The Trump administration has declared Anthropic’s newest AI models national security risks, imposing immediate export controls that lock out all international users. Claude Fable 5 and Claude Mythos 5 are now restricted solely to domestic entities and approved Project Glasswing partners.
Anthropic launched both models on June 9, 2026. While Fable 5 serves as the safety-guarded public version, Mythos 5 is a powerful cybersecurity tool intended for vetted government defenders. The crackdown followed reports that a third party successfully bypassed Mythos 5 security protocols, triggering federal intervention.
These restrictions severely impact Anthropic’s financial outlook. The company recently doubled pricing to $10 per million input tokens and $50 per million output tokens. With the global market now inaccessible, competitors like OpenAI and Google DeepMind are positioned to capture displaced international demand.
This regulatory action signals a shift in how Washington treats frontier AI, equating it with advanced semiconductor technology. For investors, the unilateral restriction of global reach introduces significant geopolitical risk. Valuation models assuming worldwide adoption must now account for potential export controls, likely increasing discount rates on future AI cash flows.