Build, a London/NYC AI-driven consultancy that claims to speed up large infrastructure projects with its platform, has raised an $8.5 million seed round led by Index Ventures, as investors continue to bet that AI will move into handling physical infrastructure.
Pebblebed, Puzzle Ventures and Tiny.vc also participated, alongside angels including OpenAI CFO Sarah Friar, Blackstone CTO John Stecher and senior figures from OpenAI, Meta AI Research and Google Maps.
Founded by architect James Stirrat-Ellis and AI researcher Ben McClusky, Build is attacking one of the biggest bottlenecks in infrastructure projects, namely the fragmented, consultant-heavy process of deciding where and how major projects should get built.
The immediate infrastructure boom, ironically, AI itself, or rather the explosion in data center build-outs to service the ravenous tech industry need for compute.
“We often imagine the internet to be this invisible thing. But AI is much more compute intensive than the rest of the internet has been. To grow that compute, you need to build data centres. Data centres need energy,” Stirrat-Ellis told Pathfounders.
Build’s pitch is that the old infrastructure process is no longer fast enough for that world. The company says its AI systems can cut due diligence timelines by more than 95%, turning weeks of work into hours. Its platform analyses more than 1,600 data sources covering planning, environmental, power, political and site constraints, then produces work that is reviewed by human domain experts before being sent to customers.
Stirrat-Ellis describes the company as “a digital labour platform” rather than another SaaS tool.
“The way we work is not a bit of software,” he said. “We’re actually selling outcomes.”
Build wants to sit inside the consultancy services market, which real estate and infrastructure firms still rely heavily on, but with much of the labour automated. Right now those consultants and internal teams use tools such as Google Maps, AutoCAD and specialist real estate software.
“We work as a services firm, but behind the services firm is actually mostly software,” said Stirrat-Ellis. “That means we’re able to be very competitive in price and quality, as well as speed.”
The company says it has already worked on more than 100 infrastructure projects across 15 countries, supporting governments, Fortune 500 companies and institutional real estate groups, including Tishman Speyer. Its customers collectively manage more than $2 trillion in assets.
A government or developer looking for the best sites for data centres can send Build a brief. The company’s agents ingest the project, run the majority of the work autonomously over several hours, and then pass the output to qualified domain experts before delivery.
“That’s the same way you would interact with a services firm,” said Stirrat-Ellis. “Similar prices, comparable prices, but you’re getting the thing back at 10 times the speed.”
The wider market is large. Build is targeting the development services industry that supports data centres, energy projects, industrial sites, multifamily housing and other institutional assets. Stirrat-Ellis points to large infrastructure and real estate services firms such as CBRE and others as evidence that the category already accounts for hundreds of billions of dollars in annual spend.
But Build is also entering a politically charged market. Data centres are increasingly running into community opposition, grid constraints and energy concerns, particularly in Europe.
Stirrat-Ellis argues that automation could help developers address those issues earlier rather than after a site has effectively already been chosen.
“Often the political and community side of things is taken way down the process,” he said. “That is way too late, given the impact of these data centres on communities. What Build offers is the ability to do this stuff right at the beginning.”
Before starting Build, Stirrat-Ellis worked as a designer at Heatherwick Studio, including on Singapore’s Changi Airport Terminal 5. He later taught himself to code and moved into engineering. McClusky brings a background in multi-agent systems and reinforcement learning.
Build’s platform is model-agnostic, using frontier models from providers such as OpenAI and Anthropic, combined with its own internal tooling, ontology, geospatial capabilities and long-running agent workflows. The company says the important part is not the model layer, but the ability to turn AI work into institutional-grade output.
“Companies don’t hire us to produce AI,” said Stirrat-Ellis. “They hire us to get the results.”
Build is headquartered in New York but has roots in London, where Stirrat-Ellis says it still generates a significant part of its revenue and is now hiring. The decision to build from the US, he said, reflects where infrastructure and data centre development is moving fastest.
“The US is the most forward on building, especially building data centres and infrastructure,” he said. “It’s the place to be to really drive the company forward.”
The new funding will be used to expand Build’s engineering and infrastructure teams, accelerate R&D and grow across North America and Europe.
Its long-term ambition is to automate more of the development lifecycle, from site selection and due diligence through permitting, engineering, pre-construction and asset management. The company calls this “agentic real estate.”
Martin Mignot, partner at Index Ventures, commented in a statement that Build represents “a new generation of AI companies focused on delivering actual work rather than simply improving software workflows.”
Ophelia Cai, partner at Tiny.vc, added: “What stood out immediately about Build was the ambition of the vision and the speed of execution.
