London has overtaken San Francisco in AI company density as the UK government pours £1.1 billion into sovereign AI infrastructure and US tech giants flood the capital with office space.
London has overtaken San Francisco in AI company density as the UK government pours £1.1 billion into sovereign AI infrastructure and US tech giants flood the capital with office space.

London has overtaken San Francisco in AI company density as the UK government pours £1.1 billion into sovereign AI infrastructure and US tech giants flood the capital with office space.
London has overtaken San Francisco in the number of AI companies with 50-plus employees, as the UK government's £1.1 billion hardware plan and a wave of US tech expansion turn the capital into Europe's undisputed AI hub.
"The UK is one of the places where AI is being built, deployed, and adopted at the same time — by startups, by enterprises, and by the public sector," Arkady Volozh, founder at Nebius, which committed £1.75 billion to UK cloud capacity, said.
AI firms leased 565,000 square feet of London office space in the first four months of 2026, nearly triple the 211,000 square feet for all of 2025, according to leasing data compiled by brokerages. Anthropic took space for 800 people in the Knowledge Quarter, while OpenAI opened its first permanent UK office and Google moved teams into a new 11-story building in King's Cross. The UK government's plan includes £750 million for a supercomputer at the University of Edinburgh and targeted support for domestic chip startups.
The clustering creates a self-reinforcing cycle of talent and capital, but also risks squeezing local startups as well-funded US firms bid up engineering salaries. UK AI startups raised £8.2 billion in venture capital in the first half of 2026, per Tech Nation, while the broader UK tech sector was valued at £1.2 trillion. The question is whether Britain ends up owning a meaningful slice of the industry or simply hosting everyone else's operations.
Government money meets private capital
The UK's new AI hardware plan, announced at London Tech Week, directs £750 million toward a national supercomputer in Edinburgh — a "heterogeneous mixed chip system" slated to open by 2030 as part of the AI Research Resource network. Of that sum, £400 million will go to purchase next-generation AI chips, including £150 million for inference chips from British firms and startups. A separate £120 million hardware innovation program gives startups funding flexibility to develop new chip designs.
The British Business Bank is backing a £150 million fund led by Silicon Valley investor Playground Global, which will open its first office outside the US in the UK. AMD committed £2 billion to UK operations over five years, working with the University of Cambridge on the Zenith and Sunrise supercomputer systems. Dutch neocloud provider Nebius pledged £1.75 billion to expand its UK data center footprint to 65 megawatts by 2027, deploying Nvidia's latest full-stack AI factory technology.
Omdia forecasts IT spending across Europe will increase 8.2 percent in 2026, its fastest pace since 2021, reaching $1.3 trillion. The UK government has also set aside £45 million for AI training, bringing its total skills fund to £80 million, and plans to fund 500 PhD students annually in semiconductor design.
The startup pipeline shows promise
Beyond the headline funding numbers, a cohort of British AI startups is gaining traction across multiple segments. Oriole Networks, working with AMD at the government-backed inference lab, has developed PRISM — a photonic network that replaces electronic switches with optical circuit switching, reducing GPU power consumption by 81 percent and cutting GPU idle time from 60 percent to less than 1 percent. The company expects wider industry rollout in 2027.
Cosine, selected to lead the UK's first fully sovereign frontier AI model called Lumen Sovereign, has rallied a coalition including BAE, BT, HSBC, and the Alan Turing Institute to build an open-weight model deployable in fully air-gapped environments. The model targets release in late 2026 using the Isambard-AI supercomputer in Bristol.
PhysicsX, a London-based startup developing AI digital twins for industrial design, announced an oversubscribed $300 million Series C at a $2.4 billion valuation, led by Singapore's Temasek with participation from Nvidia and Siemens. The company has doubled year-over-year revenue and employs 300 people across London, New York, and the Bay Area.
For investors, the London AI boom presents a dual picture. The influx of US giants — Anthropic, OpenAI, Cursor, Databricks, Salesforce — validates the talent pool but compresses margins for local startups competing for engineers. UK-based chip design giant Arm, owned by Softbank, stands to benefit from any expansion in domestic chip production. The government's role as an early customer for novel chip designs could create a pipeline of public-sector AI procurement worth billions. But the real test will come when the current wave of VC funding matures and London's startups must prove they can generate revenue at scale without being acquired by the US firms they now share office buildings with.
This article is for informational purposes only and does not constitute investment advice.