Judge's gavel on a wooden block, symbolizing court ruling against Anthropic's attempt to block Trump blacklisting.

Editorial illustration for Court Denies Anthropic's Bid to Block Trump Blacklisting

Court Denies Anthropic's Bid to Block Trump Blacklisting

By AI Daily Post Edited by Brian Petersen, Editor-in-Chief

September 25, 2026 • 3 min read

A federal appeals court sided with the Trump administration on Thursday, ruling that the government can keep Anthropic off a list of approved AI suppliers for the military, even without proof the company acted in bad faith. The 2-1 decision from the US Court of Appeals for the District of Columbia Circuit ends, for now, a fight that started in March when Anthropic sued after Trump and Defense Secretary Pete Hegseth barred federal agencies from using its products and blocked defense contractors from buying them too.

At issue is a set of Claude features the Pentagon wanted for battlefield use, ones Anthropic refused to build in. The company argued that withholding those features was a legitimate safety call, not grounds for blacklisting under the Supply Chain Security Act. The two judges who ruled against Anthropic, Gregory Katsas and Neomi Rao, both served under Trump's first administration, Katsas as deputy White House counsel and Rao at the Office of Management and Budget.

The same court had already turned down Anthropic's request for an emergency stay back in April, so this ruling confirms a pattern rather than reversing one. The panel's reasoning gets into just how thorny the underlying military tradeoffs actually are.

A US appeals court today approved the Trump administration’s blacklisting of Anthropic technology. Judges decided the US had authority to blacklist Anthropic for withholding certain AI features even if Anthropic had no malicious intent.

Court rules Trump can blacklist Anthropic for refusing to enable Claude features - Ars Technica AI

Why this matters

This ruling gives the executive branch real teeth to compel AI companies to build specific capabilities into their models, not just refrain from harmful ones. Anthropic wasn't accused of sabotage or bad faith. It just declined to enable certain features, and that refusal alone was enough to trigger a blacklisting the D.C.

Circuit just upheld 2-1. For developers and founders building on Claude or any frontier model, the precedent matters more than the specific dispute: compliance with government requests for military or defense-adjacent functionality may no longer be optional, even when a company has legitimate safety or policy objections. The dissent presumably saw this differently, and the panel's own language about "profoundly difficult questions" suggests even the judges aren't fully comfortable with where this leaves things.

Worth watching whether Anthropic appeals further, and whether other labs facing similar pressure decide it's easier to just build what Washington asks for rather than risk their own blacklisting fight. The two judges who sided against Anthropic both served in Trump's first administration, which will fuel arguments about how politically insulated this kind of tech policy really is.