Anthropic and London court each other as Pentagon backlash deepens

Anthropic’s clash with the Pentagon has quickly spilled over into a competition between jurisdictions to host frontier AI labs. Sadiq Khan’s overture is more than symbolic politics: it is a signal that London is willing to position itself as a haven for firms that push back on US national‑security demands, so long as they fit within the UK’s emerging AI rulebook. For Anthropic, which is already present in London, the letter strengthens its leverage in negotiations with Washington by underscoring it has friendlier options abroad.

For the broader race to AGI, this episode highlights how political risk is now tightly coupled to frontier model development. Governments are starting to sort into different camps on how comfortable they are with military applications, surveillance uses, and corporate self‑governance. Cities and countries that can credibly promise both regulatory stability and ethical alignment will gain an edge in attracting high‑end researchers and data‑center investment. At the same time, the US government’s “supply chain risk” label is a reminder that access to defense and classified workloads can be weaponized against labs that resist certain deployment paths.

Over the next few years, we should expect more explicit bidding wars for AGI labs between North America, the UK and the EU, with tax incentives, talent visas and public‑sector contracts all on the table. Where Anthropic ultimately concentrates headcount and infrastructure will tell us a lot about which jurisdiction is winning that contest.