Anthropic is hitting the gas on its European expansion. This week, the AI unicorn announced it's moving into a massive new 158,000-square-foot office in London's King's Cross district—a space that can hold 800 employees, quadrupling its current capacity.
The new digs put Anthropic right in the thick of what's become one of the world's densest AI clusters, nestled among neighbors like Google DeepMind, OpenAI, Meta, Wayve, and Isomorphic Labs. Pip White, Anthropic's head of EMEA North, says the company needs to scale up to meet demand from Europe's largest enterprises and fastest-growing startups, which are increasingly choosing its Claude AI models.
"The UK has both ambitious companies and institutions that understand the importance of AI safety, and an exceptional pool of AI talent," White said. "That's exactly why Anthropic is choosing to deepen our roots here."
This move comes just two years after Anthropic first planted its flag in London. And here's the kicker: with space for 800, Anthropic's new office could soon eclipse its rival OpenAI's planned London presence. In February, OpenAI announced it would make London its largest research center outside the U.S., with plans to house over 500 people there.
Geopolitics in the background
There's more to this expansion than just real estate. According to previous reports, the UK government has been actively courting Anthropic, leading the company to consider not only a larger London office but also a potential dual listing. This courtship comes at a delicate time for Anthropic, which is currently locked in a legal dispute with the U.S. Pentagon.
The company was placed on a U.S. government supply chain risk list and had a $200 million military contract terminated after it refused to allow its models to be used for mass surveillance or autonomous weapons systems.
As part of its UK push, Anthropic says it will deepen its collaboration with the UK's AI Safety Institute. That institute just this week released a risk assessment report on Anthropic's latest model, Claude Mythos Preview. The UK government is one of the few entities in Europe granted access to the model, which Anthropic has restricted to select partners over concerns about potential misuse by cybercriminals.
The cluster effect
UCL Vice-Provost Geraint Rees observed that this concentration of AI power in King's Cross didn't emerge from some top-down planning document. "It's because serious researchers and companies understand the importance of geographic proximity," he said. That clustering creates a pipeline that turns research into products.