A California federal judge has reportedly temporarily blocked the Pentagon from designating AI company Anthropic as a supply chain risk, delivering what sources suggest one former Trump administration AI advisor called “a devastating ruling for the government” that exposes the structural gap between executive rhetoric and legal authority over the private AI sector.
What the court found
Reports indicate that a judge issued an opinion ruling that the government bypassed proper legal procedures and violated Anthropic’s First Amendment rights, as reported by MIT Technology Review. The court reportedly found the administration had “set out to publicly punish Anthropic for its ‘ideology’ and ‘rhetoric,’ as well as its ‘arrogance’ for being unwilling to compromise those beliefs.”
The government’s central claim — that Anthropic could implement a “kill switch” cutting off government access to its AI — reportedly had no evidentiary support, according to court documents.
The timeline tells the story
Reports suggest federal agencies used Anthropic’s Claude model throughout 2025 without complaint. The conflict reportedly only surfaced when direct contracting discussions began — at which point Anthropic’s usage policy became a sticking point.
Reports indicate President Trump posted on social media referencing the company and directed federal agencies to stop using the company’s AI. Defense Secretary Pete Hegseth reportedly then announced he would direct the Pentagon to label Anthropic a supply chain risk. Anthropic vowed to fight the designation.
Public statements outran legal authority
The sequence matters. The public declarations of punishment preceded the legal process required to enact it. Federal contracting rules have specific procedural requirements for removing contractors — requirements that reportedly were not followed here. Reports suggest the government itself admitted in court that the Secretary of Defense lacks the legal power to enforce a ban on contractors doing business with Anthropic.
Sources indicate that Dean Ball, who worked on AI policy for the Trump administration and wrote a brief supporting Anthropic, described the ruling as finding Anthropic “likely to prevail on essentially all of its theories for why the government’s actions were unlawful and unconstitutional.”
What this actually reveals
The case exposes a structural tension at the centre of the government-AI relationship. The executive branch depends on a small number of frontier AI companies for critical capabilities, yet the legal architecture governing federal procurement sharply limits the government’s ability to punish those same companies for policy disagreements.
As observers have noted, according to MIT Technology Review, there are “mechanisms the government can use to apply some degree of pressure without breaking the law.” The operative phrase being: without breaking the law. The court’s finding suggests the government chose the louder path first and the legal one not at all.
For AI companies globally, the ruling sets an early precedent: executive displeasure, however vocal, does not automatically translate into enforceable procurement sanctions. The machinery of government contracting has its own logic, and it does not bend easily to social media posts.
Feature image by Towfiqu barbhuiya on Pexels
