Washington Lifts Short-Lived Export Ban on Anthropic's Most Powerful AI Models

A Sudden Lockdown and National Security Fears

In mid-June 2026, the United States Department of Commerce abruptly imposed strict export controls on Anthropic, restricting access to its most powerful artificial intelligence models, Claude Fable 5 and Mythos 5. The sudden regulatory action, initiated by a letter from Commerce Secretary Howard Lutnick to Anthropic Chief Executive Officer Dario Amodei on Friday, June 12, 2026, designated the models as too dangerous for foreign use. This decision was triggered after a competitor claimed it had successfully jailbroken Mythos 5, raising alarms within the administration regarding potential national security vulnerabilities. Although government officials initially requested that Anthropic halt the release of these advanced systems, the company proceeded, prompting the federal export restrictions. Under those rules, any export, re-export, or domestic transfer of the models required a license, and the company was forced to apply for individually validated licenses. To ensure compliance and avoid severe civil and financial penalties, Anthropic deactivated access to the models entirely. The sweeping ban prevented foreign governments, international enterprises, and even Anthropic employees who were not U.S. citizens from utilizing the technology. This regulatory challenge occurred alongside a separate Pentagon blacklist that had already categorized Anthropic as too hazardous for direct government applications.

The Collaborative Path to Reversing the Ban

Following two and a half weeks of intense negotiations and technical evaluations, the Trump administration reversed its stance. Late on Tuesday, June 30, 2026, the Department of Commerce officially lifted the export controls on both Claude Fable 5 and Mythos 5. According to Secretary Lutnick, federal officials spent the preceding two weeks collaborating closely with Anthropic to thoroughly analyze Fable 5 and establish robust security safeguards. As part of the resolution, Anthropic committed to proactively identifying and mitigating security hazards associated with the models and agreed to notify the federal government of any malicious activity detected on its platforms. However, the administration made it clear that the lifting of these controls is conditional. In his correspondence, Lutnick cautioned that the government retains the authority to reinstate the export restrictions if security conditions shift or if Anthropic fails to meet its safety commitments. White House Chief of Staff Susie Wiles publicly expressed appreciation for the cooperation, noting that the resolution aligns with the administration's Executive Order on Promoting Advanced AI Innovation and Security, which prioritizes deploying advanced technology rapidly while maintaining strict safety guardrails.

Restoring Global Access to Fable and Mythos

With the restrictions removed, Anthropic announced it would begin restoring access to its flagship models starting Wednesday, July 1, 2026. The company stated that Claude Fable 5 is being made available to its global user base on the Claude platform. Meanwhile, access to Mythos 5 has been restored for select organizations within the United States, and Anthropic is actively working with federal regulators to safely expand access to additional domestic and international partners. The return of these models comes at a highly competitive moment, as industry observers note the emergence of a new Chinese AI model that reportedly rivals or exceeds the capabilities of both Fable 5 and Mythos 5. Simultaneously, Anthropic is expanding its focus beyond conversational AI, recently launching an internal drug discovery initiative led by its head of life sciences to build specialized AI tools for pharmaceutical manufacturers. This policy shift reflects a broader evolution in U.S. trade strategy, which has increasingly utilized export controls—such as the December 2025 approval of specific advanced chip licenses—as leverage to balance international economic competitiveness with national security priorities.

As Washington increasingly treats advanced neural networks as dual-use national security assets, the rapid cycle of banning and then reinstating Anthropic's flagship models suggests that the boundary between protecting state secrets and maintaining American commercial dominance remains highly volatile.