Appeals Court Maintains Anthropic’s Supply-Chain Risk Label Despite Lower Court Block

A federal appeals court has ruled to maintain Anthropic's designation as a supply-chain risk for now, denying the AI company's request to pause the Pentagon's label while a broader government ban on its Claude model remains blocked by a lower court judge. This decision, reported by Bloomberg, leaves Anthropic in legal limbo amid conflicting rulings, with the U.S. military facing uncertainty over using the technology.^[2]

The saga began earlier this year when the Trump administration, led by Defense Secretary Pete Hegseth, invoked a rare military authority under 10 USC 3252 to declare Anthropic a supply-chain risk. According to Anthropic's own statements, this followed the company's refusal to allow unrestricted military applications of Claude, prompting what the firm described as an "unlawful campaign of retaliation."^[3] In March, U.S. District Judge Rita Lin in California issued a temporary block on the designation and on President Trump's directive barring federal agencies from using Claude, calling the measures "arbitrary and capricious" and potentially crippling to the San Francisco-based company.^[1]

Lin's ruling specifically halted the Pentagon's "broad punitive measures," which she said appeared designed to punish Anthropic rather than address legitimate supply-chain concerns, as the Department of War could simply stop using Claude if needed.^[1] Anthropic had sued, seeking an emergency order to remove the "stigmatizing" label, arguing it was unprecedented and mismatched with the narrow statute intended to protect government contracts, not broadly punish suppliers.^[2][3] The company emphasized that the designation affects only Claude's use in direct Department of War contracts, leaving most customers unaffected.^[3]

Now, a separate federal appeals court in Washington, D.C., has declined Anthropic's bid to pause the risk label pending further review, creating a split with Lin's decision.^[1] As Wired reports, this conflicting stance means the U.S. military cannot clearly determine if or how it can integrate Claude, heightening operational uncertainty.^[1] The Pentagon had defended its position, insisting it could use the AI as it deems lawful, while Hegseth and officials publicly pressured Anthropic via social media to accept "all lawful" uses.^[2]

This dispute underscores broader tensions over AI governance in national security, pitting a leading U.S. AI firm against its own government. Anthropic, known for its safety-focused models, affects federal contractors, employees, and potentially military operations reliant on advanced AI. A narrower case remains pending in the D.C. appeals court, which could clarify the label's scope, while Lin's block on the wider ban holds for now.

The outcome matters for the AI industry, as such designations could deter partnerships and signal risks in refusing military applications. Stakeholders, from defense contractors to Anthropic's commercial users, await resolution, with no set timeline for full appeals. Until then, the limbo persists, balancing supply-chain protections against what courts have deemed potentially overreaching retaliation.