Artificial-intelligence lab Anthropic sued the U.S. Defense Department on Monday to overturn the Pentagon’s designation of the company as a “supply chain risk,” a label the company says was imposed after it refused to allow unrestricted military use of its AI system, Claude.
The dispute began when the Defense Department demanded Anthropic remove safety guardrails that prevent uses such as autonomous weapons and domestic surveillance. Anthropic CEO Dario Amodei warned Defense Secretary Pete Hegseth that deploying untested AI in autonomous warfare would be dangerous and declined to lift the restrictions before a February 27 deadline.
The Pentagon responded that private technology firms should not dictate defense policy. Hegseth posted on X that U.S. troops “will never be held hostage by the ideological whims of Big Tech.”
Anthropic says the supply-chain-risk designation is unprecedented and unlawful and has filed lawsuits in California, where it is based, and in Washington, D.C., asking courts to vacate the label and block its enforcement. In its California complaint, reported by the Wall Street Journal, Anthropic accused the government of “seeking to destroy” the company’s economic value.
The company argues that even advanced AI models are not reliable enough for automated weapons and that using its technology for surveillance would threaten fundamental rights. The Pentagon, by contrast, says it needs full access to AI functionality for “any lawful” use and views Anthropic’s refusal as a private company imposing policy constraints on defense operations.
Observers say the designation is notable because Anthropic is the first U.S.-based firm to receive the “supply chain risk” label, which had previously been applied largely to foreign companies seen as security threats, such as Huawei. Under U.S. law, the term covers systems that could “sabotage” or maliciously introduce unwanted functions.
The designation effectively bars Anthropic from doing business with federal agencies and could ripple to government contractors and suppliers. President Donald Trump announced a government-wide ban on Anthropic technology, criticizing the company as run by “left wing nutjobs.” Anthropic’s Amodei said the label has a “narrow scope” and that many businesses can still use Anthropic tools for non-Defense projects.
Anthropic had been negotiating usage limits with the Pentagon after signing a $200 million contract in July 2025; that agreement was canceled amid the dispute. Despite the legal fight and the risk designation, Claude reportedly remains embedded in some Defense Department intelligence systems, and U.S. media have said the model played a substantial role in planning a recent U.S.-Israel attack on Iran.
Edited by: Jenipher Camino Gonzalez