DC Circuit Upholds Pentagon's Anthropic Risk Label, 2-1

The U.S. Court of Appeals for the District of Columbia Circuit voted 2-1 on Friday to uphold the Pentagon's "supply chain risk" designation of Anthropic, a classification that bars the Defense Department and its contractors from using Claude. A separate ruling on a related designation, issued by a federal judge in San Francisco in August, remains in Anthropic's favor.

Circuit Judges Gregory Katsas and Neomi Rao formed Friday's majority; Circuit Judge Karen LeCraft Henderson dissented. Katsas wrote that "the Department had ample support for its conclusion that the continued integration of Claude into the Department's information systems ... presented a statutorily covered national-security risk," according to TheNextWeb and Investing.com, which both reviewed the opinion.

Two designations, two courts, one split outcome

The Pentagon issued its risk label on Anthropic in March 2026, and the company sued within days, according to TheNextWeb. The department had relied on two separate legal designations to justify the block, and the two ended up litigated in two different courts: the D.C. Circuit case decided Friday, and a second case a federal judge in San Francisco ruled on in August 2026, striking that designation down as unlawful, according to Investing.com.

Designation	Court	Outcome	Date
First (D.C. Circuit case)	U.S. Court of Appeals, D.C. Circuit	Upheld, 2-1	Sept. 25, 2026
Second (N.D. Cal case)	Federal court, San Francisco	Struck down	August 2026

That split means Friday's ruling does not fully resolve Anthropic's Pentagon business. The San Francisco order remains in effect for the designation it covers, Investing.com reported, so the practical question of whether Claude can reach military systems still depends on which designation a given contract falls under.

What the March 2026 label actually bars

The underlying dispute is over Anthropic's contract terms, which prohibit deploying Claude for lethal autonomous weapons and mass domestic surveillance. Anthropic refused to drop those terms, and the Pentagon's Department of War argued no private company should dictate military operating rules, TheNextWeb reported. The appeals panel concluded the Pentagon reasonably feared Claude's built-in safeguards could fail during live operations, according to TheNextWeb's account of the ruling.

Anthropic Chief Executive Dario Amodei has defended keeping the restrictions in place, saying the company maintains "strict prohibitions against deploying Claude for mass domestic surveillance and lethal autonomous warfare," according to Investing.com. The outlet traced the dispute in part to a sensitive military operation in Venezuela in January 2026, after Secretary of War Pete Hegseth had mandated an "AI Dominance" push for the department earlier in the year.

What happens to the case next

An Anthropic spokesperson said Friday the company "remains confident in our position and is considering all options, including further review," according to TheNextWeb. Anthropic did not say whether it will petition the U.S. Supreme Court, which is the company's most direct path to challenging Friday's outcome rather than accepting it.

The ruling lands during an active week for Anthropic elsewhere. The company cut Claude Opus 5.5's price by 20 percent against Fable 5.1's benchmarks on Sept. 22, and it had already pulled its Claude for Financial Advisors plugin from GitHub without explanation on Sept. 21, leaving a Sept. 23 bug report about the removal unanswered as of Friday.

The San Francisco ruling on the second designation was not before the D.C. Circuit and is untouched by Friday's decision, Investing.com reported. Whether the Pentagon appeals that separate loss, or Anthropic seeks Supreme Court review of this one, is the next thing to watch.