The US Department of Commerce compelled Anthropic to immediately cut off foreign national access to its two most advanced AI models, Fable 5 and Mythos 5, on June 12 regarding a jailbreak vulnerability classified as a threat to critical infrastructure.
Anthropic executed a global shutdown of both models by June 13. The company determined that building a real-time nationality verification system across its infrastructure was technically unfeasible, choosing a complete operational halt over compliance with a partial ban.
The directive contradicted the voluntary cybersecurity collaboration framework President Trump outlined just ten days prior in an executive order. The sudden shift from light-touch guidance to forced product shutdown signals severe regulatory volatility for centralized AI platforms.
Market consequences were immediate. Anthropic’s pre-IPO shares fell sharply as investors reassessed the geographic vulnerability of centralized AI assets. Conversely, decentralized AI tokens experienced a sharp surge. Venice (VVV) and Morpheus (MOR) rallied as the market priced in the value of censorship-resistant infrastructure immune to single-point government intervention.
The event establishes a clear dichotomy for institutional portfolios: centralized AI models now carry demonstrable geopolitical shutdown risk, while permissionless, decentralized protocols become the hedge against regulatory disruption.