Build raises $8.5M to automate data-center due diligence

The AI boom is running out of places to live. Before a single wall goes up for a data centre, power line, or factory, months of paperwork must be completed. A British-founded startup now claims its AI can cut that due-diligence grind by 95 percent.

That startup is Build, which announced on Tuesday that it has raised $8.5mn (€7.4mn) in seed funding. The round was led by Index Ventures, with participation from Pebblebed, Puzzle Ventures, and Tiny.vc. Notable angel investors include OpenAI’s chief financial officer Sarah Friar and Blackstone’s chief technology officer John Stecher, Build confirmed to Tech Funding News.

Build operates at the unglamorous front end of construction. Before any data centre or power plant gets built, consultants spend weeks on site sourcing, technical due diligence, power assessments, and environmental checks. Build automates that entire process. It pulls from more than 1,600 data sources and assesses planning, power, and political risks in parallel rather than sequentially. The company says this approach cuts due-diligence timelines by more than 95 percent, turning what once took weeks into hours.

Instead of selling software, Build sells the work itself. The twist lies in its business model. While most AI startups push software licences, Build offers its services on a monthly retainer, positioning itself as a replacement for consultants rather than a tool for them.

“We achieve software margins by automating much of our work, but our focus is on services spend, not software spend,” co-founder and chief executive James Stirrat-Ellis told Tech Funding News. A human still reviews each assessment before delivery, ensuring quality.

This model is gaining traction in property tech. Paris-based Davis follows a similar path, handing clients completed feasibility studies instead of a subscription. Build’s raise is part of a broader surge. According to EU-Startups, roughly €112mn flowed into built-environment and construction AI across 2026.

Timing is everything. The scramble to build AI capacity has turned land, power, and permits into the industry’s tightest bottleneck. It is a multi-trillion-dollar infrastructure problem, and demand is outstripping what traditional developers can deliver. Hyperscalers are expected to pour close to $700bn into data centres by 2026.

New sites are rising fast, from Europe’s largest buildouts to Blackstone’s first AI-era data-centre REIT. Every one of them requires the early groundwork Build sells. That has attracted serious customers quickly. Build has run more than 100 projects across 15 countries, for clients including property giant Tishman Speyer. It recently landed its first hyperscaler, which Stirrat-Ellis calls “the top customer you can get” in data-centre work. The firm remains small, with around 10 staff across four continents, and plans to hire more.

The idea came from personal frustration. Stirrat-Ellis trained as an architect and worked on the S$13bn expansion of Singapore’s Changi Airport. There, he watched lengthy consultant reports and scattered data drag projects out. He left a master’s programme at Harvard, taught himself to code, and founded Build in 2024 with AI researcher Ben McClusky.

Index Ventures, fresh from deals such as its $60mn bet on Conduct, moved quickly. “The round came to us. We got a term sheet within a day,” Stirrat-Ellis said.

Whether Build can hold its lead is the open question. Big consultancies such as CBRE, Arup, and JLL are adding AI of their own. Build’s wager is that replacing them beats helping them, and that the firm which maps the ground fastest wins the race to build.