A federal appeals court has allowed the Anthropic Pentagon blacklist to remain in place, handing the Defense Department a major legal victory in its dispute with one of the country’s leading artificial intelligence companies. The U.S. Court of Appeals for the District of Columbia Circuit declined to block the Pentagon’s national-security supply-chain risk designation, leaving Anthropic exposed to restrictions affecting defense business while the underlying litigation continues.
The 2-1 decision marks a significant reversal in momentum after Anthropic won a separate ruling in California federal court last month. The appellate panel concluded that the Pentagon had a sufficient national-security basis for the designation after Anthropic refused to remove safeguards limiting the use of its Claude models for autonomous weapons and mass domestic surveillance.
Anthropic Pentagon Blacklist Survives Appeal
The D.C. Circuit’s Sept. 25 decision left the Pentagon’s supply-chain risk designation in effect while Anthropic continues challenging the government’s action. The majority rejected the company’s request for emergency relief and found that the Defense Department had articulated a national-security rationale tied to the availability and reliability of AI systems used by the military.
The ruling does not end Anthropic’s case on the merits, but it allows the designation to keep operating during the litigation. That distinction is important because Anthropic has argued that the designation is already causing substantial commercial harm, including lost defense-related revenue and uncertainty among private customers that do business with the federal government.
American News Brief previously reported that a California federal judge struck down a parallel Pentagon action against Anthropic on Aug. 27. That earlier ruling found problems with the government’s use of a separate legal authority, meaning the latest appellate decision does not erase Anthropic’s California victory but leaves a different Pentagon designation intact.
Why the Pentagon Blacklisted Anthropic
The dispute began after months of negotiations over how the military could use Anthropic’s Claude models. The company said it was willing to support national-security missions but would not remove restrictions designed to prevent its systems from being used for fully autonomous weapons or broad domestic surveillance of Americans.
Defense Secretary Pete Hegseth designated Anthropic a supply-chain risk in March after those negotiations broke down. The administration argued that a military supplier could not impose restrictions that might interfere with lawful military operations, while Anthropic maintained that its safeguards reflected technical and civil-liberties concerns rather than opposition to defense work.
Anthropic has said current AI systems are not reliable enough to make autonomous lethal decisions without meaningful human control. The company has also objected to the use of its technology for mass domestic surveillance, arguing that such uses create risks that go beyond ordinary contractual disagreements over how software is deployed.
The Pentagon has taken the opposite view on the central procurement question. Its position is that the government must be able to rely on critical technology suppliers under lawful operational requirements, particularly when those systems could become embedded across intelligence, logistics and weapons-related infrastructure.
Two Courts Are Pulling in Different Directions
The most important legal complication is that Anthropic is now dealing with two related but distinct government actions under different statutes. U.S. District Judge Rita Lin in San Francisco blocked one designation in August, while the D.C. Circuit has now permitted the separate supply-chain risk action to continue.
That split does not necessarily mean the courts disagree about every underlying fact. Each case involves different statutory authority, procedural requirements and legal claims, so a government action can fail under one law while surviving review under another.
The latest appellate majority also rejected Anthropic’s argument that the Pentagon was retaliating against the company for protected speech. The government has maintained that the dispute arose from Anthropic’s refusal to accept military-use terms rather than from public criticism of the administration, while Anthropic has argued that officials used punitive measures after the company defended its safeguards publicly.
Those competing arguments will remain central as the litigation proceeds. The appeals court’s refusal to block the designation at this stage is not a final determination of every constitutional or administrative-law claim Anthropic has raised.
What the Ruling Means for Military AI
The case has become an important test of who sets operational boundaries for advanced AI used by the federal government. Technology companies can build safeguards into their products, but the Pentagon controls procurement standards and can decide whether a supplier’s restrictions make its technology unsuitable for sensitive military use.
That tension is likely to grow as AI systems become more capable and more deeply integrated into defense networks. Questions that once concerned ordinary software licenses now touch weapons autonomy, intelligence analysis, surveillance and the reliability of systems that may operate under battlefield conditions.
The commercial stakes are also substantial. Anthropic executives previously told a federal court that the government’s actions could reduce 2026 revenue by multiple billions of dollars if contractors and other customers interpret the blacklist broadly, with hundreds of millions of dollars in defense-related business potentially at risk.
The Pentagon has nevertheless left open limited exceptions when Anthropic technology is considered mission-critical and no viable alternative exists. That carve-out underscores the practical difficulty of removing a widely used AI platform from complex systems while the legal dispute is still unresolved.
Anthropic Can Still Seek Further Review
Anthropic said it respectfully disagreed with the appellate decision and was considering its legal options. Those options can include asking the full D.C. Circuit to rehear the case or eventually seeking Supreme Court review, although neither step guarantees that the designation will be suspended while litigation continues.
The company also continues to benefit from the separate California ruling, which limits part of the government’s earlier effort to exclude it from federal business. The result is a fragmented legal landscape in which one Pentagon action has been blocked while another remains enforceable.
For the Defense Department, the appellate ruling preserves leverage in negotiations over how AI vendors structure safety restrictions for military customers. For Anthropic, it raises the cost of maintaining safeguards that the company says are essential to responsible deployment, while intensifying pressure from investors and commercial partners that want certainty about its federal standing.
The Anthropic Pentagon blacklist fight is therefore no longer only a procurement dispute. It has become a broader test of how much control private AI developers can retain over the downstream use of their systems when the customer is the U.S. military, and how far the government can go in penalizing a supplier that refuses to remove those limits.
