benefits are evidenced, and the balance genuinely favors the patient. * **The regulatory lead** signs that the file is complete against the requirements of the markets you're entering. When an investigator picks up your report in a pre-approval inspection, the signature block is very often the first thing they read. They want to know who made these judgments, and whether those people had the competence and authority to make them. A single signature from a junior engineer is a flag that gets pulled on. A cross-functional panel, with documented training records behind each name, is a quiet sign the organization takes the decision as seriously as it deserves. Two more rules bite here. First, the date of the last required signature is the official date of the report, and it has to fall after every activity the report summarizes is complete, and before the device is released. Sign in January, run a design change in February, and your report now describes a device that no longer exists. You need a fresh review. Second, the report is a controlled document under version control. If the design changes after you sign, you don't quietly edit the old report. You open a new revision, re-run the review, and re-sign. A report that floats free of a revision number is one a reviewer has no reason to trust. ## The one sentence Element four deserves more attention than any other line, because it's the sentence that tells the regulator, the auditor, and your own management that this device is safe enough to put in front of a patient. Here's the weak version and the strong version, side by side. Weak: _The overall residual risk of the WearPump appears to be acceptable based on our review of available information._ Count the problems. "Appears to be" is a hedge; the risk either is acceptable or it isn't, and you're the one supposed to know. "Available information" tells a reviewer nothing: not what was reviewed, not whether it was complete, not which criteria you applied. That sentence wouldn't survive fifteen minutes with an FDA reviewer or fifteen seconds with a notified body assessor, because it's a shrug wearing a lab coat. Strong: _The overall residual risk of the WearPump, model WP-100, is acceptable as defined by the risk acceptability criteria in the risk management plan, based on the evidence in the hazard analysis, the risk control records, the residual risk evaluation, and the benefit-risk analysis, each cited by document and revision. Every one of the eleven hazardous situations has been addressed. One, the silenced-alarm situation, was accepted through a documented benefit-risk argument. No residual exceeds the acceptable threshold without that documented justification. The overall residual risk is acceptable._ That's a verdict a regulator can trace, line by line, back to a real document. And watch what a reviewer does with it: they don't nod, they take it apart. They pull the plan and confirm the criteria say what you claim. They pull each cited document and confirm the revision matches. They count your hazards and check every one is accounted for. A strong statement invites that, because every claim is checkable and every check comes back clean. A weak statement can't survive it, because there's nothing underneath the words to pull on. Write your acceptance statement to the strong bar, then read it back and cut every hedge until it's a sentence you'd defend out loud in a room full of people who disagree with you. ## Two different audits That report gets read by two very different reviewers, and understanding the difference lets you write one document that satisfies both. An **FDA reviewer** is checking, first, that the required content is present and the conclusions are supported. Does the report exist as its own document, not buried as "see attached"? Does it trace to the plan? Is the acceptance statement clear? And then, for a device like the WearPump, two things they lean on hard: software, and human factors. The pump runs firmware that controls insulin delivery and an app that drives the interface, both governed by the software standard, and the report has to show those software risks were analyzed and controlled, not treated as a footnote. Under that standard you classify each piece of software by how badly it can hurt someone if it fails, and the firmware that commands a dose is the most serious class there is. Human factors is the other one: the pump can be engineered perfectly and still hurt someone if a tired, frightened person misreads a screen at two in the morning and confirms the wrong dose. The regulator ties the use-error study directly to the risk file, and your report has to show that connection. A **European notified body** is doing something different: they assess the process, not just the outcome. FDA largely asks, is the content here and supported. A notified body assessor asks, do you have a working system, documented, followed, and wired into the rest of your quality system. The place they most often find a gap is the seam between your clinical evaluation and your risk management. The assessor reads your benefit-risk conclusion, the one that says the pump prevents more severe lows than injections do, and asks the obvious question: where does that clinical claim come from? If the answer is a clinical evaluation report, the follow-up is instant: and where, in this risk management report, do you cite it? If the honest answer is nowhere, that's a nonconformity, because your two most important safety documents aren't speaking to each other. Wire them together on purpose, name the evaluation by its document number, and the loop is closed. ## The pre-release walkthrough, and how it fails Run the WearPump check the way the report will. Process confirmation first: the plan listed its required activities, and every one is named, its output cited, its status confirmed complete, hazard identification through overall residual risk, plus the discipline-specific pieces, the software analysis, the use-error analysis, the biocompatibility work on the silicone adhesive. Then the residual picture, stated honestly: eleven hazardous situations, most closed as individually acceptable, the motor driver and connectivity path carrying residuals that came down hard but aren't zero, and the silenced-alarm situation accepted on a benefit-risk basis. Not a wall of green. An honest account of what was reduced, what was accepted, and why. Then the post-production confirmation, all active and assigned and referenced by document number. And finally the recommendation: the WearPump is recommended for release, one clear call, signed by the four functions, dated after the last activity closed. The report is one document in a technical file that also holds your design records, clinical evaluation, software documentation, and manufacturing information, and it's the thread that ties the safety story together and points out to all of it. A reviewer often starts right here, because a good report is a map of the whole file. Which is one reason to make it clear and honest: it's frequently the first impression your entire submission makes. These are the patterns that generate deficiency letters: 1. 1\. **The post-production section is simply missing** (the most common by far), because the team assumed the surveillance plan was a separate document that didn't need to appear here. 2. 2\. **The acceptance statement is vague or hedged** ("appears acceptable"), which reads as uncertainty about whether the device is safe. 3. 3\. **Software risk isn't integrated**, treated as a footnote to the hardware. 4. 4\. **The human factors results are never referenced.** 5. 5\. **The report is dated before the last activity was truly complete.** 6. 