Anthropic has initiated legal action against the Trump administration, urging federal courts to overturn the Pentagon’s decision to label the artificial intelligence firm a “supply chain risk” after it declined to permit unrestricted military applications of its technology.
The company filed two separate lawsuits on Monday — one in a federal court in California and another before the federal appeals court in Washington, DC — with each case challenging different elements of the Defence Department’s actions.
The Pentagon formally applied the designation to the San Francisco–based technology company last week following a highly public disagreement over the potential wartime use of Anthropic’s AI chatbot, Claude.
In its court filings, Anthropic described the government’s move as unlawful and retaliatory. The company argued that constitutional protections prevent the government from using its authority to penalise a firm over protected speech and maintained that no federal law authorises the measures imposed. It said it had turned to the courts as a final remedy to defend its rights and stop what it characterised as executive overreach.
Anthropic said it had placed limits on how its technology could be used, specifically blocking applications related to mass domestic surveillance and fully autonomous weapon systems. Defence Secretary Pete Hegseth and other officials reportedly maintained that the company must permit all lawful uses of Claude and warned of consequences if it refused.
Being designated a supply chain risk effectively bars Anthropic from defence-related work under authorities typically used to shield national security systems from foreign threats. This marks the first known instance of the federal government applying the classification to a US-based company.
President Donald Trump also stated that he would direct federal agencies to discontinue use of Claude. However, the Pentagon has been given six months to phase out the system, which is currently integrated into classified military infrastructure, including platforms used in the Iran conflict.
While contesting the Pentagon’s actions, Anthropic has reassured clients that the restriction is limited in scope and affects only military contractors using Claude for Department of Defense work.
Clarifying this boundary is critical for the privately held company, as the bulk of its projected 14 billion dollars in annual revenue comes from commercial enterprises and civilian government agencies that deploy Claude for software development, automation and related tasks.
According to investment disclosures, more than 500 customers pay Anthropic at least one million dollars per year for access to Claude, and recent funding rounds have valued the company at approximately 380 billion dollars.
