London's AI Pitch to Anthropic: Stability or Empty Promise?

A handwritten invitation from London's mayor has landed on the desk of one of Silicon Valley's most prominent AI chiefs, following a week that saw an American AI company become the first domestic firm ever designated a 'supply chain risk' by its own government. The pitch is simple: abandon the chaos of Trump's Washington for the supposed stability of post-Brexit Britain.

Sadiq Khan's letter to Anthropic CEO Dario Amodei represents Britain's opening gambit in what could become a defining test of whether the UK can genuinely capitalise on American political turbulence to attract marquee tech investments. The context makes the timing irresistible: Anthropic refused Pentagon demands for military access to its Claude AI system, citing concerns over mass surveillance and autonomous weapons. Within days, its lucrative defence contract was terminated and Trump slapped the company with a designation previously reserved for foreign adversaries like Huawei.

What's interesting here is that Khan is effectively asking whether principled opposition to military AI applications might find a more sympathetic home in London than California. His letter praised Anthropic's 'steadfastness' and promised a 'stable, proportionate and pro-innovation environment'. The subtext is unmistakable: American tech policy has become weaponised and unpredictable under Trump's second administration.

Enjoying this article?

Get stories like this in your inbox every week.

Join the Community

The practical obstacles no one's mentioning

Strip away the diplomatic niceties and this pitch faces substantial complications. Anthropic has raised billions from Amazon and Google, two American giants whose cloud infrastructure underpins the company's entire operation. Amazon Web Services hosts Anthropic's models; Google Cloud provides additional compute capacity. Relocating such deeply embedded infrastructure dependencies isn't a matter of switching office locations.

Venture capital firms and corporate backers don't typically fund companies that then relocate primary operations outside US jurisdiction, particularly when those operations involve cutting-edge technology with potential national security implications.

The company's investor base poses another constraint. Moving Anthropic's headquarters would require unprecedented investor consent and could trigger thorny questions about corporate governance, IP ownership, and regulatory compliance across multiple jurisdictions.

Khan's letter requested a meeting to discuss how Anthropic could 'expand operations further' in London. That phrasing matters. Expansion differs dramatically from relocation, and the distinction suggests even the mayor's office recognises the impracticality of Anthropic abandoning the US altogether.

Britain's AI infrastructure reality check

The UK government has spent considerable political capital positioning Britain as an 'AI superpower', a phrase that sounds increasingly aspirational when measured against material constraints. Britain lacks the scale of compute infrastructure that American hyperscalers have built. The country's capital markets, whilst substantial, cannot match the depth of American venture funding for frontier AI development.

According to figures from Dealroom, UK AI companies raised £3.9bn in 2023, impressive by European standards but a fraction of the estimated £42bn that flowed to American AI firms in the same period. British regulatory posture towards AI also remains uncertain. The government has resisted the EU's comprehensive AI Act approach, favouring sector-specific regulations that don't yet exist in concrete form.

Whether that framing lands with Amodei depends on whether he views undefined British regulations as preferable to politically volatile but technically familiar American ones.

This creates its own ambiguity, precisely what Khan's letter attempts to position as clarity by contrast with American chaos. Amodei has already indicated his intention to challenge the supply chain designation in US courts, describing it as exceeding statutory authority. That legal strategy suggests commitment to remaining within American jurisdiction rather than relocating to avoid it.

What happens if Britain actually lands this

Attracting Anthropic would represent the most significant tech coup for the UK since the country's departure from the EU. The company employs hundreds of researchers and engineers, precisely the high-skilled workforce Britain claims to want. Its presence would validate claims that London can compete for frontier AI development, not merely host European sales offices for American platforms.

The knock-on effects matter more than the immediate win. If Anthropic establishes substantial operations in London, other AI labs facing American regulatory pressure might view Britain as a credible alternative. OpenAI, which quickly secured the Pentagon contract Anthropic rejected, has already expanded its London presence. DeepMind, whilst owned by Google, maintains significant British operations from its origins as a London startup.

The more likely outcome involves Anthropic expanding its existing small London footprint without relocating primary operations. That would give Khan a symbolic victory and Britain some additional high-value jobs, but wouldn't fundamentally shift where frontier AI development occurs. American compute infrastructure, capital availability, and talent density remain structural advantages that political chaos hasn't yet overcome.

Khan's gambit tests whether 'stable governance' has become valuable enough to American tech firms that they'll accept Britain's smaller market, less developed infrastructure, and uncertain regulatory future. The answer will reveal whether Brexit Britain's promise to operate as a nimble, business-friendly alternative to both American chaos and European regulation amounts to genuine competitive advantage or merely consolation narrative.

More articles by Ross Williams