6\. **The signature scope is undefined**, so nobody can tell what each signer was certifying. If you fix only two things before you file, fix the first two: the missing post-production confirmation, because it's so common that finding it tells a reviewer the team stopped paying attention at the finish line, and the hedged acceptance statement, because "appears acceptable" isn't a conclusion, it's an anxiety. And here's the encouraging part: every one of these is a document problem, not a device problem. None require you to change a single thing about the pump. They're cheap to fix before you file, and brutally expensive to fix after a deficiency letter, when the clock is running and a reviewer is already skeptical. ## The recap Clause 9 is two halves: the review (confirm the plan was executed, the overall residual risk is acceptable, and the post-production machinery is live) and the report (a _shall_ document, traceable to the plan, with a clear acceptance statement, signed). The file is the evidence; the report is the judgment about it. Build the report from its seven elements, write the acceptance statement so every claim is checkable, get four functions to sign with defined scopes and a date that follows the last activity, and satisfy both the FDA content read and the notified body process read by wiring the clinical evaluation and the risk file together. Avoid the six failure patterns, especially the missing post-production section and the hedged sentence. And there's a longer reason to get this right, past your submission: the report outlives the team that wrote it. Years from now, if something goes wrong with this device in the field, an investigator or a lawyer opens this exact document and reads it cold, with nobody from the original project in the room to explain it. What they find is either a clear, dated, signed judgment that shows a serious organization made a careful call with the evidence it had, or a vague paragraph that looks like nobody was quite willing to own the decision. You're writing for that reader too. So here's the question to sit with. Think about the last risk management report you produced, reviewed, or inherited. Could you point an investigator to a single sentence, not a paragraph, one sentence, that states without qualification that the overall residual risk of that device is acceptable? If you can, what makes it strong? If you can't, what's the first thing you'd change? Module 12, the finale, builds the machinery this report kept pointing at: post-market surveillance. What you collect, how you watch it, and what happens the day the real world sends you a signal your risk file never saw. \--- If this way of closing out a device is useful to you, I write more about operating a business effectively in my newsletter, _The Build_, at [davesaunders.net](https://davesaunders.net), and the rest of my work is at [baserealitygroup.com](https://baserealitygroup.com). Subscribe so the final module in this series finds you when it lands. [← Module 10Overall Residual Risk Evaluation: ISO 14971 Clause 8](/guides/iso-14971/overall-residual-risk) [Module 12 →Post-Market Surveillance: ISO 14971 Clause 10](/guides/iso-14971/post-market-surveillance) Get the next guide ISO 14971 is the first Operator’s Guide. ISO 13485, IEC 60601, and the GHG Protocol are in the queue. Leave your email and each new guide lands in your inbox the day it ships. Nothing else, no drip sequence. Email address Send me each guide Done. You’ll get the next guide when it ships. The standard is free. So is this course. Building the product is the work. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) --- # What Is ISO 14971? Why Medical Device Risk Management Exists | Base Reality Group Source: https://www.baserealitygroup.com/guides/iso-14971/what-is-iso-14971 [← All ISO 14971 modules](/guides/iso-14971) Module 01 28:12 video + article Watch this module Module 1 of the video course. The article below covers the same ground in written form, so you can watch, read, or both. In 2018, a patient in a hospital outside Chicago received a fatal insulin overdose. The pump was FDA-cleared. The hospital had protocols and trained staff. The device had been on the market for years without incident. The investigation found that the dose-confirmation interface was ambiguous under certain lighting conditions. A decimal point was hard to read. A nurse confirmed a dose that was ten times the intended amount. The patient died. Nobody set out to harm that patient. Every person in that chain was doing their job. The failure was upstream, at a design table, probably years earlier, where someone made a decision about interface contrast without asking the right question. Not because they didn't care. Because no process forced the question. ISO 14971:2019 is the international standard that creates that process. It's the framework regulators expect you to have worked before your device reaches a patient. This is Module 1 of a 12-module course that builds a complete risk management file from scratch on a single running device: WearPump, a wearable subcutaneous insulin pump for Type 1 diabetes. ## What ISO 14971 actually is Most companies describe ISO 14971 as a compliance requirement. That framing is accurate, but it's also the reason risk management files end up as the most expensive PDFs in the building. They get written to pass an audit, not to catch a decimal point. ISO 14971:2019 is the international standard for applying risk management principles to medical devices, published by the International Organization for Standardization. It's been mandatory-by-reference in the EU Medical Device Regulation and formally expected by the FDA since its first edition in 2000. In the United States, FDA expects manufacturers to have a risk management process consistent with ISO 14971 as a condition of market clearance or approval. In the EU, compliance is required under the MDR, not just encouraged. The companies that get this right treat it as a design tool, not a compliance artifact. It forces a specific set of questions early, when design decisions are still cheap to change. Companies that work the process properly finish audits faster, catch problems earlier, and submit regulatory packages that hold up under scrutiny. Companies that treat it as paperwork tend to find out why that approach fails at the worst possible time. ## How the standard is structured ISO 14971:2019 has 10 clauses and three normative annexes. Clause 1 defines scope. Everything counts: the physical device, the software, the accessories, the drug-device interface, the packaging. If it can contribute to a patient outcome, it's in scope. Clause 3 is terms and definitions. This is where the standard becomes precise and where most people, having learned the words informally, discover they've been using them wrong. Clause 4 sets up the risk management system: your policies, your responsibilities, your verification criteria. Clauses 5 through 9 are the operational core: hazard identification, risk estimation, risk evaluation, risk control, and the review of overall residual risk. Clause 10 adds production and post-production activities: vigilance, complaint handling, post-market surveillance. Risk management doesn't end at market clearance. Everything you generate through this process, the analyses, the decisions, the evidence, goes into one document: the risk management file. That file is what an auditor reviews. It's also what you review when a patient complaint comes in and you need to know whether you anticipated that scenario. ## The four definitions your auditor will test you on Before Clause 4, before the file, before any analysis at all: four words. Get these wrong and every risk analysis that follows is built on a cracked foundation. A hazard is a potential source of harm, not the harm itself. A sharp edge on a device enclosure is a hazard. The bacteria that can grow in a catheter hub is a hazard. Electrical energy in a powered device is a hazard. A hazardous situation is the set of circumstances in which a person is exposed to a hazard. Not the hazard in the abstract, but the specific context where exposure becomes possible: a patient using the device in a bathroom while showering, a clinician who misidentifies a connector under low light, a child who accesses the device's cartridge compartment. Harm is the actual physical injury or damage to health that results. Overdose. Infection. Electrocution. These are harms. They are outcomes, not scenarios. Risk is the combination of the probability that a harm will occur and the severity of that harm if it does. ISO 14971 doesn't ask you to eliminate risk. It asks you to estimate it, control it, and evaluate what's left against a defined acceptability threshold. These four words are not synonyms. An auditor will ask you to define them in that order, and to demonstrate that your risk analysis uses them in that order. The insulin pump case at the start of this piece shows what happens when they blur together: the hazard (ambiguous interface) was present; the hazardous situation (a nurse confirming a dose in a clinical setting) was predictable; the harm (fatal overdose) was documented in human-factors literature for similar devices. The question was never asked. ## The regulator's view Regulators work with incomplete information. They can't test your device in every use scenario, and they can't audit every design decision you made. What they can evaluate is your process. Did you have a systematic method for identifying what could go wrong? Did you document your reasoning? Did you apply controls proportional to the risks you found? Did you revisit all of it when something changed? When you can answer yes to those questions with evidence, submissions move faster. When a post-market complaint lands, your response is grounded in documented prior analysis rather than reconstructed memory. The companies that invest in this process don't do it only because regulators require it. They do it because it's the only systematic way to make design decisions before a patient does the testing for you. ## What "in scope" means for your device