Appeals court upholds Pentagon’s designation of Anthropic as a supply chain risk

Appeals court upholds Pentagon’s designation of Anthropic as a supply chain risk

A federal appeals court upheld the Pentagon’s decision to label Anthropic a supply‑chain risk, allowing the Defense Department to continue limiting the AI company’s tools on military contracts. The ruling from the U.S. Court of Appeals for the D.C. Circuit lets the Pentagon remove Anthropic’s Claude models from its systems and bar contractors from using them on Defense work, The Washington Post reported.

Wired and Defense News (reporting from Reuters) said the court found the department had support to conclude that Anthropic’s built‑in usage restrictions could interfere with lawful, contract‑authorized tasks. The government argued those limits—tied to Anthropic’s policies against uses such as autonomous weapons or mass domestic surveillance—posed an unacceptable risk inside military systems. According to those accounts, the court rejected Anthropic’s claims that the designation violated its free‑speech and due‑process rights, treating the dispute as one over contract terms rather than retaliation for the company’s views.

Coverage differed on who wrote the opinion and which judges joined or dissented, but the bottom line in all accounts is the same: the designation stands, and Pentagon restrictions may remain in place.

The decision arrives alongside a separate track of litigation. In a separate California case, a federal judge struck down a parallel government action under a different statute and blocked a broader, government‑wide ban on Anthropic’s technology. This D.C. ruling concerns a separate supply‑chain security law and affects Defense work specifically, not the entire federal government, The Washington Post noted. Bloomberg Law likewise reported that the California order remains in effect even as the Pentagon’s designation survives in the D.C. case.

Anthropic said it respectfully disagrees with the ruling and is weighing its options, according to Wired, The Washington Post, and Bloomberg Law. Courthouse News Service’s account summarized parts of the majority opinion about why the department prevailed, noting instances where Anthropic’s guardrails may have blocked government tasks and concerns that the company could adjust the model in ways the Pentagon could not fully monitor inside its own systems.

The immediate practical effect is continuity. The Pentagon has been removing Anthropic tools and, with this ruling, can keep directing its agencies and contractors to swap out Claude for other approved AI systems, The Washington Post reported. That transition process—and any knock‑on impacts in existing Defense projects—may continue while appeals play out.

Why it matters

If you work for a defense contractor, this could mean audits of AI tooling, contract changes, and extra work to replace Anthropic‑based features with alternatives. More broadly, the ruling signals that, at least for now, the government has wide latitude to set AI terms in national‑security contracts—even when a vendor’s safety policies conflict. It’s also a test case for how Washington could police fast‑moving AI in real‑world procurement, not just in policy speeches.

September 26, 2026 (0)