Anthropic sued the Pentagon and other federal agencies on Monday after the Pentagon announced last week that it would designate AI companies as threats to national security and ban their products from being used for defense purposes.
In late February, President Donald Trump announced that other federal agencies would also ban the use of Anthropic products.
Anthropic filed two lawsuits in the U.S. District Court for the Northern District of California and the District Court of Appeals for the District of Columbia Circuit, alleging that the federal government’s move to separate the company goes beyond a standard contract dispute and is instead an “unlawful retaliatory campaign.” The company said its “reputation and core First Amendment freedoms are under attack” and would work to prevent the Trump administration from enacting the ban.
Anthropic said the supply chain risk designation and messaging from the White House have already “put hundreds of millions of dollars at risk,” illegally disregarded necessary procedures and overstepped the president’s authority.
“Seeking a judicial review does not change our long-standing commitment to leveraging AI to protect national security,” a spokesperson for Anthropic told NBC News. “However, it is a necessary step to protect our business, customers, and partners. We will continue to pursue all avenues toward a resolution, including dialogue with the government.”
The lawsuit was filed after months of tense negotiations between the company and the Department of Defense over how to allow the military to use Anthropic’s advanced AI systems. Anthropic CEO Dario Amodei had asked the Pentagon to use the system for “all lawful uses,” but wanted stronger assurances from the Pentagon that its systems would not be used directly for lethal autonomous weapons or domestic mass surveillance.
In addition to the Department of Defense, Anthropic is suing several other federal agencies and their executives, including the Treasury, State, and Commerce departments. Anthropic named as defendants, among others, Treasury Secretary Scott Bessent, Health and Human Services Secretary Robert F. Kennedy Jr., and outgoing Homeland Security Secretary Kristi Noem.
Anthropic’s flagship AI system, Claude, is reportedly being used by the Department of Defense on classified networks to support intelligence assessments, targeting recommendations, and combat simulations as part of a partnership with data analytics company Palantir. Claude is used across federal agencies to assist with data analysis and other administrative functions, as well as consumer-facing chatbots.
Amodei said supply chain risk labels have historically been limited to untrustworthy foreign adversaries and affiliates in important industries, but have never been publicly applied to U.S. companies.
President Donald Trump announced on Truth Social that all federal agencies must “immediately cease use of Anthropic’s technology” after the two parties were unable to reach an agreement by the February 27 deadline set by the Department of Defense.
Shortly after, Secretary of Defense Pete Hegseth announced on X that he was directing the Pentagon to classify humans as a “supply chain risk to national security.” On Wednesday, Mr. Hegseth made good on his threat and formally notified Anthropic that the company was prohibited from doing business with the Department of Defense and its contractors for defense purposes.