Clause 1 defines scope. The implication most teams miss: you can't scope out the parts of your device that are inconvenient to analyze. For WearPump, scope covers three layers that interact with each other. The physical device includes the pump body, cannula, reservoir, insertion mechanism, skin adhesive, housing materials, and any component a patient or clinician touches. The software includes the dose calculation algorithm, the BLE stack connecting the device to the companion app, the app's UI logic, and the alarms and alerts. Software is a medical device under both FDA and EU MDR frameworks, and its failure modes are in scope. The drug-device interface is the third layer: WearPump delivers insulin, a high-alert medication where dosing precision is directly tied to patient outcome. The interaction between the delivery mechanism and the drug, including flow rate variability, occlusion detection, and the consequences of under- or over-delivery, is in scope. You can't run a risk analysis only on the hardware. You can't exclude the software because it "runs on the phone." You can't hand the drug-device interaction to a different team and assume it isn't your risk to manage. Clause 1 won't let you. ## Acceptable risk, not zero risk There's a sentence in ISO 14971's introduction that most risk management training skips: the goal is not to eliminate risk. The goal is to reduce risk to an acceptable level. No medical device is risk-free. Risk is not a failure of design; it's a property of intervening in biology. What ISO 14971 requires is that you think clearly about the trade. For every risk you identify, you weigh it against the clinical benefit the device provides. For WearPump, the benefit case is strong. Type 1 diabetes requires continuous insulin delivery to survive, and WearPump provides more precise, more convenient delivery than multiple daily injections. The risks don't disappear because the benefit is high; they have to be estimated, controlled, and documented. But the benefit is real, and it counts. Acceptable risk is risk that has been evaluated in the context of clinical benefit and found to be within the criteria you established before the analysis began. Not zero, not trivial, within criteria. That's the standard. ## Building the risk file before you've found a single hazard The first four entries in a risk management file have nothing to do with hazards. They exist before any hazard identification begins, because without them you have no framework for deciding what to do with the hazards when you find them. The first entry is your intended use and intended users. For WearPump: continuous subcutaneous insulin infusion for Type 1 diabetics, ages 18 and older, self-administered. This definition determines the use scenarios you're obligated to analyze and the population whose physiological variability your device has to accommodate. The second is your scope of risk management activities. What's in (hardware, software, drug-device interface) and what's out, explicitly stated, signed, and dated, not implied. The third entry is the risk management team: names, roles, expertise. A team with no clinical representation can't credibly analyze clinical risks. A team with no software expertise can't analyze software failure modes. Team composition is a commitment that your analysis will actually have the coverage it claims. The fourth entry is your risk acceptability criteria: the thresholds that define what you will accept as residual risk and what you won't. These have to be established before the hazard analysis begins. If you set them after, you've given yourself the option to move the goalposts to wherever your analysis lands. That's not risk management; it's reverse-engineering a pass. ## The WearPump risk landscape WearPump's full risk analysis will develop across all 12 modules. Three categories matter from the start. Drug delivery risks are the highest-severity category. Overdose and underdose are the failure modes with the most direct path to patient death. The decimal point in the opening case belongs here, as do occlusion detection failures and software miscalculations around dose confirmation. Software risks are substantial and consistently underestimated. The dose calculation algorithm, the BLE stack, the companion app: each has its own failure modes, from incorrect calculation to dropped connection to misleading UI state. Software failures in a drug delivery device show up regularly in FDA adverse event reports. Human factors risks reflect the reality of who uses this device. WearPump users are managing a chronic condition, often tired, often under stress, often in imperfect environments. The device has to work for someone at 3am with crashing blood glucose, not an engineer in controlled test conditions. Human factors analysis is the process by which you find the decimal points before a patient does. ## What's coming in Modules 2-12 Each module adds a layer to the WearPump risk file. Module 2 covers risk management planning. Module 3 covers intended use, intended users, and use scenarios. Module 4 is risk estimation: for each hazard you identify, you assign probability and severity. Module 5 covers risk evaluation against the acceptability criteria. Module 6 covers risk controls and the required hierarchy for applying them. Modules 7 and 8 cover residual risk and the overall residual risk review. Modules 9 through 12 close with production and post-production activities, the benefit-risk determination, and the complete file review. By the end of the course, you'll have a complete risk management file built from first principles on a real device. The framework transfers to yours. ## Action items for this week Before Module 2, three specific things are worth doing now. First, get ISO 14971:2019 and read Clause 1 and Clause 3. Not the whole standard, just those two clauses. Scope and definitions. Thirty minutes at most. Second, draft your intended use statement for your device. One sentence: what the device does, for whom, in what context. Then ask whether your current risk analysis covers the full population and the full range of use environments that sentence implies. Third, write down the names of five people who belong on your risk management team. Check whether the necessary roles are covered: clinical, engineering, software, regulatory, human factors. If one is missing, that's your first risk management action. ## Before Module 2 Think about the device you're working on right now. What is the highest-consequence thing that could go wrong, not the most likely, but the most severe? Do you have a documented process for managing that risk? Is it in your risk file, or is it in your head? If it's only in your head, that's what this course is for. \--- _Dave Saunders works with medical device companies on regulatory strategy, risk management, and the specific decisions that determine whether a device reaches the market and performs as intended. More at [baserealitygroup.com](https://baserealitygroup.com)._ [Module 2 →Hazard, Hazardous Situation, and Harm: ISO 14971 Clause 3 Definitions](/guides/iso-14971/hazard-hazardous-situation-harm) Get the next guide ISO 14971 is the first Operator’s Guide. ISO 13485, IEC 60601, and the GHG Protocol are in the queue. Leave your email and each new guide lands in your inbox the day it ships. Nothing else, no drip sequence. Email address Send me each guide Done. You’ll get the next guide when it ships. The standard is free. So is this course. Building the product is the work. [Work With Dave](/work-with-me) [Prefer a smaller first step? Book a $500 one-hour working session →](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Dave Saunders is the founder of Base Reality Group and a Fractional CPO for product companies. He was a founder and operator at Galen Robotics, where the surgical-robotics platform earned FDA De Novo authorization in 2023, and he managed a 35-patent portfolio licensed from Johns Hopkins. He wrote _Founders Who Finish_ and publishes _The Build_. [More about Dave →](/about) --- # Base Reality Group — Dave Saunders | Fractional CPO for Product Companies Source: https://www.baserealitygroup.com/ New **Plan to Market:** a six-week cohort to leave board-ready and fundable. Five seats. [See the cohort →](/plan-to-market) Fractional CPO for Product Companies # Great businesses are built _in base reality._ The fantasy says raise a billion or shut it down. Base reality says: prove people truly want what you're building, then build the business that sustains the people who build it. Sometimes that's an IPO. More often it's a company that feeds ten families and throws a great Christmas party every year. Both count. I step in as a fractional CPO and run the product function: the roadmap, the priorities, the risk discipline, and the path to first revenue. Forty-plus products over thirty-five years: desktop software, SaaS, embedded systems, network hardware, and a surgical robot through FDA De Novo clearance. I've walked this road from the inside. [Book a $500 Working Session](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) One hour on your problem, plus follow-up homework. Booked and paid up front. [Apply to Work With Dave](/contact?interest=fractional) The Fractional CPO engagement. $10K to $25K per month, six-month minimum. 