Promise Robotics and D.R. Horton announced a multi-year commercial agreement on 13 October under which Horton will use Promise’s Factory-as-a-Service capabilities “to support the production of 1,250+ homes,” according to the release, which carries a Phoenix dateline. The release describes D.R. Horton as “America’s largest homebuilder by volume” and calls it the first major US homebuilder to use Promise’s production platform. Both characterisations are the release’s; The Builder’s Daily, which runs on housingwire.com, also calls Horton the nation’s largest homebuilder.
The number to hold onto is not 1,250. It is zero — the number of factories D.R. Horton has to build, staff, finance or depreciate in order to get robotic production into its pipeline, on the model the release describes.
The Model Is Capacity, Not Equipment
Factory-as-a-Service is Promise Robotics’ own registered term, and the release describes it as an integrated production platform that takes a builder’s existing home designs and turns them into a coordinated process spanning robotic manufacturing, factory operations and onsite installation. The explicit promise is that builders “can access automated production capacity without developing and operating their own factories.”
That sentence is the entire commercial argument. Offsite construction’s record, on this publication’s reading, has turned more often on the balance sheet than on the engineering: a builder who commits to a panel plant has bought a fixed cost that has to be fed through a housing cycle it does not control. Promise’s structure leaves the plant on Promise’s side of the line and sells Horton throughput.
The AI claim is specific about where the software sits in that chain. Per the release, Promise’s systems interpret existing home designs, generate production plans and robotic instructions, orchestrate manufacturing, and connect that production with site installation — combined, it says, in a single environment rather than automating isolated tasks. Note what that excludes: nothing here designs a house. The input is the builder’s own drawings.
Ramtin Attar, chief executive and co-founder of Promise Robotics, puts the capacity argument directly: “Physical AI changes that by creating a new production model that can expand capacity within and across markets,” adding that “Our goal is to make that capability accessible to builders across North America.”
Brad Conlon, Senior Vice President of Strategic Growth and Innovation at D.R. Horton, gives the buyer’s version: the company looks for innovations that improve cycle times, production capacity and consistency, and “Promise Robotics’ integrated platform gives us a new capability to support those goals.”
What Was Not Disclosed
The release gives one figure of substance — 1,250+ homes. It does not disclose the value of the agreement, its duration beyond “multi-year,” a utilisation commitment, a per-home price, or where the homes will be produced. The Builder’s Daily reports that neither the announcement nor its interview disclosed the agreement’s dollar value, factory capital costs or utilisation thresholds.
Two things fill in around that silence. The first is the dateline: the announcement went out from Phoenix, and Promise’s own site lists Calgary, Edmonton and Phoenix as the three cities where it is producing homes at scale. The release itself names no facility, so the location of Horton’s production is not established by either document.
The second is framing, and it comes from Promise rather than from the outlet carrying it. Attar told The Builder’s Daily: “We call that anchor volume, and then we say every six months as the factory increases its throughput, we’re going to offer that to you.” On his account the headline number is the floor of a ramp rather than a fixed-quantity supply contract — a characterisation the company makes in an interview and does not make in the release.
The Operational Numbers Come From Promise
Everything quantitative about performance in this story is the vendor’s, and should be read that way.
The Builder’s Daily reports Attar’s claim that the system can assemble the framing for a typical two-storey, 1,800-square-foot house with a basement and attached garage in about eight hours, with more complex duplex projects taking two to two-and-a-half days. The outlet attaches its own caveat, which is worth carrying over: those figures describe assembling structural components, not completing a whole home, and they do not establish the total cycle saving D.R. Horton will realise, which depends on how the process meshes with the trades and inspections that follow. Promise’s own site carries three headline figures: +$300 in savings per home per day through faster capital turnover, +60% faster home construction, and 2x homes built with the same labour. None is independently verified in the sources reviewed for this piece, and the site does not state the baseline any of them is measured against.
The same coverage records a constraint the vendor does not control: permitting and inspection rules limit how much of a house can legally be completed in a factory, and Promise hopes to get more factory processes certified over time. That is a real governor on offsite construction in the United States, and robotics does not move it.
Two financial details in the trade account are worth noting because start-ups in this category rarely volunteer them. Attar says Promise is paid frequently as projects progress, limiting its working-capital exposure, and that “now every project that Promise produces, we actually are margin positive.” He also describes a relatively concentrated investor base including AI investment funds and pension capital with longer horizons, with conventional lending intended once individual factories prove their economics — “it can be self-funded on its own,” as he puts it. Those are company statements, not audited figures, and the outlet notes that project-level margins do not establish companywide profitability.
Promise’s Funding, for Scale
Promise Robotics raised a $15 million Series A announced on 31 October 2023, led by Horizons Ventures with Radical Ventures, the Public Sector Pension Investment Board, the United Brotherhood of Carpenters, Relay Ventures and Alate Partners participating. Its own release put total funding raised to date at $25 million, a figure The Robot Report carried in its write-up two days later. The company was co-founded by Attar and Reza Nasseri, listed on Promise’s site as Chief Innovation Officer; The Builder’s Daily reports that Attar spent 13 years at Autodesk before co-founding Promise. That Series A release is datelined Toronto, and Promise’s newsroom footer gives its locations as Toronto, Calgary and Edmonton.
No updated funding figure accompanied this agreement, and the sources reviewed for this piece give no current total. A company with $25 million disclosed in 2023 signing a production agreement with the builder the release calls America’s largest by volume is an unusual ratio, and it is the ratio that makes the financing structure — factories funded as assets rather than from venture equity — the thing to watch.
Why This Is the More Interesting of Two Robotic-Homebuilding Deals Announced the Same Day
Analysis, labelled as this publication’s own. ICON’s $200 million Series D, announced on 13 October, is the larger number and the louder story. It funds, among other things, manufacturing Titan print systems in Austin so that other builders can deploy them. Promise is selling the opposite thing: not a machine a builder operates, but output a builder orders.
Both models answer the same objection — homebuilders do not want to become manufacturers — and they answer it in incompatible ways. One hands the builder a capital asset and the operating risk that comes with it. The other keeps both and charges for throughput, which is the same structure that has made capacity-priced robotics contracts legible to specialty trades: the customer pays for work completed, and a stopped machine is the vendor’s problem.
There is also a relationship here that predates this announcement, and it is on the record in this newsroom. D.R. Horton is an investor in construction software as well as a buyer of it: it appears on the cap table of Scaffold’s $15 million seed round alongside Pulte Homes and Builders FirstSource, where its president and chief executive Paul Romanowski said that “At our volume, small amounts of friction between our systems and our trade partners’ systems compound into real cost.” A builder that already treats production friction as a line item is a more plausible early adopter of robotic capacity than one being sold on novelty — and at Horton’s volume, 1,250 homes is a test, not a transformation.
