New export controls forced the AI safety company to disable its frontier models for all users, escalating a months-long feud with the federal government Anthropic, the AI company co-founded by former OpenAI researchers with a stated mission of building safe AI, just had its two most powerful models effectively shut down by the US government. The company was forced to disable access to its latest frontier models, Fable 5 and Mythos 5, after the Trump administration imposed new export controls on June 12.
The directive came directly from Commerce Secretary Howard Lutnick to Anthropic CEO Dario Amodei. It prohibits foreign nationals, including those residing in the US, from accessing the models, citing national security risks tied to their capabilities.
Here’s the thing: Anthropic didn’t just cut off foreign users. The company disabled all access to the models to avoid penalties, a move that also knocked out US subscribers and even Anthropic’s own employees. The models had been available for only days before the plug was pulled.
A feud months in the making #
This didn’t come out of nowhere. The collision between Anthropic and the Trump administration has been building since February 2026, when federal agencies were prohibited from using any Anthropic products.
The reason? Anthropic was labeled a supply-chain risk. Not because of a security breach or data vulnerability, but because of its own policies. The company has maintained strict guardrails against military use of its technology, particularly for surveillance and autonomous weapons applications.
President Trump referred to Anthropic as a “radical left, woke company” in February 2026.
What the export controls actually mean #
Export controls on AI models represent relatively uncharted regulatory territory. Traditionally, these mechanisms have been used for physical goods: weapons, semiconductor equipment, nuclear materials. Applying them to software models, particularly ones delivered via cloud APIs, raises a different set of enforcement challenges. The new controls target Anthropic’s frontier-tier models specifically. By designating these models as controlled technology, the Commerce Department is treating them more like dual-use military hardware than consumer software.
For Anthropic, the compliance math was brutal. Rather than risk penalties by trying to verify the nationality of every user in real time, the company chose the nuclear option: shut everything down. That meant US customers who were paying for access lost it too, at least temporarily. Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our