A £32m cheque from American investors isn't something London AI startups collect every day. But Fyld, a four-year-old BCG spinout, has just landed precisely that—led by New York's Energy Impact Partners—to bankroll an aggressive US expansion. The pitch? Mobile-first AI that promises to cut worksite injuries by nearly half whilst keeping sprawling infrastructure projects from spiralling off schedule.
The timing is hardly coincidental. America's construction and infrastructure sectors are drowning in federal dollars from the CHIPS Act and Inflation Reduction Act, yet struggling with the same chronic problems that have plagued them for decades: inadequate safety oversight, paper-based compliance systems, and a workforce stretched dangerously thin across massive project sites. When you're managing 40,000 field staff across multiple locations—as some of Fyld's clients now do—the old clipboard-and-Excel approach doesn't just fail. It kills people.
What's interesting here is how a sector notorious for resisting digital transformation is suddenly opening its chequebook. Construction has historically treated technology adoption with the enthusiasm of a site manager facing another health and safety audit. Yet Fyld has already secured contracts with some of America's heaviest hitters: Kiewit Corporation, Quanta Services, Emery Sapp & Sons, and Sulzer have all signed on, joining Spanish infrastructure giant Ferrovial in the client roster.
The unsexy problem with a brutal ROI
The platform itself is refreshingly straightforward. Field workers use a mobile app to analyse sites in real time, flagging potential safety and quality risks before someone gets hurt or a structural defect gets buried under concrete. Site managers receive live oversight across multiple locations without needing to hire additional supervisors—a significant cost saving when labour is scarce and expensive.
According to the company, some clients have seen worksite injuries drop by close to 50%. That figure warrants careful scrutiny—Fyld hasn't disclosed which types of injuries, the measurement timeframe, or how many clients achieved these results. But even if the actual reduction proves more modest across a broader dataset, the business case remains compelling. Construction accounts for roughly 20% of workplace fatalities in the UK despite representing just 5% of the workforce, according to Health and Safety Executive figures. The US picture looks similarly grim.
For an industry where project delays routinely run into millions in cost overruns and where a single serious injury can trigger regulatory investigations, insurance premium hikes, and reputational damage, the value proposition is straightforward. Prevention is cheaper than crisis management. Always has been.
Rémi Said, general partner at Partech's Growth Impact Fund (which joined the round), framed the investment in terms that will resonate with anyone tracking enterprise software: "Fyld is tackling large-scale infrastructure projects, a category that has been underserved by technology for decades. Its strong ROI for customers, AI-powered product and robust go-to-market engine give us confidence that Fyld can emerge as a global leader."
Why mobile-first matters for muck and steel
The shift to mobile-first AI is breaking through where previous attempts at construction tech have foundered. Desktop software required workers to return to site offices or portacabins to log information—inevitably meaning they didn't bother, or logged it hours later when details had faded. Wearable devices faced resistance from workers who (not unreasonably) balked at constant surveillance.
A phone app sits in a sweet spot: familiar, pocketable, and capable of capturing information at the moment it matters. Workers can photograph a potential hazard, flag it through the system, and move on. The AI analyses the image, categorises the risk, and routes it to the appropriate manager. Simple enough to actually get used in muddy boots and high-vis jackets.
The US infrastructure boom has created an unusually receptive market. Firms that won massive federal contracts suddenly face accountability requirements they've never dealt with at this scale. Demonstrate safety compliance or watch the funding evaporate. Prove you can deliver on time and budget, or face penalties. Technology adoption stops looking like a nice-to-have and starts looking like survival.
The British AI export question
Fyld's expansion represents a test case for a question that's been nagging at UK tech policy for years: can British AI companies build global champions, or will they inevitably get acquired by American competitors before reaching scale? The company is targeting 40% of revenue from the US by end of 2026—an ambitious but measurable benchmark. Deliver on that, and Fyld joins a relatively short list of UK tech companies that cracked America without selling out first.
The company's chief futurist Karl Simons recently attended a closed-door Westminster meeting where AI startup founders pushed the government to follow through on its AI action plan. That tells you something about the sector's current mood: cautiously optimistic about UK capabilities, but increasingly impatient with implementation gaps between policy announcements and actual support.
Whether Fyld can maintain independence whilst scaling into the notoriously competitive US market will depend partly on execution, partly on whether American giants decide they'd rather acquire than compete. But the client roster suggests they're solving a real problem with genuine urgency behind it. When construction firms with 40,000 workers are willing to rip out legacy systems for your platform, you've moved beyond pilot programmes into mission-critical infrastructure.
The next 18 months will reveal whether this £32m opens the door to a substantially larger Series C—likely from US funds—or whether one of the major construction technology players decides Fyld is worth buying before it gets too expensive. Either outcome would validate the underlying thesis: that health and safety compliance, for all its unglamorous reputation, might be exactly the kind of high-stakes, regulation-heavy vertical where AI delivers returns investors can actually measure.