White House Lifts Export Controls on Anthropic After National Security Standoff

On June 30, 2026, the United States government lifted strict export controls on Anthropic's most advanced artificial intelligence models, Claude Fable 5 and Mythos 5, resolving a high-stakes, weeks-long standoff between the Trump administration and the artificial intelligence developer. The Commerce Department's decision allows Anthropic to resume global access to Fable 5 on its cloud platform starting July 1, 2026, while access to Mythos 5 is being restored for select domestic organizations and government-approved partners.

The National Security Freeze and the Amazon Jailbreak

The dispute began in mid-June 2026, shortly after Anthropic launched Claude Fable 5 to the public. Fable 5 was designed as a consumer-ready version of its highly advanced Mythos model, equipped with specialized guardrails to prevent exploitation for cyberattacks or other malicious activities. However, a trusted partner, later identified as Amazon, discovered a jailbreak technique that could bypass these protective guardrails.

Upon learning of the vulnerability, the Commerce Department issued an export control order that restricted foreign nationals, including Anthropic's own international employees, from accessing the technology. Because of the broad nature of the government's mandate, Anthropic felt compelled to completely disable user access to both Fable 5 and Mythos 5 around mid-June to remain in compliance with national security regulations.

A High-Stakes Dispute Over Frontier AI Guardrails

The sudden shutdown sparked intense debate within the technology sector regarding the boundaries of federal AI regulation. Anthropic openly disagreed with the government's sweeping action, arguing that the jailbreak vulnerability was relatively simple and easily replicated on other publicly available models across the industry. The company cautioned that if a narrow potential jailbreak became the standard threshold for recalling commercial models, it would effectively freeze all future deployments from frontier AI developers.

To resolve the impasse, Anthropic implemented a new safeguard engineered specifically to block the vulnerabilities discovered by Amazon. Meanwhile, trade policy experts noted that the administration's actions reflected a broader trend of deploying export controls as multi-purpose trade and national security instruments. This aggressive regulatory approach has also impacted other industry leaders, with the White House reportedly requesting that OpenAI limit the release of certain upcoming models.

Resolution and Restored Access Under Project Glasswing

After two weeks of intense negotiations, federal officials cleared the models for redeployment. Commerce Secretary Howard Lutnick announced on the social media platform X that his department worked closely with Anthropic to analyze and approve Fable 5, aiming to align policies across the United States government and strengthen domestic leadership in artificial intelligence.

While the public gains access to Fable 5, the deployment of Mythos 5 remains highly restricted. Mythos 5 is a version of the model built with fewer guardrails, which experts warn can identify and exploit software vulnerabilities at an unprecedented speed. To manage these risks, the government previously allowed a partial easing of controls, permitting Anthropic to distribute Mythos 5 to a restricted group of partners under an initiative known as Project Glasswing, designed to secure critical global software infrastructure. Anthropic plans to continue working alongside the federal government to safely expand access to more domestic and international partners over time.

With the government demonstrating its willingness to freeze entire commercial model deployments over simple jailbreaks, frontier AI developers must now brace for a regulatory landscape where national security concerns can instantly override market access.