The legal fight over how far the federal government can go in policing artificial intelligence companies has become one of the most closely watched disputes in Washington. That fight sharpened on August 27, when a federal judge blocked the Trump administration from enforcing its effort to blacklist Anthropic, the AI company behind Claude, after President Donald Trump had ordered agencies to stop using the firm’s products. The ruling carries national implications because Anthropic’s software had been cut off across federal agencies and military contracting channels.
Judge blocks blacklist and agency cutoff order
U.S. District Judge Rita Lin ruled Thursday night that the Trump administration acted unlawfully when it punished Anthropic and labeled the company a national security or supply chain risk, according to the written order described by the Associated Press and coverage of the case by The Washington Post. The decision followed earlier temporary relief issued in March, when Lin had already blocked enforcement of Trump’s directive telling federal agencies to stop using Anthropic and its chatbot Claude.
The dispute began in late February, when Trump said agencies should “immediately cease” using Anthropic technology and Defense Secretary Pete Hegseth separately moved to classify the company as a supply chain risk, according to previous reporting on the administration’s actions. Anthropic then sued in federal court, arguing that the government had exceeded its legal authority and retaliated against the company for protected speech about limits on military uses of AI.
In her latest ruling, Lin said the Pentagon’s measures were “illegal and baseless,” according to the Associated Press. The order blocks the government from enforcing the blacklist while the broader legal fight continues, preserving Anthropic’s ability to do business with federal agencies and contractors for now.
The decision is national in scope because the challenged order was directed at federal agencies, not one state or one regional procurement office. That means the immediate confirmed impact is on government use of Anthropic products across the country, including any federal office, contractor, or military-related entity that had paused or avoided Claude because of the administration’s order.
What is not yet publicly clear is how many agencies had fully stopped using Anthropic tools before the injunction, or how many contracts were delayed, canceled, or revised because of the blacklist. Anthropic has publicly argued in court that the government’s actions caused broad business harm, but a comprehensive agency-by-agency list of affected federal users has not been released in public reporting.
For government vendors and technology teams, the ruling removes at least one immediate legal barrier to working with Anthropic while appeals and related litigation continue. A separate, narrower Anthropic case remains pending in the federal appeals court in Washington, according to the Associated Press, so the procurement and policy picture is not fully settled.
The administration’s break with Anthropic grew out of a larger conflict over military and national security uses of artificial intelligence. According to prior reporting in The Washington Post and the Associated Press, the company resisted government demands that its technology be available for uses including domestic surveillance and fully autonomous weapons, and the White House and Pentagon responded with unusually sweeping restrictions.
Anthropic argued that the blacklist was not a routine contracting dispute but a punishment for the company’s speech and policy positions. Legal experts quoted in earlier coverage said the government appeared to rely on authorities designed to protect federal systems from foreign-linked threats, not to sanction a U.S. company in a commercial disagreement over AI deployment terms.
For federal workers, contractors, and agencies, the practical takeaway is that the administration cannot currently enforce the Anthropic ban struck down by Lin’s order. The White House is expected to keep defending its broader AI security posture, while the case continues to test how much power the government has to cut off a major American AI supplier from public-sector use.