Trump Administration Blocks Foreign Access to Anthropic AI Models

Export Controls Target Advanced AI Systems

The Trump administration has imposed strict export controls on Anthropic's most advanced AI models, prohibiting foreign governments, companies, and individuals from accessing the technology. Commerce Secretary Howard Lutnick notified Anthropic CEO Dario Amodei on Friday that the Mythos 5 and Fable 5 models would require licensing for any export, re-export, or domestic transfer to foreign persons.

The action follows claims from another company that it successfully jailbroke Mythos, triggering national security concerns within the administration. According to an administration official, the government attempted to persuade Anthropic to delay releasing the models but was unsuccessful, prompting the Commerce Department's intervention.

Why It Matters

This represents the first time the U.S. government has used export control authority to restrict access to a commercial AI model in real time. Anthropic now occupies an unusual position: its technology is simultaneously deemed too risky for Pentagon use (via a separate blacklist) and too sensitive for foreign access. The move signals that Washington is prepared to treat frontier AI systems as strategic assets comparable to advanced semiconductors or weapons technology, potentially reshaping how AI companies operate globally.

Licensing Requirements and Penalties

Under the Commerce Department's directive, Anthropic must obtain licenses for any transfer of the restricted models. The company will need to submit applications for individually validated licenses, a process that could significantly constrain its ability to serve international customers or collaborate with foreign researchers. Non-compliance would trigger financial and civil penalties, though specific amounts were not disclosed.

An administration official indicated the restrictions would remain in place until the U.S. government's national security infrastructure is adequately prepared to handle the technology—a process expected to take several weeks.

Tension with Voluntary Testing Framework

The export control action creates friction with the Trump administration's recent executive order on AI testing, which established a voluntary framework for evaluating advanced models before deployment. White House chief AI adviser David Sacks championed the voluntary approach specifically to avoid mandatory licensing, which he views as enabling regulatory capture by dominant companies.

Anthropic participates in pre-deployment testing through a partnership with the Commerce Department's Center for AI Standards and Innovation. Yet the company now faces mandatory controls that extend beyond the voluntary testing regime. An administration official emphasized that President Trump "does not want to hurt the industry and wants innovation to continue," highlighting the tension between security concerns and economic competitiveness.

Growing Regulatory Pressure

The export controls add another layer of complexity to Anthropic's relationship with federal authorities. The company already faces restrictions on government contracts due to its inclusion on a Pentagon list of systems considered too risky for official use. The dual restrictions—barring both domestic government deployment and foreign access—place Anthropic in a uniquely constrained position among major AI developers.

These details were first reported by Axios. Anthropic did not immediately respond to requests for comment on the new restrictions.