Anthropic, the AI safety company founded by former OpenAI researchers, just had its two newest models yanked offline globally. The Trump administration ordered that access to Fable 5 and Mythos 5 be restricted exclusively to US persons, citing national security concerns.
Anthropic complied, disabling the models worldwide on June 13, 2026. The company characterized the directive as the result of a misunderstanding and expressed hope for a quick resolution.
From cold shoulder to full freeze
This latest move is not an isolated incident. It is the sharpest escalation yet in a feud that has been building since February 2026, when President Trump ordered federal agencies to stop using Anthropic’s technology altogether.
The Pentagon followed up by designating Anthropic as a “supply chain risk.” In English: the Department of Defense essentially told every government contractor that doing business with Anthropic could jeopardize their own government relationships.
The root cause of the conflict is straightforward. Anthropic refused to give the Pentagon unrestricted access to its AI technology for military applications. CEO Dario Amodei pointed to the company’s AI safety policies as justification, particularly around autonomous weapons and ethical safeguards.
Anthropic escalated things on its own end by filing a lawsuit against the Defense Department in March 2026. The company essentially argued that being blacklisted over safety concerns was both legally questionable and strategically counterproductive for the country’s AI ambitions.
Defense Secretary Pete Hegseth and President Trump have both taken to social media to criticize Anthropic, claiming the company’s stance jeopardizes national security.
The competitive fallout
Rival firms like OpenAI moved quickly to fill the vacuum. After restrictions on Anthropic were enacted, competitors secured Pentagon contracts or expanded their existing government access.
Anthropic’s entire brand is built on the premise that AI development should be cautious, deliberate, and subject to ethical guardrails. The company was literally founded because its leaders felt OpenAI was moving too fast. Now that caution is being weaponized against it, while less safety-focused competitors reap the benefits.
Dario Amodei has consistently maintained that responsible AI development and national security are not in conflict. The administration has framed the company’s safety-first approach as obstruction rather than principle.
The global shutdown of Fable 5 and Mythos 5 adds another dimension. International users and businesses that had integrated these models into their workflows are now collateral damage in what is fundamentally a domestic policy dispute.