Judge sides with Anthropic, temporarily blocks Pentagon's ban on working with the company

A federal court has temporarily blocked the Pentagon's ban on working with Anthropic. Rita F. Lin, a judge of the US District Court for the Northern District of California, granted the company's motion for a preliminary injunction. That suspends the US Department of Defense's decision to blacklist the maker of Claude for the duration of the litigation. The ruling takes effect in seven days.

The dispute has run for several weeks. The US Department of Defense (officially renamed the Department of War) designated Anthropic a "supply chain risk," a legal mechanism that effectively shuts a company out of all federal contracts. The label is normally applied to vendors with foreign ties or serious information-security violations. Anthropic's case was fundamentally different.

According to the Department of Defense's own documents, Anthropic was deemed a "risk" because of its "hostile manner in the press." The company had publicly criticized the military's position on the terms of government contracts, including its approach to using AI in lethal autonomous systems and mass surveillance tools. That public dissent is what landed the company on the blacklist.

In her ruling, Judge Lin called the actions a classic violation of the First Amendment to the US Constitution. "Punishing Anthropic for bringing public attention to the government's contracting position is classic unlawful restriction of free speech," she wrote. The court found that the government used the contracting mechanism as a tool to suppress inconvenient public statements, which directly contradicts constitutional guarantees.

Anthropic is one of the world's leading AI developers and the creator of the Claude family of models. The company was founded by former OpenAI employees and presents itself as an organization that puts AI safety first. For several years, Anthropic has spoken publicly about ethical limits on using its models for military and government purposes. That stance earned it a reputation as one of the most responsible developers in the industry, but it also created friction with the military.

A preliminary injunction is a significant but interim win. The full proceedings will continue, and the final decision may differ. Courts grant preliminary injunctions only when they see real grounds for the claim, so the fact that the court sided with Anthropic at this stage points to the strength of its constitutional arguments.

The case sets an important precedent for the entire AI industry. A tech company openly confronted the military in court and won the first round. As governments around the world draft rules for dealing with AI vendors, the ability to defend the right to public criticism in court becomes a strategically important tool. Whatever the final outcome, the episode has already entered the history of AI regulation.