Circuit Court of Appeals may lack the authority to review the Pentagon’s decision to designate Anthropic, a leading artificial intelligence company, as a national security risk. The statement came during oral arguments that could reshape the balance of power between the executive branch and domestic technology firms, as the Trump administration seeks to bar the company from government contracts while simultaneously racing to gain access to its most advanced AI model.
The dispute began on March 3, when Secretary of Defense Pete Hegseth and the Pentagon formally designated Anthropic a supply chain risk after the company refused to allow its AI tools to be used for mass domestic surveillance and in autonomous weapons systems. But the potential government-wide ban has been complicated by what lawyers described as an urgent scramble among federal agencies to obtain access to Anthropic’s Mythos model, a powerful system capable of finding and exploiting software vulnerabilities at a scale previously unseen.
Judges Gregory Katsas and Neomi Rao, both appointed by President Donald Trump, appeared skeptical that the court could intervene in the underlying designation itself. “This court can’t review designation decisions,” Rao said, adding that the only remedy available to the panel would be to declare that further actions excluding the company’s products from government use were unlawful. Katsas at one point asked government attorneys whether they wanted an “informal” pause in the case to continue negotiating a resolution with Anthropic.
The third member of the panel, Judge Karen Henderson, an appointee of President George H.W. Bush, offered a sharp counterpoint. She expressed deep skepticism about the Pentagon’s justification for targeting Anthropic, noting that the law invoked by the department was originally written to address companies tied to foreign adversaries. “I don’t see that the department has in any way supported its determination that Anthropic is a supply chain risk,” Henderson said. “To me this is just a spectacular overreach by the department.”
First Judicial Test of 2018 Law
Rao noted during the hearing that the courts are interpreting the relevant statute, passed in 2018, for the first time. A ruling in the administration’s favor could establish sweeping executive authority to label domestic AI and other technology companies as supply chain risks based on their corporate usage policies and safety restrictions, rather than evidence of ties to hostile governments. Lawyers for Anthropic argued that the designation itself carries significant “reputational stigma,” amounting to being described as the “equivalent of a saboteur.”
Katsas pressed Anthropic’s legal team on whether its current products truly avoided the risks described by the government. He voiced “conceptual difficulties” with the company’s position, adding that “AI three months from now will be totally different” from the products currently available. The comment underscored a central tension in the case: the rapid pace of AI development may outstrip the legal frameworks designed to govern it.
A decision overturning Hegseth’s orders would block the administration from cutting off Anthropic from government work and could set limits on how the executive branch applies national security designations to domestic companies. But a ruling in the government’s favor could give the White House broad new powers to effectively blacklist technology firms based on their internal policies, reshaping the relationship between Silicon Valley and the federal government for years to come.
