DC Circuit upholds Pentagon’s blacklisting of Anthropic, citing national security risks

Key points

• The US Court of Appeals for the DC Circuit ruled 2-1 that the Defense Secretary had the authority to blacklist Anthropic under 41 U.S.C. § 4713, which does not require evidence of malicious intent.
• The court found that Anthropic’s restrictions on Claude’s functionality, which prevented the model from performing tasks requested by the military, qualified as a 'supply chain risk' under the broader statutory definition.
• Judge Gregory Katsas and Judge Neomi Rao, both appointed by President Trump, formed the majority, while Judge Karen Henderson dissented, arguing the statute was intended for hostile actors, not contractors enforcing ethical restrictions.
• A separate ruling by Judge Rita Lin in the Northern District of California had previously struck down a parallel designation under 10 U.S.C. § 3252, finding it violated the First Amendment and required proof of malicious intent.
• Anthropic’s CEO, Dario Amodei, has faced public criticism from President Trump, and the company recently declined an invitation to a state dinner for Chinese President Xi Jinping.
• The court delayed the immediate effect of its decision to allow Anthropic time to seek rehearing or appeal to the Supreme Court.

Background

This dispute follows a failed negotiation in late 2025 between Anthropic and the Department of Defense over the deployment of Claude models on the military’s GenAI.mil platform. Anthropic sought assurances that its technology would not be used for autonomous weapons or mass surveillance, while the Pentagon demanded unfettered access. The conflict escalated in March 2026 when the DoD designated Anthropic a supply chain risk, leading to dual litigation in California and DC. Recent archive notes indicate Anthropic has been consolidating its product suite, merging Claude Cowork into Claude Chat, while the legal battle continues.

How outlets are covering it

Ars Technica emphasizes the legal distinction between the two statutes at play, highlighting how the DC Circuit’s exclusive jurisdiction over 41 U.S.C. § 4713 allowed it to uphold the ban despite the California court’s ruling on 10 U.S.C. § 3252. The outlet notes the political alignment of the majority judges with the Trump administration. CNBC focuses on the broader political fallout, including President Trump’s public criticism of Anthropic CEO Dario Amodei and the company’s exclusion from a state dinner. Both outlets agree on the 2-1 ruling but differ in emphasis: Ars Technica details the statutory interpretation, while CNBC highlights the personal tensions between the administration and Anthropic’s leadership.

Why it matters

This ruling sets a precedent for how the US government can restrict private AI companies based on ethical or operational disagreements, even without evidence of malicious intent. It highlights the growing tension between national security priorities and corporate ethical guidelines in the AI sector. The decision may influence future negotiations between tech firms and the military, potentially leading to more restrictive contracts or further legal challenges if companies refuse to comply with government demands for unrestricted AI access.

What to watch

Anthropic is expected to seek an en banc review by the full DC Circuit or petition the Supreme Court to overturn the decision. The company may also continue to pursue its case in the Northern District of California, where a parallel designation was struck down. The outcome of these appeals will determine whether Anthropic can resume business with the Department of Defense or if the blacklist will remain in place, potentially impacting the company’s revenue and the military’s access to advanced AI tools.