Anthropic sued the Trump administration March 9, filing two federal lawsuits to overturn the Pentagon’s supply chain risk designation — the first time the government has used that authority against an American company. Two days later, the company filed an emergency motion in the D.C. Circuit asking the court to freeze the order while the case unfolds. The first hearing is set for March 24.
The 48-page complaint in the Northern District of California names the departments of Defense, Treasury, State, and Commerce as defendants, along with individual officials including Defense Secretary Pete Hegseth. A separate filing in the D.C. Circuit challenges the designation under the federal statute that governs judicial review. Anthropic argues the government violated its First Amendment rights by retaliating against the company for its public position on AI safety, denied it due process under a statute that provides no notice or opportunity to respond, and exceeded the scope of a law designed to block foreign adversaries from national security systems.
The Pentagon maintains that restricting how the military uses a purchased tool is not a vendor’s decision to make. Officials have said the department requires AI for “all lawful purposes” and cannot allow a private company to dictate operational boundaries during a national security emergency.
The legal exposure extends beyond defense work. Hegseth directed that no contractor doing military business may conduct any commercial activity with Anthropic. Amazon and Google are both Pentagon contractors and Anthropic’s cloud infrastructure providers. When the government imposed comparable restrictions on Huawei, it required an act of Congress. Anthropic says the formal designation it received applies only to Claude’s use in Pentagon contract work.
CFO Krishna Rao filed a statement saying the government’s actions could reduce Anthropic’s 2026 revenue by multiple billions of dollars. The company projects $14 billion this year from more than 500 customers paying at least $1 million annually. Federal contracts are already being canceled.
Claude remains operational on classified military networks supporting the Iran war during a six-month phaseout. Pentagon officials acknowledged that replacement models from OpenAI and xAI are not as capable in those applications.
OpenAI struck its own Pentagon deal hours after Anthropic was banned, claiming the same restrictions Anthropic had demanded. The contract language appears to allow “all lawful use” with restrictions tied to existing law the government can change. OpenAI later acknowledged the announcement looked “sloppy and opportunistic.” A senior robotics engineer resigned days later, citing the same surveillance and autonomous weapons concerns Anthropic raised.
Dozens of scientists from OpenAI and Google DeepMind filed an amicus brief supporting Anthropic, arguing the designation could chill public discussion about AI risks and harm American competitiveness. HCA reported the initial confrontation on February 28.
The March 24 hearing will be the first court test of whether procurement authority built for foreign threats can be used against a domestic company — and where the line falls between the Pentagon’s operational authority and a vendor’s right to set terms on its own product.
Support Independent Local Journalism — High Country Advocate was created as a real alternative to regional media that too often silences dissenting voices while taking sides in the political struggle. Producing in-depth, unflinching reporting like this series and others is expensive: servers, editing, research time, and legal review all add up quickly. If these articles have informed you or given you new perspective, please consider supporting HCA with a paid subscription — every subscriber helps keep this reporting strong and independent. 