Key Points
- On March 9, 2026, Anthropic initiated two separate legal actions targeting the Pentagon and additional federal entities
- Defense officials classified Anthropic as a threat to supply-chain security following the company’s decision to maintain AI safety restrictions
- Trump issued a directive banning Claude usage across all federal government operations
- The lawsuits claim government actions breach First Amendment protections and due process requirements
- OpenAI secured a Pentagon contract following Anthropic’s exclusion from federal work
The AI developer Anthropic initiated legal proceedings against the Defense Department and multiple federal entities this Monday following its placement on a national security exclusion list.
The legal challenge consists of two distinct filings — one submitted to the Northern District of California and another to the D.C. Circuit Court of Appeals. Both filings contest the government’s characterization of Anthropic as a “supply-chain risk.”
The conflict emerged from disagreements about military applications of Anthropic’s AI platform, Claude. Defense officials sought unrestricted access for “any lawful use.” Anthropic maintained its position on safety protocols that prevent deployment for autonomous weaponry or domestic monitoring activities.
Defense Secretary Pete Hegseth issued the formal supply-chain risk designation on February 27. Anthropic received official notification on March 3.
Trump expanded the restrictions through social media posts, directing all federal agencies to cease Claude usage, extending well beyond the Pentagon’s initial scope.
The company characterized government actions as “unprecedented and unlawful,” stating its “reputation and core First Amendment freedoms are under attack.” According to Anthropic, these measures constitute retaliation for protected expression rather than legitimate national security concerns.
“The Constitution does not allow the government to wield its enormous power to punish a company for its protected speech,” Anthropic stated in court documents.
Financial Impact in the Hundreds of Millions
According to the company, the designation currently threatens “hundreds of millions of dollars” in commercial relationships. Over the previous year, the Defense Department executed contracts valued at up to $200 million with leading AI developers, including Anthropic, OpenAI, and Google.
Wedbush analyst Dan Ives cautioned that the blacklist designation may lead enterprise customers to suspend Claude implementations during the litigation process.
Dario Amodei, Anthropic’s chief executive, expressed openness to AI-enabled weapons conceptually while emphasizing that existing AI capabilities lack sufficient precision for fully autonomous deployment. He clarified that the Pentagon designation applies narrowly and leaves non-Pentagon business relationships intact.
An internal communication from Amodei, later disclosed by The Information, suggested Pentagon decision-makers were influenced by Anthropic’s failure to offer “dictator-style praise to Trump.” Amodei subsequently issued an apology regarding the memo.
The Path Forward
Anthropic emphasized that filing lawsuits does not preclude continued dialogue with government representatives. Pentagon communications officials declined commentary on active litigation, while a Pentagon representative confirmed last week that direct negotiations between parties had ceased.
The second legal filing challenges broader supply-chain legislation that could expand restrictions beyond military agencies to encompass civilian government operations. The full extent of such designation awaits completion of an interagency assessment currently underway.
OpenAI revealed an agreement providing its AI capabilities to Pentagon systems following Anthropic’s exclusion. CEO Sam Altman highlighted alignment between Pentagon requirements and OpenAI’s positions regarding human supervision of weapons systems and rejection of widespread domestic surveillance.
Investors backing Anthropic are actively working to mitigate consequences stemming from the federal agency dispute.
