The US Court of Appeals for the District of Columbia has upheld the Trump administration's blacklist of Anthropic, granting the Pentagon authority to exclude the AI provider from its supply chain—even though Anthropic has no malicious intent.
Key Facts
- 2-to-1 ruling by the US Court of Appeals for the District of Columbia Circuit confirms the Trump administration's blacklist of Anthropic
- The Pentagon argues that overly constrained AI models could cause military operations to fail; Anthropic warns that unconstrained models might hallucinate inappropriate targets for lethal force
- The court upholds the Defense Department's authority under a less restrictive legal standard (41 U.S.C. § 4713) than a California district court had applied
- Anthropic announced it is considering further legal remedies, including a petition to the Supreme Court
The AI Safety Dilemma
The court's ruling captures a fundamental tension: the government warns that "overly constrained AI models shutting down unexpectedly" could cause military operations to fail, while Anthropic cautions that unconstrained models could "hallucinate inappropriate targets for lethal military force."
"Trump and Defense Secretary Pete Hegseth must determine how best to balance the competing risks," the court wrote, concluding that the Secretary of Defense "did not transgress any limits on his authority under the Supply Chain Security Act or the Constitution."
Both judges who ruled against Anthropic were Trump appointees who served in his first administration: Gregory Katsas, who was deputy counsel to the president, and Neomi Rao, who worked in the Trump administration's Office of Management and Budget.
Two Courts, Two Outcomes
Today's ruling directly contradicts an earlier decision by a California federal court. The US District Court for the Northern District of California had ruled last month that the blacklist was unlawful because Anthropic does not meet the definition of a supply-chain risk—limited to "the risk that an adversary may sabotage, maliciously introduce unwanted function, or otherwise subvert a covered system."
The DC appeals court did not dispute the district court's primary finding but argued that the two courts were applying different laws. The California court reviewed the decision under 10 U.S.C. § 3252, which restricts supply-chain risks to malicious actions. The appeals court reviewed it under 41 U.S.C. § 4713, which imposes fewer restrictions. The DC Circuit has exclusive jurisdiction over the latter statute.
What's Next?
An Anthropic spokesperson told CNBC: "We respectfully disagree with the court's decision. Another federal court has already held the government's parallel designation unlawful. We remain confident in our position and are considering all options, including further review."
Anthropic can request an en banc review by all judges on the appeals court or petition the Supreme Court. Notably, Commerce Secretary Howard Lutnick recently stated that the Trump administration and Anthropic have resolved their differences and are now "in tune."
Implications for Global AI Vendors
This decision could have significant consequences for European AI providers working with US government agencies or defense partners. The court has effectively recognized that security concerns about frontier AI models constitute legitimate grounds for government exclusion—regardless of malicious intent. German and European companies should anticipate that similar criteria may be applied by EU or German authorities. At the same time, the lack of clear definitions around specific security requirements creates substantial compliance uncertainty for AI vendors operating across borders.
Sources
Editorially owned by Ideal Syka. Sources and method: Newsroom & method. Tips and corrections: ai@i6eal.de.
