In December 2019, Katerra shut its 200,000-square-foot factory in Phoenix and laid off 200 people. The plant had built whole walls, with the windows, insulation, wiring and plumbing already inside. Early in 2027, Promise Robotics plans to start building wall and floor framing in Phoenix, in a 16,000-square-foot building, about a twelfth the size of Katerra's. D.R. Horton, the largest U.S. homebuilder by volume, will be its first customer there. On Tuesday the two companies announced a multi-year agreement for Promise to supply framing for more than 1,250 Horton homes, most of them single-story houses around Phoenix. Promise had a buyer for the building's output before the building existed.
What Katerra paid for
Katerra raised more than $2 billion, most of it from SoftBank, and reached a valuation above $4 billion in 2019, Construction Dive reported after its collapse. It meant to own construction from design to finished building, and in 2018 it bought two architecture firms and a general contractor. Its factories needed builders and developers to send them work, and many kept using the engineers, suppliers and subcontractors they already trusted. Katerra filed for Chapter 11 in June 2021.
Katerra paid for its factories before buyers had committed to fill them. Construction Dive's sources also blamed its attempt to do everything at once, but the factory mistake is the one a hardware founder is likeliest to repeat, because new capacity looks like progress on a board slide long before anyone has agreed to buy what it makes.
How Promise sells the capacity first
Ramtin Attar, Promise's CEO, spent 13 years at Autodesk working on industrialized construction before co-founding the company in 2021, and he told HousingWire he designed it with Katerra's failure in mind. Promise studies a builder's product line and prices the work against local supply before it agrees to a schedule. The committed piece is what Attar calls anchor volume. As a factory's throughput climbs, Promise offers that builder more capacity every six months, and the same plant can take work from other builders. HousingWire describes a narrow first product, little inventory, payment as projects progress, and factories meant to pay for themselves without waiting for enormous volume. Every project Promise produces carries a positive margin, according to Attar.
The narrow product is framing: walls, floors and stairs, built from the builder's existing drawings and set on site by trained crews. Attar told HousingWire the system can frame a two-story, 1,800-square-foot house with a basement and garage in about eight hours. Windows, plumbing and inspections still run on the builder's schedule, so the gain on a finished home is smaller and nobody has measured it yet.
One plant serving several builders already works in Calgary and Edmonton, Promise's first two factories. BetaKit reports more than 200 homes finished for builders including Brookfield, Hopewell and Mattamy, and just over 2,500 under contract across several builders.
What the announcement left out
Neither the announcement nor Attar's interview disclosed the contract's value, Promise's factory capital costs, or the utilization a plant needs to cover them, HousingWire notes. A positive margin on each project says nothing yet about the company as a whole. Attar says Promise can now take a factory's economics to a bank and fund the next plant with debt instead of another venture round. Utilization is the figure that bank would underwrite, and committed anchor volume is the first thing that figure rests on.
When a hardware team shows me its first production line, I ask who has committed, in writing, to the first year of that line's output. Surgical robotics taught me why: hospitals stopped paying for science experiments, and a first customer would not buy until the payback math worked. Brad Conlon, D.R. Horton's senior vice president for strategic growth and innovation, listed what his company wants from Promise's platform: shorter cycle times, more production capacity and more predictable schedules. Promise has to deliver those for the anchor volume to grow every six months.
Dave's take
I would copy Promise's sequence in any hardware business, whether it makes houses, battery packs or surgical robots. Whether Promise's own business works is still open, and I would wait for one of its plants to run through a slow housing quarter, and for the company to publish that plant's utilization, before calling it.
From Dave’s video library
This week's Founder Tip Tuesday: how Vercel had its best sales rep write down the rules for inbound before an AI agent took the work over, and how to sort your own repeated decisions into rules and judgment.
I’m here to help you scale.
Work With DavePrefer a smaller first step? Book a $500 one-hour working session →
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 →