US bans foreigners from Anthropic's top models — company disabled them for everyone

The U.S. government has forced Anthropic to block access to its two most powerful models from foreign users. Instead of implementing selective citizenship-based blocking within a shared cloud service, Anthropic has disabled Claude Fable 5 and Claude Mythos 5 globally for everyone.

What happened

On the evening of June 12, 2026, Anthropic unexpectedly deactivated Claude Fable 5 and Claude Mythos 5 worldwide — including for users in the United States itself. The reason: a directive from the U.S. government requiring that the company's two most powerful flagship models be blocked from foreign citizens.

These are Anthropic's top products — Claude Fable 5 and Claude Mythos 5 were positioned as the company's most advanced systems, designed for complex enterprise tasks. They are the ones covered by the government directive.

Technically, implementing precise citizenship verification within a unified cloud service is a non-trivial task. It requires verifying documents from millions of users, configuring exception systems for dual citizenship, and bearing legal responsibility for identification errors. Anthropic did not develop such a mechanism on a tight timeline and simply disabled both models for everyone without exception — a more radical but technically straightforward solution.

Only three days passed between receiving the directive and the global shutdown. The government set tight deadlines, leaving no time for a smooth transition.

Why this is called a turnaround

In recent years, Anthropic has openly positioned Claude as a global service. The company expanded language support, built partnerships with European and Asian corporate clients, and attracted international developers through its API. The forced shutdown of flagship models starkly contrasts with this open-access strategy.

• Claude Fable 5 and Claude Mythos 5 have been unavailable since the evening of June 12, 2026
• The shutdown affected users in all countries, including U.S. citizens
• The formal reason — an authority requirement to block access for foreign citizens
• This is the first known case of a global shutdown of a commercial AI model by government directive

Previously, Anthropic adhered to the principle of open global access. Now that principle is violated — not by the company's decision, but by direct order from U.S. authorities.

Export controls reach models

The U.S. has long applied export restrictions on hardware: bans on shipping Nvidia H100 and H200 chips to China have become standard tools of technological policy. Now this approach is being applied for the first time directly to software AI systems.

The precedent is significant for several reasons. It demonstrates the government's willingness to regulate AI capabilities as a strategic resource on par with chips — the boundary between a commercial software product and a strategic technology is shifting. This creates legal uncertainty for all American AI companies whose significant revenue comes from international clients.

For Anthropic's corporate clients outside the U.S., the situation is particularly painful: they built products and internal processes around specific models that are now inaccessible without warning. The company has yet to announce when and under what conditions access will be restored.

What this means

The Anthropic case is the first clear signal that the U.S. government is prepared to regulate AI systems as objects of export control, not just chips. For OpenAI, Google DeepMind, and other market players, this is a troubling precedent. For international companies and developers building businesses on American AI services, it is a reminder: technological access can be interrupted at any moment for national security reasons beyond their control.