40+ Products Commercialized since 1989 FDA De Novo Clearance · July 2023 $500M/yr TAOS business unit at Ascend 35 Patents Licensed from Johns Hopkins The problem ## So here's what I see a lot. A founder has cleared the hardest technical hurdle. The product works. The science is solid. The team has built something genuinely impressive. But nobody owns the product full time. The roadmap lives in the founder's head and shifts with the loudest customer or the latest board call. Engineering keeps shipping, and nobody is holding the line between what gets built next and what the business needs next: the proof points, the first paying customers, the story that carries the next round. Underneath it, the structure is thin. User needs are vague. Risk work lives in a stack of spreadsheets. If your product can hurt someone or cost someone serious money, that gap surfaces eventually, whether a regulator is watching your category or not. And it tends to pick the worst possible moment: a diligence call, a design review, a customer incident. The companies that make it aren't always the ones with the best technology. They're the ones where somebody owns the product end to end and keeps the engineering pointed at the market. That's the job I do. How the work gets structured ## Five topics inside the engagement. _One clear methodology._ Inside a six-month Fractional CPO engagement, I run the product function and we move through these five topics together, in the right sequence, at the right depth for where your product is. The work your team produces is something they own, understand, and can defend. 01 ### Product and Market Definition Back to first principles. Intended use. User needs. Indications. System boundaries. Decide what the product actually is, and let everything else flow from that decision. 02 ### Productization Roadmap Bridge the gap between prototype and shippable. Manufacturing readiness. Supplier strategy. Design-for-manufacturability. IP coverage. The path from one working unit to a product that scales. 03 ### Quality Culture and Regulatory Pathway ISO 14971-aligned risk management and design controls installed early, as discipline rather than paperwork. Regulatory strategy defined: 510(k), De Novo, PMA, or your industry's equivalent. Even in categories no regulator touches yet, the same risk discipline is what keeps a high-consequence product safe and sellable. 04 ### Verification, Validation, and Design Discipline User needs and risks translated into clear engineering requirements, traceable all the way through to your test plan. The connective tissue that survives regulatory review and investor diligence. 05 ### Go-to-Market and Investor Narrative Align the commercial story with the technical and regulatory reality. Position the product so it can be sold, funded, and scaled, not just cleared. A development story that holds up in the boardroom the way it holds up in a submission. [See How We Work Together](/work-with-me) Why this works ## The consultants who gave you a slide deck weren't wrong about the destination. They just didn't know how to get there from inside your program. ![Dave Saunders](/brand-assets/opt/dave-headshot.webp) Dave Saunders Founder, Base Reality Group. Fractional CPO. Forty-plus products commercialized since 1989. I do. Forty-plus products commercialized over thirty-five years, and most of them were software: desktop applications, SaaS, embedded systems. The original consumer Wi-Fi hotspot, designed at Bell Labs and shipped by Apple as the AirPort. The TNT access concentrator at Ascend Communications during the dial-up era. A 30-million-line operating system, TAOS, that I branded and productized into a $500M-per-year business unit before Lucent acquired Ascend for $24 billion. In 2011, surgical robotics found me. Fifteen years and counting. The deepest chapter was eight years as co-founder and CTO of Galen Robotics, taking a Johns Hopkins prototype through three hardware generations, feedback from hundreds of surgeons, two cadaver studies, and FDA De Novo clearance in July 2023 for the world's first cooperatively controlled microsurgical robotic assistant. A 35-patent portfolio licensed from Johns Hopkins, fully granted by the time we cleared. $20M Series A. I provided the core leadership that got that device across the line. My interest has always been the same. The little gears that bring technology together, and the work on the other side of the fence to get it into customers' hands. The industries change. The problem doesn't. I know what a clean design history file looks like. I know what a defensible risk framework looks like. I know what a regulator is actually looking for, and I know what a Series B diligence team will find if the structure isn't there. When you work with me, you're not getting advice from someone who has studied this. You're getting an operator who has done it. Where I've operated ![Ascend Communications](/brand-assets/ascend.svg) ![Lucent Technologies](/brand-assets/opt/lucent.webp) ![Galen Robotics](/brand-assets/opt/galen.webp) Recent client work ![AiM Medical Robotics](/brand-assets/opt/aim.webp) ![Agilis Robotics](/brand-assets/opt/agilis.webp) Trak Surgical Corosant Medical Pharyvac In partnership with ongoing research at **Johns Hopkins University Laboratory for Computational Sensory Robotics.** What leaders in the field say > “Dave has an uncanny ability to quickly ramp up forward-looking new concepts... and commercialize products. If you are looking for a leader who takes technology-driven concepts to commercialization... Dave is your guy.” ![Dr. Russell Taylor](/brand-assets/opt/Taylor.webp) Dr. Russell Taylor Director, Laboratory for Computational Sensing and Robotics, JHU > “His greatest skill is his relentless pursuit to make something work then to make it work even better... always looking at everything through a ‘Shareholders Lens’. He thrives in chaos.” ![Dr. Bob Froehlich](/brand-assets/opt/Froehlich.webp) Dr. Bob Froehlich Fmr. Vice Chairman, Deutsche Bank > “Dave brings something most advisors can’t: he’s actually done this. His experience taking a surgical robotics program through FDA clearance has helped us think more clearly about our regulatory path and product structure. That kind of firsthand perspective is hard to find.” ![Dr. Ka-Wai Kwok](/brand-assets/opt/Ka-Wai-Kwok.webp) Dr. Ka-Wai Kwok Director, Agilis Robotics > “Dave led the team successfully to de Novo clearance with FDA... \[and\] managed a patent portfolio of 35 applicants, all of which were granted.” ![Dr. Henry Brem](/brand-assets/opt/Brem.webp) Dr. Henry Brem Director, Dept of Neurosurgery, Johns Hopkins Hospital Latest from Field Notes [All posts →](/field-notes) [ October 5, 2026 Amprius Rents Its Battery Factories. Every New Plant Has to Be Qualified Again. Amprius rents its battery factories, and its new Indianapolis partner shows what that costs: every plant has to be qualified again, one cell format at a time. Read more → ](/field-notes/2026-10-05-amprius-rents-battery-factories-qualifies-every-plant) Is this for you? This is for funded founders, seed through Series B, whose product works but whose path to revenue isn't proven yet. My deepest credential is medical devices and surgical robotics, and most of what I've shipped is software. The method travels: SaaS, desktop, embedded, diagnostics, defense, climate and energy hardware, industrial robotics, physical-AI, connected devices, and the consumer hardware world I've shipped in since the original Apple AirPort. If you're pre-prototype, looking for general business advice, or hoping someone else will do the hard part for you, this isn't a fit. Two more boundaries worth naming: I'm not a regulatory consultant; the pathway is part of the work, not the whole of it. And I'm not a fundraising advisor; I don't raise money for companies or take fees on a raise. Running something smaller by design, or not funded yet? I still want you to win. That's what my book and newsletter are for. Start at [davesaunders.net](https://davesaunders.net). If you're serious about getting your product to market and you want someone who has done this in your corner, start below. Work with Dave ## Two ways to start. 