A federal appeals court just handed the Pentagon a major win in its ongoing war with a leading AI company. The ruling could reshape who controls America’s military AI going forward.
A three-judge panel sided 2-1 with the Defense Department, upholding its decision to blacklist Anthropic from military contracts entirely.
The Fight That Split Two Federal Courts
The dispute traces back to February, when Defense Secretary Pete Hegseth demanded Anthropic strip safeguards blocking Claude from fully autonomous lethal weapons and mass domestic surveillance. Anthropic publicly refused on February 26, just hours before Hegseth’s deadline expired.
Hegseth formally designated Anthropic a national security risk on March 3. That designation canceled the company’s existing military contracts and barred other Pentagon contractors from using its technology as well.
According to the court’s decision, an Anthropic executive had separately questioned a contractor’s use of Claude during a specific military operation.
Media reports in the case record identified that mission as the January 3 operation that captured Venezuelan President Nicolás Maduro.
Defense officials cited that episode as evidence that Claude’s behavior remained too unpredictable for military reliance.
Judge Gregory Katsas, joined by Judge Neomi Rao, wrote the majority opinion. He said the Pentagon had “ample support” for its conclusion. Judge Karen LeCraft Henderson dissented.
Why Anthropic Says This Fight Isn’t Over
A separate California court reached the opposite conclusion just last month. Judge Rita Lin ruled a broader, parallel Trump administration ban unconstitutional, calling it retaliation meant to make “a public example” of Anthropic.
Anthropic responded firmly to Friday’s ruling. “We respectfully disagree with the court’s decision,” a spokesperson said, cited by CNBC. “Another federal court has already held the government’s parallel designation unlawful.”
Whether Anthropic seeks rehearing or takes the fight to the Supreme Court next remains unclear, leaving America’s newest AI battleground unresolved across two contradicting federal courts.
