Anthropic ends 18-day standoff as US lifts export controls on 2 AI models

The Trump administration lifted export controls on Anthropic's two most powerful AI models, ending an 18-day freeze that rattled the industry.

The U.S. Department of Commerce lifted export controls on Anthropic's Claude Fable 5 and Mythos 5 models, ending an 18-day freeze that had blocked foreign nationals — including the company's own employees — from accessing the systems.

"Over the past two weeks, we have worked closely with Anthropic to analyze and approve Fable 5 to ensure alignment across the US Government and strengthen America's leadership in AI," Commerce Secretary Howard Lutnick said in a post on X.

Anthropic will begin restoring access to Fable 5 on Wednesday, the company said. Mythos 5, a more powerful version, was cleared last week for about 100 trusted organizations including federal agencies. The company agreed to continue collaborating on release protocols and to inform the government of any malicious activity detected on its platforms.

The resolution removes a regulatory overhang that had threatened Anthropic's commercial rollout and directly benefits Amazon, a major investor in the company, and Broadcom, a likely chip and infrastructure partner. Amazon shares rose in early trading on the news.

The standoff began June 12, three days after Anthropic launched Fable 5 — its first publicly available Mythos-class model — and Mythos 5, which it touted as state-of-the-art across multiple industry benchmarks. The Commerce Department ordered Anthropic to suspend all access for any foreign national, citing national security concerns over a potential jailbreak vulnerability that researchers at Amazon had discovered.

Anthropic disputed the severity of the vulnerability but shut both models down globally, arguing it was technically impossible to filter users by nationality in real time. The shutdown disrupted enterprise clients across AWS Bedrock, Google Cloud, Microsoft Foundry, and direct Claude APIs.

The company dispatched co-founder Tom Brown to Washington to lead negotiations, replacing CEO Dario Amodei, who had been a vocal supporter of Kamala Harris in the 2024 election and a frequent target of administration criticism. Lutnick's letter lifting the controls was addressed to Brown, not Amodei.

Competitive Pressure From China

The freeze handed valuable time to Chinese open-source developers. Zhipu AI released a system last month that security researchers said matched Mythos in its ability to find security bugs. A coalition of information security leaders published an open letter calling for the controls to be lifted, warning that ad-hoc regulation threatened America's AI lead.

Sam Altman, CEO of OpenAI, labeled the government's desire for phased model launches with direct government approval as "bad news." OpenAI itself agreed last week to limit access to its GPT-5.6 model after a call from the administration.

Investment Impact

The deal removes a key uncertainty for Amazon, which invested roughly $4 billion in Anthropic and stands to benefit from the company's continued commercial expansion. Broadcom, which supplies custom AI accelerators and networking chips for large-scale deployments, is positioned to capture infrastructure spending tied to Anthropic's model rollout.

Anthropic said it implemented a new method of targeting the cybersecurity workaround that had worried officials. The Commerce Department's Center for AI Standards and Innovation tested and supported the new guardrails. However, the company acknowledged that its systems are now less capable than before, with some requests being downgraded to less powerful models to avoid misuse.

This article is for informational purposes only and does not constitute investment advice.