01  /  Working Session ### One hour on your problem. You bring the problem. I bring forty-plus products of commercialization experience. We work through what you're facing, and you leave with clear next steps. I send homework after the call and follow up while you work through it. If a longer engagement makes sense, we'll both know. $500 One hour plus follow-up. Booked and paid in advance. [Book a $500 Working Session](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) 02  /  Fractional CPO Engagement ### Senior product leadership inside your company. Six months minimum. I join your team one to two days a week and run the product function: roadmap, priorities, product and market definition, risk discipline, and the commercial story that carries your next round. Your team learns the method as we go and owns the structure when we're done. $10K to $25K /month Six-month minimum. Scope and stage set the number. [Apply to Work With Dave](/contact?interest=fractional) Not sure which? Start with the hour. It's the fastest way for both of us to find out. --- # Plan to Market — A Six-Week Commercialization Cohort for Founders | Base Reality Group Source: https://www.baserealitygroup.com/plan-to-market Plan to Market  /  Six-Week Cohort # Most investors give you _one meeting._ Six weeks. Five founders. Live, with me. You built a real product and earned the early signals. What's ahead runs through rooms you can't afford to fumble: a lead investor, a strategic partner, a first conversation you don't get twice. This is where you build what those rooms expect: the positioning, the commercial clock, the go-to-market economics, the traction narrative for your raise, and the investor pitch assembled from all of it. Built by you, torn apart and rebuilt weekly, until it survives. [Reserve a Seat](https://buy.stripe.com/bJeeVe6wk7dl5y0gAaao802) [See what's included](#seat) A $500 deposit holds your seat and applies to the $10,000. You then have 30 days to pay the balance or arrange payment, and if that doesn't come together the $500 comes back. Stripe handles the payment securely and offers several ways to pay at checkout. The cohort at a glance Format 6 weekly live sessions Session length 90 minutes Cohort size 5 founders, max You leave with Plan + investor pitch Price $10,000 To hold a seat $500, refundable Five seats total. Starts when three fill, closes at five. Built from work at ![Galen Robotics](/brand-assets/opt/galen.webp) ![Lucent Technologies](/brand-assets/opt/lucent.webp) ![Ascend Communications](/brand-assets/ascend.svg) ![AiM Medical Robotics](/brand-assets/opt/aim.webp) ![Agilis Robotics](/brand-assets/opt/agilis.webp) Watch the overview ## Under two minutes on what you'll leave with. Your browser doesn't support embedded video. [Watch the promo here.](/brand-assets/plan-to-market-promo.mp4) The diagnosis ## Five questions your board will ask. 01 Who buys first, and who signs that check? 02 What has to be true before you can sell unit one, and what happens if that slips six months? 03 How does unit one actually get sold, and what does that cost? 04 What evidence says this is a business and not a science project? 05 Where is the 24-month commercialization plan, on one page? **If any answer is fuzzy, that is the gap this program closes.** Most founders answer the first two with confidence and go quiet by the fourth. That silence is expensive in a boardroom. It is cheap here. How it works ## One gate per week. A deliverable every week. Nothing here is homework for its own sake. Every deliverable is something your board and your raise require anyway, and the weeks stack into one integrated plan. ![Papercraft scene of a round table with five chairs and planning documents spread across it](/brand-assets/opt/papercraft/five-founders-table.webp) Five seats. One table. Everyone leaves with the plan. 01 ### Positioning and Your Wedge Who buys first, who signs the check, and the narrowest market where you are the obvious choice: sized bottom-up and set against the competition. Positioning statement + competitive read 02 ### The Commercial Clock Every constraint between you and revenue (a certification, manufacturing, a platform, a partner) mapped to timeline and cash. Your constraint map with the cash need at each milestone 03 ### The Go-to-Market Model Direct, channel, or strategic-led, priced against reality: procurement committees, budget cycles, long closes. GTM model, unit economics, market math 04 ### The Traction Narrative Translate technical milestones into evidence investors fund. The milestone map for your raise, with proof attached to every claim 05 ### The Plan-to-Market Roadmap Gates one through four assembled into one sequenced, budgeted, 12-to-24-month plan, with your team plan. Then the investor pitch skeleton, populated from everything you built. Board-page roadmap + pitch skeleton 06 ### Board-Ready, Pitch-Ready You present the plan and the pitch. The room attacks both the way a board and a lead investor will. What wobbles gets fixed live. You leave board-ready and pitch-ready The output container ## Every section of the pitch has a week behind it. Investors expect the same seven sections in every pitch. The weekly gates were built to feed them. Week five is assembly: the pitch skeleton fills from deliverables you already wrote. Week six pressure-tests it. Intro What you sell, who pays for it, and why now, in one sentence. Week 1 Team Who executes the plan, who's missing, and when the hire lands. Week 5 Opportunity What the problem costs the customer, and a market size built from arithmetic. Weeks 1 + 3 Solution The offer, the differentiation, and why it holds under pressure. Week 1 Competition The matrix, with the status quo in it, and why customers switch. Week 1 Business Model Pricing, unit economics, and the proof you already have. Weeks 3 + 4 The Ask An amount that buys named milestones and the runway to reach them. Weeks 2 + 4 + 5 ![Dave Saunders](/brand-assets/opt/dave-saunders-headshot.webp) Who runs it ## The same process I run inside companies. I'm Dave Saunders. Thirty years commercializing products as a CTO, CPO, and co-founder across telecom, cloud, and medical devices, including eight years taking a Johns Hopkins surgical-robotics spinout from prototype to market. My six-month Fractional CPO engagements run $10,000 to $25,000 a month. Plan to Market is the cohort version of the same commercial thinking, for founders who want to build the plan themselves with my hands on it weekly. 40+ products taken to market 30 yrs in the operator's seat, not the advisor's chair 5 founders per cohort, maximum. Everyone presents, nobody hides. What leaders in the field say ## The people who have watched me do it. > Dave has an uncanny ability to quickly ramp up forward-looking new concepts... and commercialize products. If you are looking for a leader who takes technology-driven concepts to commercialization... Dave is your guy. ![Dr. Russell Taylor](/brand-assets/opt/Taylor.webp) Dr. Russell Taylor Director, Laboratory for Computational Sensing and Robotics, JHU > His greatest skill is his relentless pursuit to make something work then to make it work even better... always looking at everything through a ‘Shareholders Lens’. He thrives in chaos. ![Dr. Bob Froehlich](/brand-assets/opt/Froehlich.webp) Dr. Bob Froehlich Fmr. Vice Chairman, Deutsche Bank > Dave brings something most advisors can’t: he’s actually done this. His experience taking a surgical robotics program through FDA clearance has helped us think more clearly about our regulatory path and product structure. That kind of firsthand perspective is hard to find. ![Dr. Ka-Wai Kwok](/brand-assets/opt/Ka-Wai-Kwok.webp) Dr. Ka-Wai Kwok Director, Agilis Robotics Terms ## Five seats. $10,000 each. Six weekly live sessions of 90 minutes, capped at five founders so every session works on your specific plan. Direct access to me between sessions, every deliverable reviewed weekly. All sessions recorded; you keep the recordings. Two things I ask of everyone in the first cohort: honest feedback while I'm running it, and a testimonial at the end if the program earns one. What it's worth Retaining me to build this inside your company is a Fractional CPO engagement: **$10,000 to $25,000 a month**, six-month minimum. Plan to Market runs the same commercial process, pressure-tested every week, for five founders at once. So the work you would pay roughly $25,000 for one-on-one is $10,000 here. Built privately, one-on-one ~$25,000 at my engagement rate Plan to Market, six-week cohort $10,000 five founders share the room $10,000 A $500 refundable deposit holds a seat. Balance due within 30 days. 5 seats Starts when three seats fill. Enrollment closes at five. What a seat includes * Six 90-minute live sessions * A cohort capped at five founders * Every deliverable reviewed weekly * Direct access to me between sessions * All sessions recorded, yours to keep * A board-ready plan and investor pitch you present in week six The guarantee Finish the six weeks, do the work, and if you don't leave with a board-ready commercialization plan, I keep working with you until you do. Same promise I make on every engagement, and one condition, same as all my work: you show up and do the work. [Reserve a Seat](https://buy.stripe.com/bJeeVe6wk7dl5y0gAaao802) Or take the call first: 30 minutes, and I'll tell you honestly whether this fits where you are. [Email me](mailto:dave@baserealitygroup.com?subject=Plan%20to%20Market). A $500 deposit holds your seat and applies to the $10,000. You then have 30 days to pay the balance or arrange payment, and if that doesn't come together the $500 comes back. Stripe handles the payment securely and offers several ways to pay at checkout. Questions ## The ones founders actually ask. Who is this for?