Sovereign AI: America Built an AI Kill Switch — Then Aimed It at Allies

On June 12, 2026, the United States did something it had never done to a piece of software: it treated a frontier AI model the way it treats a weapon.

A Commerce Department order forced Anthropic to cut its two most capable models off from every foreign national on Earth — not just hostile states, but the company's own visa-holding engineers, allied firms, and allied governments. Within a day the models went dark for everyone, because no provider can sort foreign users from American ones in real time.

The kill switch everyone had theorised about turned out to be a single phone call from Washington — and the people it caught first were America's closest friends. A proposed UK exemption died; AUKUS-partner Australia was locked out with no warning and no backup plan.

That reframes the question almost every government is answering wrong. The reflex is “build data centres on our own soil.” But the thing switched off was permission, not hardware — a company in Frankfurt was as cut off as one in Virginia, because the control attached to the model, not the metal. This report traces how the most powerful software on the planet became an export-controlled good, where Europe actually stands when the music stops, why Mistral quietly stopped racing the frontier, and what sovereignty over AI really requires once you know the plug exists. The honest answer is contested — some argue for building the whole stack at home — but the evidence leans toward something narrower and more achievable: guaranteed, un-revokable access plus leverage of your own, not a compute empire you can never out-spend.

I'm Dan. The AI revolution is moving way too fast for anyone to keep up — so I built a custom stack of AI tools to chase down the questions I can't stop thinking about and share what I find. Building this is the whole point: it's the revolution helping you keep pace with the revolution. AI-assisted, fact-checked, and always worth a second look.

The full written report — with the charts, sources and interactive exhibits — is on the site: read it here. Browse every episode and all of Dan's shows at connectiveshift.com.

This show is AI-generated and fact-checked — deeply researched, and worth a second look. Follow Dan's AI Intel — it's free, and the single best way to help a small independent show grow. Thoughts or pushback? [email protected] — I read every one.