Runway, the AI video generation startup valued at $5.3 billion after its $315 million Series E in February 2026, has chosen London as its European headquarters and committed over $200 million to the UK’s AI sector by the end of 2028. The round was led by General Atlantic, with Nvidia, AMD Ventures and Fidelity among the participants. Runway’s co-founder cited proximity to European clients and access to research talent as the reasons for the London choice.
It’s the third time in under a year that a major US AI lab has made the same decision. In April 2026, OpenAI opened its first permanent London office at King’s Cross, with capacity for 544 workers, having designated the city its largest research hub outside the US in February. That same month, Anthropic expanded its London presence to 158,000 square feet in the Knowledge Quarter, scaling from 200 to 800 staff capacity.
When three top-tier AI firms choose the same city in a single year, it begs the question: why this location, and is the momentum as solid as it looks?
The Data Behind The Hype
London is home to about 758 AI companies, according to a report released during London Tech Week 2026, which is more than Paris and Berlin combined. Almost 30% of Europe’s new generative AI startups are based in the city.
London AI startups raised a record $3.5 billion in VC investment in 2024, a 52% increase from the year before, making the city third globally for AI venture capital behind New York and the Bay Area. The UK AI market had a combined valuation of $230 billion in Q1 2025, the largest in Europe, with 20 AI unicorns and more than 2,300 VC-backed AI startups.
The investment is unusually concentrated geographically: 90% of UK AI investment over the past five years went into the London-Cambridge-Oxford triangle, with London alone accounting for 71% of total UK AI investment and $13.3 billion over five years. The UK government has backed the momentum with £2 billion committed to the AI sector over four years, approximately £1 billion toward sovereign compute capacity with a 20-fold expansion target by 2030, and up to £500 million for a Sovereign AI Unit. On talent, the UK ranks second globally for tech talent, just behind San Francisco.
Why US Labs Keep Landing Here
The reasons aren’t hard to find – London shares a time zone with European clients, has the deepest research talent pool outside the US, and offers labour costs considerably below San Francisco. Google DeepMind is London-based. Synthesia, one of Europe’s most valuable AI companies, is headquartered in the Knowledge Quarter alongside Anthropic’s new office. The concentration creates a pull: more AI companies mean more talent, which means even more AI companies.
Regulatory strategy is part of the calculation too. London gives them a base inside the European market without full exposure to EU regulation, access to the UK’s relatively pragmatic AI policy environment, and proximity to financial services, insurance and legal sectors where enterprise AI has the clearest near-term commercial value. According to data covering 2021 to 2024, London is the second most attractive destination globally for AI-related foreign direct investment, behind only Dubai.
But Let’s Not Get Carried Away
The celebration is warranted, but the picture isn’t uniformly positive.
Access to capital and talent remain the two biggest growth barriers cited by UK AI startup leaders, despite the record investment numbers. The $3.5 billion raised in London in 2024 is impressive in European terms and negligible compared to the Bay Area’s $60.7 billion in the same year. The US labs choosing London are doing so partly because of what the city offers and partly because it’s the least difficult European option – which is a different kind of endorsement.
The arrival of OpenAI, Anthropic and Runway is either going to create the dense commercial and research environment that pulls in the next generation of AI startups and talent, or the infrastructure investment will flow mostly to the US labs themselves while the domestic startup base continues to struggle for capital at scale. Those are two very different versions of the same headline, and the data won’t make clear which one is playing out for another two years at least.