⌄ Founders with a built or near-built product, pre- or early-commercial, typically raising within the next 12 months. Any sector; hard-tech and regulated products are welcome. It is not for idea-stage companies or founders who want the plan done for them; that is what the six-month engagement is for. How much time does it take?⌄ One 90-minute live session per week, plus two to three hours of work between sessions. Every deliverable is something your board and your raise require anyway. This is not homework on top of the job. It is the job, with structure. I'm in a regulated space. Does this cover regulatory strategy?⌄ Your regulatory or certification path enters the program as a constraint we map onto your roadmap, cash, and raise in week two. Deep pathway strategy stays with your regulatory advisors; I'm not selling that expertise. What you build here is the commercial plan around those constraints, which is the part most regulated founders are missing. How is this different from working with you directly?⌄ In the six-month Fractional CPO engagement I join your team and run the product function with you. In Plan to Market, you build the plan yourself with my review and pressure-testing every week, alongside four other founders doing the same. Different depth, different price, same commercial thinking. Many founders start here and some go deeper afterward. What does the $500 deposit do?⌄ It holds one of the five seats, and it applies to the $10,000. From there you have 30 days to pay the balance or arrange payment with me. If that doesn't come together inside the 30 days, you get the $500 back. Will there be a self-study version?⌄ Later, at around $2,500, without live access. In the cohort you get me live, working on your specific plan every week. Need it done with you inside the company instead? That's the [six-month Fractional CPO engagement](/work-with-me). ## The program on twelve slides. The overview deck: the five gates, the six weeks, what you walk out with, and the terms. Useful if you're taking this to a partner, a co-founder, or your board. [Download the deck](/downloads/plan-to-market-overview.pptx) PowerPoint, 85 KB. ## Five seats. Then enrollment closes. Tell me where your product is and what you're raising. If it isn't a fit, I'll say so. [Reserve a Seat](https://buy.stripe.com/bJeeVe6wk7dl5y0gAaao802) A $500 deposit holds your seat and applies to the $10,000. You then have 30 days to pay the balance or arrange payment, and if that doesn't come together the $500 comes back. Stripe handles the payment securely and offers several ways to pay at checkout. --- # Podcast Guest — Dave Saunders | Base Reality Group Source: https://www.baserealitygroup.com/podcast-guest For podcast hosts and producers # Your audience gets real stories, with _numbers, names, and outcomes._ I've spent 35 years getting new technology out of the lab and into the world: the first consumer Wi-Fi hotspot at Bell Labs, a $500M product line at Ascend Communications, and the FDA De Novo clearance at Galen Robotics. Today I run Base Reality Group as a fractional CPO for product companies. I've made every mistake I warn founders about. Some of them more than once. I wrote _Founders Who Finish_ for the founders living those mistakes now. Your audience won't get theory. They'll get what happened, what went wrong, what worked, and what to do differently in their own companies. [Invite Dave on Your Show](mailto:dave@baserealitygroup.com?subject=Podcast%20guest%20invitation) [Bios & Headshot](#bios) ![Cut-paper diorama of a broadcast microphone, headphones, and sound waves in Base Reality Group's papercraft style](/brand-assets/opt/papercraft/podcast-guest.webp) What you can expect ## I come prepared. I show up on time. I make your job easy. Podcast hosts tell me the conversation has energy. That I bring passion without performance. That I follow their lead and let the interview go where it needs to go. Twenty-plus years in Toastmasters International means I can hit a point fast when you need it, or take your audience deeper when the moment calls for it. The pacing works either way. I'm direct, but I'm not a jerk about it. I've worked with brilliant people and absolute nightmares. I've raised money from billionaires and I've bootstrapped. I've shipped products that made it into operating rooms, and I've shipped products that died before launch. That range means I can meet founders where they are. ![Founders Who Finish book](/brand-assets/opt/fwf-book-cover-3d.webp) The book ## Founders Who Finish. A field manual for entrepreneurs who are done with the startup unicorn B.S. It's written for bootstrap-first founders, stalled founders grinding without progress, and corporate refugees who left the 9-to-5 only to build a worse version of it for themselves. The book cuts through startup mythology and focuses on what actually matters: finishing what you start, knowing your numbers, and building systems that work without you. I also publish _The Build_, a monthly printed newsletter for founders building in regulated markets. [The book at davesaunders.net](https://davesaunders.net/book) Episode topics ## Nine conversations I'm ready to have. ### The Product Narrative Problem Why founders can build the thing but can't explain the thing, and how a weak product narrative quietly caps fundraising, hiring, and sales all at once. What I learned positioning a surgical robot nobody had a category for. ### Problem-Market Fit Before Product-Market Fit How I learned to validate products in an industry where mistakes literally kill people, and how those same principles work for any B2B product trying to figure out if people will actually pay. ### When a Startup Needs a CPO But Can't Hire One What product leadership actually does at the seed-to-Series-A stage, what founders get wrong when they try to cover it themselves, and how to tell if the gap is strategy, narrative, or execution. ### The Three Numbers That Keep Companies Alive Revenue, costs, runway. Why 47 KPIs on a dashboard is a symptom, not a strategy, and the story of a company that died with beautiful metrics and an empty bank account. ### From Prototype to Cleared Product The unglamorous work behind taking a Johns Hopkins lab prototype to FDA De Novo clearance, and what founders in any hard-tech space can steal from it. ### When to Walk Away vs. When to Push Through I've shut down profitable companies and pushed through situations that looked hopeless, including walking away from a billionaire investor who was destroying the company. The framework behind that decision. ### Startup Theater vs. Building Real Businesses How to tell if you're making progress or just looking busy. Includes the mahogany desk test, and why I've seen companies spend more on office furniture than customer acquisition. ### Why Most Founders Shouldn't Raise VC Money The real cost of taking someone else's money and how to build without it. This isn't anti-VC dogma. It's math and psychology about what happens when your cap table has the wrong people on it. ### From Bottleneck to Builder How founders become the constraint in their own companies and what to do about it. Includes the four-box delegation matrix and why "no one can do it as well as I can" is slowly killing your business. The record ## Where the stories come from. Career span 35+ years of product commercialization across telecom, cloud, and medical devices; **40+ products** to market Galen Robotics **Co-founder and CTO, 2016–2023.** Johns Hopkins prototype to **$20M Series A** and **FDA De Novo clearance (July 2023)** for the world's first cooperatively controlled microsurgical robotic assistant. 