US lifts export controls on Anthropic’s top AI models

The U.S. Department of Commerce announced it has removed export controls on Anthropic’s most advanced large‑language models, Claude Fable 5 and Mythos 5, after a prior shutdown over security concerns about potential misuse for cyber attacks. Anthropic noted it received “the notification that the Department of Commerce lifted the export controls” and that the models are now back online for U.S. users, while access for foreign partners remains highly restricted.

The CIA’s director, John Ratcliffe, later described the capabilities of such frontier AI systems as “digital nuclear weapons,” underscoring the strategic view of the technology. In a related move, the U.S. also limited OpenAI’s newly announced GPT‑5.6 model to a closed set of approved companies and partners.

These actions illustrate a growing U.S. technocratic chokehold on critical AI infrastructure, prompting European officials to warn that reliance on American providers could leave the continent vulnerable and to call for the development of its own sovereign AI capabilities.