Federal Judge Rules Pentagon’s Blacklist of Anthropic Unlawful, Bars Contract Exclusion
A U.S. district judge vacated the Pentagon’s designation of AI firm Anthropic as a national “supply‑chain risk,” finding the move unlawful and blocking related federal penalties while the company’s challenge proceeds.
A federal judge in the Northern District of California issued a 59‑page order late Thursday vacating Defense Secretary Pete Hegseth’s decision to label Anthropic a national “supply‑chain risk” and barring enforcement of the Pentagon’s exclusion of the company from certain federal contracts. The judge concluded the government’s actions were unlawful and set aside the punitive measures while the litigation continues.
U.S. District Judge Rita Lin ruled that the government’s labeling of Anthropic amounted to unlawful retaliation tied to the company’s public criticism of Defense Department policy on AI use. Lin wrote that the decision was “illegal and baseless” and warned that “the empty invocation of national security is not a blank check to punish and retaliate against government critics.”
The dispute centers on negotiations earlier this year over using Anthropic’s Claude models for military applications. Anthropic told the court it would not permit certain uses — including support for lethal autonomous weapons and the use of its models for domestic mass surveillance — and sued after the Pentagon moved to bar the company from some government contracting pools in February.
Judge Lin’s order also halted other administrative steps tied to the designation. The ruling removes penalties imposed across multiple agencies and blocked enforcement of a prior directive that had instructed federal agencies to stop using Anthropic’s technology while the designation stood. The court emphasized that the ruling does not force the Pentagon to use Anthropic’s products; it prevents the government from using the supply‑chain label as a punitive tool.
The government has defended its actions in court filings, arguing that Anthropic’s contractual refusals could create operational uncertainty and risk for military systems. Officials are expected to appeal; the Justice Department has previously litigated a related and narrower challenge in the U.S. Court of Appeals for the D.C. Circuit, where a separate question about the statutory process for supply‑chain labels remains pending.
The ruling is the latest flashpoint in a broader national debate about how the U.S. government and private firms should manage the military uses of advanced AI. Anthropic executives and supporters in the tech sector have argued that companies should be able to impose ethical guardrails; Pentagon officials have argued that contractors cannot unilaterally constrain military uses that may be lawfully ordered later.
Legal analysts say Lin’s decision rests on two intertwined theories: that the record shows the designation was motivated by disapproval of Anthropic’s viewpoints, creating a First Amendment retaliation problem, and that the government failed to follow the statutory and procedural limits on applying a supply‑chain designation. The case will likely move quickly on appeal, preserving its role as a test of executive authority over procurement and national‑security labeling.
Why it matters
The ruling sets an important legal limit on the executive branch’s ability to use obscure procurement‑safety statutes and national‑security rhetoric to punish or exclude U.S. technology firms for policy disagreements; it also reshapes the practical terrain for Pentagon procurement, private sector cooperation on AI safety, and congressional oversight of tech‑national security disputes.
What to watch
Expect an expedited government appeal in the coming days and continued litigation in a parallel D.C.‑circuit case; Congress and industry groups may press for hearings or legislative fixes that would clarify how and when the government can use supply‑chain designations against domestic firms and what procedural protections vendors must receive.