35-patent portfolio licensed from Johns Hopkins. Bell Labs Research manager. My group built the original consumer Wi-Fi hotspot; Apple shipped it as the AirPort. Ascend Communications Branded and productized **TAOS**, the operating system inside Ascend's access concentrator line, into a ~$500M/year business unit. Today Fractional CPO at **Base Reality Group**, serving product companies. Deepest credential: surgical robotics and advanced interventional devices. Endorsed by Dr. Russell Taylor (widely considered the father of surgical robotics), Dr. Henry Brem (former Chief of Neurosurgery, Johns Hopkins Hospital), Dr. Bob Froehlich (former Vice Chair, Deutsche Bank) Speaking 20-year member of Toastmasters International Book Author of _Founders Who Finish_, at [davesaunders.net/book](https://davesaunders.net/book) Newsletter _The Build_, monthly printed and mailed to subscribers, at [davesaunders.net/free-trial](https://davesaunders.net/free-trial) For your show notes ## Bios, headshot, and past appearances. ![Dave Saunders headshot](/brand-assets/opt/dave-saunders-headshot.webp) Right-click to save the headshot, or email me for a high-resolution version. Both bios below are yours to use as written. ### Short bio (50 words) Dave Saunders co-founded Galen Robotics, leading the company to a $20M Series A and FDA De Novo clearance for the world's first cooperatively controlled microsurgical robotic assistant. His 35-year career spans early internet infrastructure, Bell Labs, and surgical robotics. He's the author of Founders Who Finish and runs Base Reality Group, a fractional CPO practice for product companies. ### Longer bio (150 words) Dave Saunders co-founded Galen Robotics, where he led the team from Johns Hopkins prototype to a $20M Series A and FDA De Novo clearance for the world's first cooperatively controlled microsurgical robotic assistant. He managed a 35-patent portfolio licensed from Johns Hopkins, commercializing technology developed in Dr. Russell Taylor's laboratory. Dave's career spans 35 years. He helped build early internet infrastructure at Ascend Communications, worked on the first consumer Wi-Fi hotspot at Bell Labs, and pivoted to surgical robotics in 2011. Today Dave runs Base Reality Group, a fractional CPO practice for product companies, with his deepest credential in surgical robotics and advanced interventional medical devices. He's the author of Founders Who Finish and publishes The Build, a monthly printed newsletter for founders. He lives in Baltimore. Past appearances [Founders Who Finish: A Conversation with Author Dave Saunders (The Independent Minds)](https://open.spotify.com/episode/1befgbuO3gFAEvdn6L3YZc) [Building the Internet, Then the Robots: Lessons From a 40-Product Founder (Unconventional Money Moves)](https://open.spotify.com/episode/1E0QXGhk8Jte86y3NNSHIP) [From Surgical Precision to Farm Waste Power: The Innovation Playbook with Dave Saunders](https://youtu.be/X1ljE5XRrEI) [Robotic Surgery: Transforming Healthcare - Can Robots Replace Doctors?](https://youtu.be/vbvxLGnFvIk) [MD&M West Podcast Recording with Dave Saunders of Galen Robotics](https://www.youtube.com/watch?v=5Pvk0dxvqVc) [Dave Saunders, Galen Robotics (EP.110)](https://youtu.be/H_rewfAYIPc) Booking ## If your audience builds real things, let's talk. I respond to all serious podcast inquiries within 24 hours. Tell me a little about your show and audience, and I'll come back with angles that fit. [Invite Dave on Your Show](mailto:dave@baserealitygroup.com?subject=Podcast%20guest%20invitation) [Connect on LinkedIn](https://www.linkedin.com/in/nemock) [The book and newsletter live at davesaunders.net](https://davesaunders.net) --- # Privacy — what this site collects | Base Reality Group Source: https://www.baserealitygroup.com/privacy # Privacy This site counts visits. There's no account to create, no login, no ad network, and nothing here is sold to anyone. Below is every piece of software that runs when you load a page, and what each one sees. ## Before you agree to anything Two things run on every page, and neither stores anything on your device. **Vercel** hosts the site. Like any web server it logs the request: your IP address, the page you asked for, your browser's user-agent string, and the time. That log is how a host tells a reader from a bot hammering the site. **Vercel Web Analytics** counts page views in aggregate. It uses no cookies and sets no device identifier, so it can't connect one visit to the next. I can see that forty people read a Field Note, not which forty. The fonts, the stylesheet, and the images all come from this site's own server. Nothing on the page is fetched from anyone else until you press play on a video. ## Only if you accept Google Analytics stays off until you press Accept on the banner. Google's Consent Mode denies analytics storage before the Analytics script loads, so nothing is written to your device and no event is sent. Press Accept and Google Analytics starts, with IP anonymization on. It sets two cookies, `_ga` and `_ga_VYH2Y37WNV`, which last two years and tell Google Analytics you've been here before. Press Decline and it stays off. Either way your answer is saved in your browser's local storage under the key `brg-consent`. That's not a cookie. It never leaves your machine, and it exists so the banner stops asking. ## The contact form The application form on the [contact page](/contact) posts to Formspree, which emails me what you wrote: your name, your email, and your answers. Nobody else reads them. ## Video The [Operator's Guides](/guides) and some [Field Notes](/field-notes) embed YouTube players through youtube-nocookie.com, which holds YouTube's tracking cookies back until you press play. Press play and Google's privacy policy takes over from there. ## Changing your mind Clear this site's data in your browser and both the consent choice and the Google Analytics cookies go with it. The banner returns on your next visit. ## Questions Write to [dave@baserealitygroup.com](mailto:dave@baserealitygroup.com). If you want to know what this site has on you: a line in a server log, plus one number inside an aggregate report if you accepted. There's no profile to send. Last updated 11 September 2026. --- # My Values for Helping Good Clients Win — Dave Saunders | Base Reality Group Source: https://www.baserealitygroup.com/values Dave Saunders — Base Reality Group # My Values for Helping _Good Clients Win_ I have been inside enough startups to know what it feels like when the work is going well and what it feels like when it isn't. When it's going well, there's a kind of forward momentum that's hard to describe to anyone who hasn't experienced it. The team is aligned. The decisions feel defensible. You can walk into a room with investors or regulators and explain your development logic without hesitation. You know where you're going and why. When it isn't going well, there's a different feeling. A low-grade tension that doesn't go away. You're moving fast but you're not sure the foundation underneath you is solid. You're answering hard questions from your board with answers that feel slightly thinner than they should. You've built something genuinely impressive and you're starting to wonder whether the structure behind it can hold up to the scrutiny that's coming. Most founders I work with are somewhere in that second place when we first talk. They aren't failing, and they aren't in crisis. They're just aware that something is missing, and not quite sure how to fix it without slowing everything down. That awareness is a good sign. It usually means you're paying attention to the right things. ## I believe you don't need a unicorn. There's a toxic idea loose in startup culture: that your only real choice is to build some magical unicorn with a billion-dollar valuation, and if you can't manage that on the first try, you should dissolve the company and accept that you failed. I reject that culture. If you've truly got something magical and the network effect is on your side, congratulations. Ride it. But not everybody has that, and not everybody has to. You can build a great business as a solopreneur. You can build a great business with five or ten people who all get bonuses, throw a great Christmas party every year, and go home to their families with their lives intact. I love that for you. It's what I want for myself, too. A business that sustains the people who build it is a success, whether it runs a surgical robot, a software product, or a laundromat. ## I believe the best part is the moment an idea becomes a product. I love being in that moment when you've got an idea, maybe a few rough prototypes, and now you have to prove it out. You run it through the mom test. You sharpen the messaging. You put it in front of potential customers and show that this is something people truly want. Then you build the infrastructure and think hard about the box, whether it's a literal box or a metaphorical one, and what that first moment with your product will feel like for a brand-new customer. A young product is a seedling. The work is finding it good, healthy soil, the right drainage, the right amount of sun, so the thing you believe in actually grows. Helping a founder do that is the best work I know. ## I believe the thinking work cannot be outsourced. You can hire people to write documents. You can hire consultants to produce deliverables. What you cannot do is hand someone else the responsibility of truly understanding your product, your customers, your market context, and your development logic, and have that understanding live somewhere useful when the diligence team or the regulator starts asking