US Government Blocks Anthropic's Most Powerful AI Models

On June 12, 2026, the Trump administration issued an export control directive forcing Anthropic to shut down access to its two most advanced AI models — Fable 5 and Mythos 5 — for all foreign nationals worldwide. The order came just days after the models' public release and was triggered by what reports describe as a jailbreak that raised national security concerns.

The directive is unprecedented in its scope. It doesn't just block foreign governments or sanctioned entities — it suspends access for any non-US citizen, anywhere on the planet. Anthropic complied immediately, disabling the models globally while retaining access for US persons.

This is the first time the US government has used export controls to restrict a consumer-facing AI model at this granularity. Previous AI export restrictions targeted hardware (Nvidia chips) or specific country embargoes. The Anthropic order signals a new doctrine: AI models themselves are now treated as dual-use technology subject to individual-level access controls.

For developers and companies building on Claude APIs outside the US, the implications are serious. Any advanced model can now be revoked without notice. Teams that relied on Fable 5 for production workloads had to migrate overnight. The incident also raises questions about whether OpenAI and Google will face similar restrictions on their frontier models, and whether ally nations will receive exemptions.

The policy could fragment the global AI landscape. Developers in MENA, Europe, and Asia may need to plan for model access as a fragile dependency, not a guaranteed service.