There are not enough homes, and there are not enough people to build them. Across the US, UK, and EU, roughly 15 million homes are missing. Closing that gap the traditional way would take around 7.5 billion additional hours of construction labour, and that labour does not exist. The US is short 400,000 to 500,000 construction workers. The UK will need about a million more by 2029. The EU faces a two-million-worker gap by 2030. The workforce is ageing, fewer young people enter the trade, and tighter immigration policy across the West is making the shortage worse, not better.
So homebuilders have two options. Hire more people or change how homes get built. The first cannot scale. That leaves the second.
We invested in AUAR (Automated Architecture) because we believe they have finally cracked the second option, after decades of others failing at it. And the reason they succeed comes down to a single idea: instead of sending the building to a factory, send the factory to the building.
The thesis: Industrialising construction works when capacity can flex with demand
The dream of building homes like cars, on a production line, with the cost curve that industrialisation brings, is decades old and fundamentally sound. Industrialised production makes almost everything cheaper over time, from televisions to cars, even as inflation pushes other costs up. There is no reason housing should be the exception.
The hard part has never been the principle. It is the fit between how you produce and how demand actually behaves. Housing demand is lumpy, cyclical, and spread across a country. Any production system that assumes steady, high volume in one fixed place is fighting that reality. The most capital-intensive version of the dream, the single megafactory that ships large three-dimensional house sections across long distances, runs straight into this: it needs to run near full capacity for years to pay back the hundreds of millions it cost, and transport economics limit how far its output can travel. When the pipeline dips or the geography spreads out, the math breaks.
So the interesting question was never whether to industrialise construction. It was how to do it in a way that bends with demand instead of breaking against it. That is the question AUAR set out to answer, and their answer is what convinced us.
Change that, and the whole model opens up.
What AUAR is building
AUAR turns the old model inside out. Rather than one enormous plant that homes travel to, AUAR builds small robotic factories, which they call MicroFactories, that travel to the homes.
A MicroFactory is a compact robotic cell that fits inside two shipping containers. It arrives on site, is set up within about a week, and manufactures the timber panels (the walls, floors, and roofs) that form the structure of a house. One unit can produce the panels for roughly 200 homes a year. When a project finishes, the factory is packed up and moved to the next site.
The machine itself is deliberately built from standard, off-the-shelf industrial parts: a robotic arm, a saw, standard actuators. That is a feature, not a shortcut. It keeps each unit cheap to build, makes the supply chain resilient, and lets AUAR concentrate its own engineering on the parts that actually create an advantage. Those parts are two. First, the specialised tools at the end of the robotic arm that let a general-purpose robot work reliably with natural timber, which is inconsistent and warps in ways metal never does. Second, and more importantly, the software.
That software, called MasterBuilder, is where the real durability of the business sits. Think of it as the operating system for the whole production process. A builder feeds in their design, even as a PDF or a paper plan, and MasterBuilder simulates the entire build, accounts for local building codes and material tolerances, writes the code that drives the robot, and then runs the factory during production. It compresses what traditionally takes weeks of engineering handovers into less than a day. The hardware is the body; MasterBuilder is the brain that makes it useful. And because it learns from every deployment, each new site makes the whole system a little better. This is physical AI in the most literal sense: artificial intelligence that reasons about, and then physically acts on, the messy real world.
Why this matters: removing the reasons builders say no
The cleverest part of AUAR is not the robot. It is the business model, because it dismantles the three specific reasons builders have always refused to adopt automation.
Picture a mid-sized homebuilder with a pipeline of a few hundred homes a year. Historically, buying into factory construction meant committing millions in capital for machinery, hiring specialist engineers to run it, and taking on the risk of owning expensive hardware that could be obsolete in a few years. Three barriers: the capital, the complexity, and the risk of being stuck with a depreciating asset.
AUAR removes all three. The builder buys no machine and hires no robotics engineers. AUAR owns the factory, deploys it, operates it, and moves it on afterwards. The builder simply pays for output, per square metre of panel produced, much as they already pay for any other material. The upfront commitment converts into production credits they can draw down against specific projects, which also solves a quieter but real problem: fitting a big machine purchase into project-level budgets that were never designed to carry it.
The comparison we keep coming back to is EquipmentShare, which grew into a multi-billion-dollar construction services business by selling the use of equipment rather than the equipment itself, embedding into how contractors already work. AUAR applies exactly that logic to manufacturing capacity. Sell the output, not the machine. And once a MicroFactory is woven into a builder's live construction schedule, taking it out would mean replanning the entire programme, which makes customers naturally sticky without locking them into contracts.
Why now
Three things have come together in 2026 that were not true a few years ago.
The labour shortage has moved from a manageable pressure to a structural break: the homes simply cannot be built by hand anymore. Carbon regulation is pushing builders toward timber, as mechanisms that raise the cost of carbon-heavy concrete and steel make low-carbon materials the economically rational choice, not just the responsible one. And the technology itself has finally matured: deployable robotics and physical AI have reached the point where a small, mobile, software-governed factory actually works at a viable cost. The earlier generation failed in part because the tech was not ready. Now it is.
Why this founder team
AUAR was founded by Mollie Claypool and Gilles Retsin, who have worked together for more than fifteen years, first as professors at the Bartlett at University College London, consistently ranked among the world's best architecture schools, where they ran AUAR as a research lab before turning it into a company in 2022.
What convinced us is the division of labour between them and the pace at which they move. Mollie, as CEO, works at the seam between the company and the market: partnerships, commercial strategy, policy, and the institutional relationships that a business selling into housing needs. Gilles, as CTO and chief architect, owns the seam between the technology and the product, making sure the factory, the software, and the on-site assembly work as one coherent system. Those are precisely the two problems a DeepTech hardware company has to solve at the same time: make the technology work, and make the market adopt it.
The pace matters most. AUAR took their factory from its first version to its current third generation in 24 months. In industrial robotics, the gap between major hardware generations is normally three to five years. They did it in two, and did it by rebuilding honestly around the failures of earlier versions rather than papering over them. Around the two founders sits a deliberately assembled bench of operators, including a COO who has scaled high-growth companies before and US construction leadership who know how American homebuilders actually buy. This is a team that has already lived every hard lesson this problem has to teach.
Why we're excited
AUAR sits exactly where PT1 Ventures likes to invest: at the point where the physical, real-world economy has been stuck for decades, and where the right combination of robotics, software, and a rethought business model finally breaks it open. The case for industrialised construction was always sound. What was missing was a way to do it without the crippling capital, the logistics, and the stranded-asset risk that sank everyone who tried before.
A small factory that travels to the site, paid for by the metre, getting smarter with every home it builds. That is a genuinely new answer to one of the oldest problems in the built world.
Find out more about Scy AI over at www.auar.io.