questions. The thinking has to happen inside your team. My job is to make sure it happens correctly, in the right sequence, so the work your team produces is something they genuinely understand and can defend. That distinction matters to me. I'm not here to build something for you that you'll have to explain later. I'm here to lead your team through building something they own. ## I believe good clients deserve a guide who has actually done this. Someone who has lived inside a real program, under real investor pressure, with real consequences for getting it wrong, rather than someone who has only studied the process or advised on it from the outside. I know what a risk framework looks like when it's built correctly and what it looks like when it isn't. I know what an FDA reviewer is actually looking for and what a diligence team will find if the structure isn't there. I know those things because I've been in those rooms for fifteen years in surgical robotics, including eight years at Galen Robotics taking the world's first cooperatively controlled microsurgical robotic assistant to FDA De Novo clearance in July 2023. That experience is what you're getting access to when we work together. It's pattern recognition from having actually done this, not a methodology or a framework someone else developed. ## I believe the companies that make it are not always the ones with the best technology. This is the thing that keeps founders up at night, even if they don't say it out loud. You've built something technically impressive. You know that. Your team knows that. Your investors wrote checks because they believed it. But you also know that impressive technology and a defensible product program are not the same thing. And somewhere between the prototype and the operating room, the gap between them has to get closed. The companies that make it are the ones that close that gap deliberately, early enough to matter, with someone who knows what closing it actually looks like. That's the work I do. And it's the work I care about. ## I believe you deserve honesty more than you deserve comfort. I will tell you what I see. If the risk framework has gaps, I'll tell you. If the regulatory strategy needs rethinking, I'll tell you. If something your team has built is going to create problems at submission, I'd rather tell you now than have you discover it later. That's not a pleasant thing to hear in the moment. But it's far better than the alternative, which is finding out when the stakes are higher and the options are fewer. The founders I work best with understand that honest feedback is a form of respect. They want to know what's actually true about their program, because they know that's the only foundation worth building on. ## I believe this work is worth doing well. Hard-tech commercialization is the hardest version of the get-something-out-of-the-lab problem. Regulated products like medical devices sit at the most demanding end of it: the consequences of getting it wrong land on the patients, operators, or users who will eventually depend on what you're building, not just on the cap table. The combination of high stakes and real difficulty is what drew me to surgical robotics fifteen years ago. The seriousness carries over no matter what you're building. A five-person software company deserves the same care as a surgical robot, and I bring it to every engagement. If any of this resonates with where you are right now, I'd like to talk. The fastest way to start is a paid one-hour working session: $500 for an hour of my full attention on whatever you're facing, plus homework after the call and follow-up while you work through it. If it makes sense to work together longer after that, we'll talk about what that looks like. You can book that conversation at the link below, or if you're further along, the application for the fractional engagement is there too. Either way, I'm glad you found your way here. [Work With Me](/work-with-me) --- # Work With Me — Dave Saunders | Base Reality Group Source: https://www.baserealitygroup.com/work-with-me Work With Me # Two ways to _work with me._ One is an hour on your problem. One is six months of me inside your company, running the product function. Start with the one that fits where you are. [01  Working Session](#consultation) [02  Six-Month Engagement](#engagement) Option 01  /  One-Hour Working Session ## Sometimes you need to talk to someone who has actually done this. You bring the problem. I bring forty-plus products of commercialization experience across telecom, cloud, and medical devices, including eight years inside Galen Robotics, from Johns Hopkins prototype to FDA De Novo clearance. We work through what you're facing and you leave with a clear sense of what needs to happen next, and whether working together longer makes sense. This is not a sales call. You pay for my time and you get my full attention. After the call I send homework, and I follow up while you work through it. $500 One hour plus follow-up. Booked and paid in advance. [Book a $500 Working Session](https://book.stripe.com/14A7sM5sg69hgcE2Jkao800) Working sessions point naturally toward the six-month Fractional CPO engagement for companies ready to commit to sustained work. Option 02  /  Six-Month Fractional CPO Engagement ## The primary engagement. A six-month minimum engagement for funded founders who need senior product leadership inside the company, not advice thrown over the wall. I join your team one to two days a week, about ten hours, and run the product function: roadmap, priorities, product and market definition, risk discipline, and the commercial story that carries your next round. We move through the five topics below in the right sequence for where your product is. Your team does the work with me and learns the method as we go, drawing on forty-plus products commercialized across software, telecom, cloud, and medical devices, including the FDA De Novo clearance I helped lead at Galen Robotics. When the engagement ends, the structure belongs to them. That's the point. The full terms live on one page you can read in two minutes: [the engagement term sheet →](/fractional-cpo) ### The five topics inside the engagement. 01 #### Product and Market Definition Back to first principles. Intended use. User needs. Indications. System boundaries. Decide what the product actually is, and let everything else flow from that decision. Your team owns it. 02 #### Productization Roadmap Bridge the gap between prototype and shippable. Manufacturing readiness. Supplier strategy. Design-for-manufacturability. IP coverage. The path from one working unit to a product that scales. 03 #### Quality Culture and Regulatory Pathway ISO 14971-aligned risk management and design controls installed early, as discipline rather than paperwork. Regulatory strategy defined: 510(k), De Novo, PMA, or your industry's equivalent. For MedTech founders, strategic roadmapping that uses a simpler first filing as a predicate for richer ones to follow. 04 #### Verification, Validation, and Design Discipline User needs and risks translated into clear engineering requirements, traceable all the way through to your test plan. The connective tissue that survives regulatory review and investor diligence. 05 #### Go-to-Market and Investor Narrative Align the commercial story with the technical and regulatory reality. Position the product so it can be sold, funded, and scaled, not just cleared. A development story that holds up in the boardroom the way it holds up in a submission. The guarantee By the end of six months, your company has a clear, actionable plan to reach your next real gate. For medical devices, that means a defined path through FDA and to market. For other regulated or safety-critical hardware, the equivalent certification and market path. For everything else, first revenue and the next funding milestone. If you don't have that plan, I keep working with you at no additional cost until you do. One condition: your team shows up and does the work. $10,000 to $25,000 /month Six-month minimum commitment. Price varies by company stage and scope. [Apply to Work With Dave](/contact?interest=fractional) Is this the right fit? ## Before you apply, here's what I'm looking for. You're building a real product: software, SaaS, connected devices, medical devices, robotics, or any product where quality and reliability decide whether it sells. My deepest credential is medical devices and surgical robotics; most of what I've shipped is software. You have a working prototype. You have funding. You have a team. You're serious enough about getting to market that you're willing to invest real money in getting the structure right. I work with a small number of companies at a time, by design. The work needs full attention, and the advice only stays honest when I'm not overextended. If that lines up with where you are, apply below. [Apply Now](/contact)