On June 12, Commerce Secretary Howard Lutnick ordered Anthropic to block foreign access to its advanced Fable 5 and Mythos 5 AI models, citing national security risks. Anthropic chose to shut down the models entirely rather than develop nationality-based access controls.
The abrupt cutoff affected allied governments, European companies, and research centers. The European Union quickly raised the issue with Washington. French President Emmanuel Macron, speaking at the G7 summit in Evian, advocated for a “trusted partners” framework that would allow vetted allied nations to access U.S. AI systems without security concerns.
Anthropic co-founder Tom Brown engaged in talks with National Cyber Director Sean Cairncross to restore access. The directive was triggered by concerns over potential jailbreak vulnerabilities in the models.
The incident has intensified European calls for AI sovereignty. Macron’s proposal offers a pragmatic near-term fix, while full independence remains a long-term goal. For investors, the disruption creates immediate uncertainty for businesses built on Anthropic’s API, and the company’s sweeping shutdown raises questions about its technical capabilities or strategic alignment with U.S. government priorities.