D.R. Horton, America’s largest homebuilder by volume, has signed a multi-year commercial agreement with Toronto’s Promise Robotics under which the robot factory startup’s Factory-as-a-Service (FaaS) platform will support the production of more than 1,250 homes. The deal, announced October 13, makes D.R. Horton the first major US homebuilding enterprise to adopt Promise’s integrated robotic manufacturing system, according to The Robot Report and the company’s own announcement.
The model is simple to state and unusual for construction: the builder doesn’t buy a factory. Promise operates robot microfactories — currently being set up in Edmonton, Calgary and Phoenix — that take a builder’s existing home designs, translate them into robotic production plans, manufacture floor and wall assemblies, and coordinate delivery to site. Its floors run 45 feet, its walls 20 feet with windows and doors already in them, and it also makes stair cages for fast assembly. The sweet spot is 3-4 story townhomes. Founder and CEO Ramtin Attar says the system can produce a two-story, two-family home with basement and garage in a single day of factory work — electronics, plumbing and window installation excluded — versus the two to three weeks that framing that shell typically takes on site. That is the company’s claim, not a verified throughput figure; the D.R. Horton volumes will be the first large-sample test.
Both sides are on the record. “Homebuilding has long lacked the production infrastructure needed to scale with the speed, predictability, and efficiency of modern manufacturing,” Attar said in the announcement. “Physical AI changes that by creating a new production model that can expand capacity within and across markets.” Brad Conlon, D.R. Horton’s senior vice president for strategic growth and innovation: “Promise Robotics’ integrated platform gives us a new capability to support those goals.” Housingwire, whose coverage frames the deal as a capital-discipline bet rather than a technology story, notes the timing: builders are facing persistent labour constraints, with immigration enforcement disrupting framing crews in some US markets and completion dates slipping as a result. A builder that rents factory capacity rather than owning crews is partly hedging that risk.
As of publication, this is a company announcement plus trade coverage — Housingwire’s independent reporting and analysis is the only substantive third-party read we found; no wire-service pickup had appeared at press time.
Two things make this worth more than a passing glance at an unusually dry press release. The first is the graveyard it is walking into. Industrialised construction has burned through fortunes: Katerra raised roughly $2 billion from SoftBank before collapsing into bankruptcy in 2019, and every prefab startup since has been priced against that memory. Housingwire poses the question directly: can this generation of construction technology overcome the capital intensity and housing-market cyclicality that defeated the last one? Factory-as-a-Service is the sharpest answer we’ve seen, because it restructures the risk: Promise carries the factory capex and the utilisation problem across many builders, while D.R. Horton buys capacity the way it already buys subcontracted framing — opex, cancelable if the numbers disappoint.
The second is the shape of the bet. Where factory-built housing has historically meant one giant centralised plant feeding distant markets (Katerra’s model), Promise is building small, code-adaptive microfactories local to their regions — its software re-tunes output per market, and Attar notes Phoenix needed 50% less sheeting and different bracing than Canadian plants. Distributed capacity over one mega-project mirrors the logic that’s working in other AI-at-scale domains: a thousand robot farms beat one robot mega-factory, as Google and Fanuc’s AI-factory push shows on the manufacturing side. There’s also a data story: Attar says every home produces 300 new images and 5,000 new tactile data sources, so volume compounds the model — the Flywheel housebuilders can’t replicate onsite.
For New Zealand readers, the relevance is friction, not fashion. NZIER’s youth-employment analysis put one in four NZ teenagers out of work this spring, while the trades struggle to hold crews — a shortage-and-spare-labour mismatch that factory-built framing attacks from the supply side. A country short of builders and consenting capacity doesn’t need a robot revolution; it needs framing walls to arrive ready-made, with windows in them, on a schedule that doesn’t slip when a crew quits. If FaaS prices that as a service line rather than a demonstration plant, it stops being a robotics story and starts being a housing-supply one. The proof arrives with Phoenix’s first output in early 2027 — and with whoever signs after D.R. Horton.
Source note: story draws on The Robot Report’s reporting and interview with Promise Robotics, Housingwire’s independent analysis, and Promise Robotics’ official press release.
