Anthropic Leases 158,000‑Square‑Foot London Space, Plans to Quadruple Workforce
Key Points
- Anthropic leases 158,000 sq ft office in London, capacity for 800 staff.
- Expansion aims to quadruple the company's London workforce from 200.
- Location sits alongside DeepMind, OpenAI, UCL and other AI firms.
- Legal dispute with the U.S. Pentagon over model use in surveillance and weapons.
- UK government and AI Security Institute to collaborate on Claude Mythos Preview.
- Company seeks to meet rising European demand for safe, trustworthy AI.
- Move positions Anthropic to compete directly with OpenAI’s London expansion.
Anthropic announced it will occupy a new 158,000‑square‑foot office in London, enough to house up to 800 employees—four times its current head count. The move aims to deepen the company’s research and commercial presence in Europe amid a talent race with other AI labs. The expansion comes as Anthropic faces a legal dispute with the U.S. Pentagon over its refusal to allow its models in mass‑surveillance or weapon systems, while the U.K. government seeks closer cooperation on AI safety and security.
Anthropic, the San Francisco‑based AI lab behind the Claude series, signed a lease for a 158,000‑square‑foot headquarters in London’s tech corridor. The space, located near Google DeepMind, OpenAI and a host of AI research outfits, can accommodate up to 800 staff, effectively quadrupling the company’s London workforce from its current 200.
“Europe’s largest businesses and fastest‑growing startups are choosing Claude, and we’re scaling to match,” said Pip White, head of EMEA North at Anthropic. “The UK offers an ambitious ecosystem and a deep pool of AI talent—we want to be where that comes together.”
The new premises will place Anthropic in the same neighborhood that hosts University College London, Wayve, Isomorphic Labs and other innovators. Geraint Rees, vice‑provost at UCL, noted that the clustering of researchers and firms is not the result of a planning directive but a natural gravitation toward proximity, which accelerates the translation of lab breakthroughs into marketable products.
Anthropic’s expansion coincides with heightened friction with the United States government. Earlier this year the company refused a Pentagon request to deploy its models for mass‑surveillance and autonomous weapons, prompting a legal battle that has strained relations with the U.S. administration. British officials, meanwhile, have been courting the firm, hoping its presence will bolster the nation’s AI safety agenda.
As part of the London rollout, Anthropic will deepen collaboration with the UK’s AI Security Institute. The institute recently released a risk evaluation of Claude Mythos Preview, the latest model in the Claude line. The UK government is among the few European entities granted access to the model, a privilege Anthropic has limited to select partners over concerns about misuse by cybercriminals.
The move signals Anthropic’s bet that Europe, and particularly the United Kingdom, will become a pivotal market for its products. By scaling up its European operations, the company hopes to compete more directly with rivals such as OpenAI, which announced its own London expansion last month.
Industry observers see the concentration of AI firms in the same district as a catalyst for rapid innovation. Proximity allows for talent sharing, joint research projects and quicker feedback loops between academia and industry. While the legal dispute with the Pentagon casts a shadow over the company’s U.S. prospects, the London office positions Anthropic to tap into a talent pipeline that includes graduates from top British universities and a growing pool of AI specialists.
Anthropic’s next steps will involve hiring engineers, researchers and product teams to fill the new space, as well as rolling out Claude Mythos Preview to UK customers under strict usage guidelines. The company’s leadership says the expansion reflects a broader strategy to align AI development with rigorous safety standards while meeting the growing demand for trustworthy generative models.