Anthropic is opening a 158,000-square-foot office in London. The new site will house up to 800 staff.

Big move into a growing AI cluster

Anthropic is moving into a much larger London office next to the city's established AI hub. The company, which set up its first London base in 2023, says the new footprint will be able to accommodate roughly 800 people—about four times its current headcount. The office is in the same district as DeepMind, OpenAI and other AI groups — a cluster that has grown up around nearby universities and research labs.

Being close to other labs and campuses helps companies recruit researchers and move ideas into products.

Pip White, head of EMEA North at Anthropic, said Europe's largest firms and its fastest-growing startups are adopting Anthropic’s Claude models, and that the company is expanding to match that demand. White framed the move as both a commercial step and a signal that Anthropic wants to be close to customers, regulators and talent across the region.

The neighbourhood is already home to a number of well-known AI groups and companies. Anthropic will be neighbours with established research teams and several commercial labs—a mix that the company says helps translate academic work into products.

Politics, safety and a contested US relationship

The move comes after a public dispute between Anthropic and US officials. The company drew a line by refusing to let its models be used for mass surveillance or autonomous weapon systems. That stance contributed to a legal dispute with the Pentagon.

British officials reportedly engaged with Anthropic after those developments in the United States. The UK government has been positioning itself as an active regulator and partner on AI safety, and it has sought closer contact with leading AI firms based outside Britain.

As part of its London push, Anthropic has said it will deepen collaboration with the UK’s AI Security Institute. That government-linked body recently published a risk appraisal of Anthropic’s latest model, Claude Mythos Preview. The institute’s review is part of the UK’s wider effort to assess advanced models before broader deployment.

Mastheads

Reports in other outlets indicate the UK government is among a small group in Europe that has been given access to the Mythos Preview model. According to reports, Anthropic has only shared the Mythos Preview with select partners, citing worries that powerful models could be abused by cybercriminals and other malicious actors.

Talent, proximity and the research pipeline

Because universities and labs are nearby, this part of London has become a draw for AI researchers and engineers. Geraint Rees, vice-provost at University College London, said the cluster didn’t spring from a planning document but from people and organisations choosing to be near one another. He argued that closeness makes it easier for research to feed into real products—though the route from idea to market is often messy and human rather than linear.

Anthropic’s new premises put it within walking distance of UCL’s campus and other academic centres. That matters for recruitment: British universities remain an important source of AI researchers and engineers, and firms often cite local campuses when explaining why they choose particular neighbourhoods.

Competition for experienced AI workers in London has intensified as multiple labs expand their footprints. OpenAI has also announced a London expansion recently, and the area hosts a mix of established tech companies and smaller startups working across the same technical problems. Anthropic’s larger office gives it room to scale quickly if hiring meets the company’s plans.

Commercial aims and safety commitments

Anthropic says the expansion is driven by customer demand in Europe and by the need to align product work with safety practices. Pip White framed the move as a response to both enterprise interest and a growing regulatory focus. Anthropic’s stated refusal to permit certain military or mass-surveillance applications is part of how it explains its approach to safety and deployment.

The company has restricted distribution of some models and opened select channels for national security reviews and government assessments. The UK’s AI Security Institute review of Claude Mythos Preview is an example of the kind of scrutiny Anthropic says it welcomes while limiting broader distribution until risks are better understood.

Those choices have consequences. Tightening access may slow product launches, but it shows Anthropic is ready to work with regulators and subject its models to outside review — something some officials and customers welcome for high-risk systems.

What the expansion could mean for London

The new office adds another large employer to an already busy part of the city. For London, the growth of AI teams is both an economic opportunity and a test of how the capital balances innovation with oversight. Local universities and research centres may see closer ties with industry—through hiring, joint projects or informal collaboration—while regulators will have more companies on their doorstep to consult.

Choosing to scale in London highlights the city's pull as a European centre for AI jobs and research. The firm’s public comments about safety and selective access to models also feed into an ongoing debate about how to manage advanced systems without stifling innovation.

For now, Anthropic is signalling it wants to be part of London’s unfolding AI ecosystem. The company will be able to grow rapidly if hiring goes according to plan, and it intends to pair that growth with deeper engagement with UK safety authorities.

Related Articles

Anthropic’s new London office will offer roughly 158,000 square feet of space and room for about 800 employees.

This article was created with AI assistance.