The U.S. government ordered Anthropic on Friday to suspend access to its two most advanced AI models, Fable 5 and Mythos 5, citing national security concerns over a potential jailbreak. The directive immediately triggered a sell-off in the company's pre-IPO market, raising questions about how regulatory interventions will impact the deployment of frontier artificial intelligence systems.
The national security directive
Anthropic received the order at 5:21 p.m. ET on Friday. The directive, citing national security authorities, required the company to suspend all access to both models for foreign nationals worldwide. The company subsequently executed a full shutdown for all customers.
The government informed Anthropic that it had discovered a method to bypass or jailbreak Fable 5. Anthropic reviewed the technique and concluded it was narrow. The company argued the method only involved identifying a small number of previously known, minor vulnerabilities that other publicly available models, including OpenAI's GPT-5.5, can find without any bypass.
Anthropic's pushback
The company said the government has so far provided only verbal evidence of a potential narrow jailbreak. Anthropic described the flagged task as essentially asking the model to read a codebase and fix software flaws, a routine task for software defenders.
Anthropic warned that applying this regulatory standard across the industry would have severe consequences. The company stated that such a move "would essentially halt all new model deployments for all frontier model providers." This public dispute highlights a growing tension between safety-first AI development and the practical realities of model deployment, a critical consideration for those studying AI for Policy Makers.
Market reaction and next steps
Traders reacted negatively to the shutdown. The Anthropic perpetual contract on Hyperliquid, which tracks where traders expect the company's equity to price, dropped 3.7 percent on Saturday to approximately $1,627. Open interest on the contract sits near $8.6 million, a meaningful figure for a company that has not yet filed for an initial public offering.
Anthropic plans to share more technical details about the specific jailbreak within 24 hours. The market is currently pricing the shutdown as a negative for the company's IPO case. The trajectory of the company's public listing ambitions will depend on whether the government's order is reversed, narrowed, or extended to other model classes once Anthropic publishes its technical rebuttal.
Why this matters for government professionals
Regulatory actions based on verbal evidence or narrow vulnerabilities set immediate precedents for how agencies oversee emerging technologies. When federal authorities mandate a full shutdown of commercial AI infrastructure, it establishes a baseline for future compliance expectations. Officials managing AI for Government initiatives must anticipate similar friction as public sector adoption of frontier models scales. Understanding the technical boundaries of these systems is necessary to draft directives that balance security with operational continuity.
Your membership also unlocks:
