Court Upholds Pentagon Authority to Blacklist Anthropic Over AI Safety Concerns

A federal appeals court has ruled that the Department of Defense has the authority to blacklist Anthropic from federal contracting, siding with the Trump administration's decision to exclude the AI company from military work after it refused to enable certain Claude features. The ruling, issued Tuesday by the US Court of Appeals for the District of Columbia Circuit, marks the latest development in an escalating legal and policy battle over how much control the government should exert over AI companies whose safety restrictions conflict with defense requirements.

The Ruling and Its Reasoning

In a 2-1 decision, the three-judge panel upheld the blacklisting under the Supply Chain Security Act, finding that Defense Secretary Pete Hegseth did not exceed his statutory or constitutional authority. The majority opinion acknowledged the gravity of the dispute, describing the case as raising "profoundly difficult questions about the appropriate military uses of an almost unimaginably powerful new technology."

The court laid out the core tension in stark terms. On one side, the government argued that overly constrained AI models risk shutting down unexpectedly during critical military operations. On the other, Anthropic warned that unconstrained models could hallucinate inappropriate targets for lethal force. The judges concluded that balancing these competing risks fell within the executive branch's authority.

"In doing so here, the Secretary did not transgress any limits on his authority under the Supply Chain Security Act or the Constitution," the ruling stated. "Accordingly, we deny the petitions for review."

The Judges Behind the Decision

The two judges who ruled against Anthropic were both appointed by President Trump and served in his first administration. Judge Gregory Katsas previously served as deputy counsel to the president, while Judge Neomi Rao held a position in the Trump administration's Office of Management and Budget. The dissenting judge offered no opinion in the publicly released ruling.

What Anthropic Refused to Do

Anthropic sued the Trump administration in March after the president ordered federal agencies to stop using Anthropic's products and banned defense contractors from conducting any business with the company. The dispute centered on Anthropic's refusal to remove safety guardrails that would allow Claude to be used for autonomous weapons systems or surveillance applications without meaningful human oversight.

The company maintained that its position was grounded in safety considerations rather than opposition to military use of AI. Anthropic's stance reflected its founding commitment to building AI systems that are helpful, honest, and harmless, even when doing so creates friction with government customers.

A Split Legal Landscape

The ruling creates a notable legal split. A separate federal court has already held that the government's parallel designation of Anthropic was unlawful, suggesting that the judiciary has not reached a consensus on whether the executive branch's actions exceeded its authority. This disagreement between courts increases the likelihood that the case will ultimately reach the Supreme Court.

Anthropic has indicated it may appeal, either by requesting an en banc review before all judges on the appeals court or by petitioning the Supreme Court directly. "We remain confident in our position and are considering all options, including further review," a spokesperson said.

The Political Dimension

Despite the ongoing legal battle, Commerce Secretary Howard Lutnick has publicly stated that the Trump administration and Anthropic have repaired their relationship and are "in tune." The comment suggests that while the legal confrontation continues, the practical dispute may be moving toward resolution through negotiation rather than litigation.

The case sits at the intersection of several unresolved policy questions: whether AI safety constraints can be overridden by national security needs, how far executive authority extends under the Supply Chain Security Act, and what role the judiciary should play in mediating disputes between technology companies and the military.

For now, the blackling stands, and Anthropic remains excluded from federal contracts and defense partnerships while the legal process plays out. The outcome will shape not only the relationship between AI companies and the Pentagon but also the broader framework for how AI safety commitments interact with government procurement power.