Anthropic Takes Pentagon and White House to Court Over 'Supply Chain Risk' Label

Artificial intelligence developer Anthropic went to federal court on Monday to challenge what it describes as an "unlawful campaign of retaliation" by the Trump administration. The dispute stems from the company's refusal to allow unrestricted military use of its technology.

The San Francisco-based firm filed two separate suits, one in California federal court and one in the federal appeals court in Washington, D.C. Each targets a different part of the government's actions. Anthropic wants the courts to reverse the Pentagon's decision, made last week, to label it a "supply chain risk," and to undo President Donald Trump's order directing federal employees to stop using its chatbot, Claude.

"These actions are unprecedented and unlawful," the complaint states. "The Constitution does not allow the government to wield its enormous power to punish a company for its protected speech. No federal statute authorizes the actions taken here." The company says it is turning to the judiciary "as a last resort." The Defense Department declined to comment, citing its policy on pending litigation.

At the heart of the conflict is Anthropic's effort to bar the use of its technology for mass surveillance of Americans and for fully autonomous weapons. Defense Secretary Pete Hegseth and other senior officials publicly insisted that the company accept "all lawful" uses of Claude. They threatened punishment if it refused and criticized the firm and its CEO, Dario Amodei, on social media.

The supply chain risk designation ends Anthropic's defense work. It relies on an authority created to keep foreign adversaries from harming national security systems, and it is the first known instance of the government applying it to a U.S. company. According to the lawsuit, Hegseth wrote in a March 4 letter that the step was "necessary to protect national security." Trump also said federal agencies would stop using Claude, but gave the Pentagon six months to phase it out of classified military systems, including those used in the Iran war. The suit also names agencies such as the Treasury and State departments, whose officials told employees to stop using the product.

Michael Pastor, a professor at New York Law School and a former New York City general counsel who helped shape its technology contracts, said Anthropic has strong First Amendment and due process arguments in a case that has "escalated beyond comprehension." He added: "I've never seen a case like this."

The financial stakes are considerable. Anthropic is privately held and projects about $14 billion in revenue this year, most of it from businesses and government agencies using Claude for coding and other work. A recent investment announcement valued the company at $380 billion and noted more than 500 customers paying at least $1 million a year. Anthropic has tried to reassure clients that the designation is narrow and applies only to military contractors when they use Claude on Defense Department work. Its lawsuit alleges the government's actions are "jeopardizing hundreds of millions of dollars" in contracts.

Anthropic said its usage policy has long prohibited "lethal autonomous warfare without human oversight and surveillance of Americans en masse," and that it has never tested Claude for such purposes. It also noted that it has allowed uses by the military that are not open to civilians, including operations and analysis of lawfully collected foreign intelligence. Until recently it was the only company in its peer group approved for classified military systems. The Pentagon is now looking to move that work to Google's Gemini, OpenAI's ChatGPT and Elon Musk's Grok.

The fight has rippled through the industry. OpenAI struck its own Pentagon deal within hours of the government's move against Anthropic, which CEO Sam Altman later conceded was rushed and appeared opportunistic. Claude downloads surged, putting it ahead of ChatGPT and Gemini in popularity for the first time. OpenAI's head of robotics, Caitlin Kalinowski, resigned over the deal, writing that surveillance without judicial oversight and lethal autonomy without human authorization "deserved more deliberation than they got."

On Monday, more than 30 AI developers from OpenAI and Google, including Google chief scientist Jeff Dean, also filed a legal brief backing Anthropic, saying they acted in their personal capacities. They argued that national security is not served by "reckless designations" of American technology partners or by suppressing public discourse on AI safety. Anthropic said seeking judicial review does not change its commitment to using AI to protect national security, but is "a necessary step to protect our business, our customers, and our partners."