Anthropic criticizes US export controls on AI models, urges firms to relocate

AI

Anthropic criticizes US export controls on AI models, urges firms to relocate

The AI company suspended global access to two flagship models after a Commerce Department directive, sparking a broader debate about where AI development should happen.

by
Editorial Team

Jun. 22, 2026

Share

Add us on Google

Anthropic, one of the most prominent AI research companies in the world, is pushing back hard against the US government. On June 12, the Commerce Department issued a directive ordering Anthropic to cut off access to its most advanced AI models for all foreign nationals, including those physically present in the US.

Anthropic complied. Then it went public with its frustration, arguing that AI companies should consider moving operations outside the US entirely to protect free speech and dodge government overreach.

What happened

The Commerce Department invoked national security authorities to demand Anthropic suspend access to two of its frontier models: Fable 5 and Mythos 5. The directive came down on June 12, and Anthropic disabled access to both models globally the same day.

Advertisement

The directive reportedly cited a narrow jailbreak vulnerability as the catalyst. But Anthropic says the government failed to provide adequate technical details about the issue, or any workable framework for enforcing nationality-based restrictions across its user base.

Anthropic has warned that this kind of action doesn’t just hurt one company. It could overshadow the deployment of new frontier models across the entire industry.

The sovereignty backlash

European stakeholders have seized on the episode as a cautionary tale about depending on American AI infrastructure. The directive has renewed calls for AI sovereignty across Europe, with policymakers and industry leaders pushing for domestic alternatives that wouldn’t be subject to US export control policy.

Anthropic’s public response went further than most corporate complaints about regulation. The company didn’t just criticize the specific directive. It suggested that AI firms should seriously consider relocating operations outside the US to protect their ability to operate freely.

Why this matters beyond AI

The US has long maintained export controls on everything from advanced semiconductors to encryption software. But applying that framework to AI models, which are essentially software accessible via the cloud, creates enforcement problems that don’t exist with physical goods. The government wanted surgical restrictions but had no scalpel, only a sledgehammer. And the result was a global service disruption that hurt allied nations and American users alike.

For investors with exposure to AI companies, the implications are concrete. Companies heavily dependent on US infrastructure and regulatory goodwill now carry a new category of risk: the possibility that their most advanced products could be suspended with minimal notice and thin justification.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.