Anthropic and Trump Administration Reach Truce as Export Controls on Claude Fable 5 and Mythos 5 Are Lifted
On Tuesday, June 30, Anthropic announced that the U.S. Department of Commerce had lifted export controls on its Claude Fable 5 and Mythos 5 models, bringing a dramatic close to one of the most consequential confrontations between an AI company and the federal government in the industry's history. As CNBC reported, the company confirmed the controls had been removed and said it would restore access to Fable 5 on Amazon Web Services, Google Cloud and Microsoft Foundry as soon as possible. The standoff began on June 12, when the Commerce Department's Bureau of Industry and Security, under the signature of Commerce Secretary Howard Lutnick, ordered Anthropic to suspend all access to the two models by any foreign national — anywhere in the world, including foreign nationals employed by Anthropic itself. Because the company could not reliably distinguish foreign nationals from U.S. persons across its global user base in real time, it abruptly disabled both models for every customer worldwide, just three days after their public launch.
The trigger for the government's intervention was a reported jailbreak technique capable of bypassing Fable 5's safety guardrails, which risked turning what Anthropic had billed as a consumer-facing model into an unrestricted cyber tool. White House AI adviser David Sacks publicly claimed that Anthropic had refused to address the security flaw, a charge the company disputed, criticising the government's process as opaque and lacking specific technical grounding. According to Forbes, Amazon CEO Andy Jassy was among several tech leaders who raised concerns about security risks in the models to senior Trump administration officials. The episode also intersected with a longer-running conflict: since February 2026, Anthropic had been at loggerheads with the Department of Defense, which had demanded the company allow Claude to be used for lethal autonomous weapons and mass domestic surveillance — terms Anthropic refused, leading to its designation as a supply chain risk, a label historically reserved for foreign adversaries.
The partial resolution came in stages. On June 26, Commerce Secretary Lutnick wrote to Anthropic co-founder Tom Brown — notably not to CEO Dario Amodei, who had been sidelined from negotiations given his political profile — stating that appropriate safeguards were in place to permit roughly 100 trusted companies and federal agencies access to Mythos 5. As CNBC confirmed, Tuesday's announcement extended that relief fully, ending the export controls and allowing Anthropic to restore both models across its major cloud partners. The company said it would continue working with the government through its Glasswing cybersecurity programme, which provides selected organisations access to advanced Claude models for defensive security testing. Anthropic also updated its privacy policy to introduce government-issued ID verification, a mechanism analysts expect will serve as the primary compliance tool for any future nationality-based access controls.
The geopolitical stakes of the standoff were underscored by events beyond the courtroom. As CNBC reported, the government's crackdown on Anthropic coincided with a swift rise in Chinese open-source models that are proving nearly as capable and significantly cheaper than leading U.S. offerings. A number of tech executives and investors warned publicly that restricting Anthropic's rollout effectively gifted Chinese developers valuable time to close the capability gap. China's Zhipu and other frontier labs moved quickly to market their availability to enterprise customers who had previously relied on Claude. The episode crystallised what many in the industry had long feared: that export control frameworks designed for physical hardware were being stretched, without clear statutory authority, to cover the deployment of commercial software at a global scale.
The legal battles are far from over. Anthropic's lawsuits challenging its supply chain risk designation continue in two federal courts, and a federal judge in San Francisco has already granted the company a preliminary injunction barring enforcement of the Claude use ban across the Department of Defense. The outcome of those cases will determine whether the government can impose the kind of open-ended restrictions it sought against Anthropic on any AI lab that declines to comply with military demands. Analysts at TechPolicy.Press have warned that a governance regime premised on perfect technical control of frontier models is ultimately incompatible with deploying useful AI at all, and that a more realistic approach requires layered defences combining technical safeguards, monitoring, user vetting, and government oversight. As the next generation of frontier models approaches release — with Anthropic's updated privacy policy taking effect July 8 and key executive order deadlines falling on August 1 — the truce between the AI industry's most safety-focused lab and the world's most powerful government will face its next test.