Anthropic has taken the US government to court following the Pentagon’s decision to put the artificial intelligence company on a national security blacklist. This is a dramatic shift in the escalating war between AI developers and the US military over the level of advancement in AI systems.
The company filed the lawsuit in federal court in California on Monday. It argues that the Pentagon acted outside the law when it labeled Anthropic a “supply-chain risk.” The designation limits how the U.S. military and its contractors can use Anthropic’s AI systems, including its Claude model.
Anthropic says the move violates its constitutional rights. In the filing, the company claims the government punished it for refusing to loosen safety rules on its technology. Those rules block the use of its AI for fully autonomous weapons and for domestic surveillance.
The company sought to have this designation canceled by the court and to enjoin federal agencies from enforcing it.
In its public statement, Anthropic said, “The government should not use its power to penalize a company for its speech or its policies.” The company asserts that it has constitutional rights to regulate how its technology is used.
The dispute began when there were tense negotiations between the Defense Department and the company for months. The Defense Department wanted greater access to the AI technology and fewer restrictions on its use in warfare operations by the military.
Pentagon vs. Anthropic, The Fight for AI Control
Defense Secretary Pete Hegseth approved the blacklist decision last week. According to reports, the Pentagon had already used Anthropic’s AI in some military work linked to operations involving Iran. Officials argued that strict limits on AI use could weaken military capability.
The Pentagon has not commented on the lawsuit. A defense official said earlier that the U.S. government must have full flexibility to use AI tools for any lawful purpose. The department believes a private company should not dictate how the country defends itself.
The clash has drawn attention across the tech sector. It raises a larger question about control over powerful AI systems. Governments want access to new technology for defense and intelligence. AI companies want to place limits on how their systems operate.
Anthropic has said current AI models are not reliable enough to run autonomous weapons. The company believes machines still make too many mistakes. It argues that using them without human control could cause serious harm.
The blacklist could have major business effects for the firm. Government contracts form a large part of the AI market. Anthropic told the court that the designation may cut billions of dollars from its expected revenue in 2026.
Company executives also warned that the damage could spread beyond defense work. Some private firms may delay or cancel projects that rely on the Claude model while the legal fight continues. Analysts say the uncertainty could slow adoption among large enterprises.
Court filings already show signs of impact. One partner with a large annual contract has switched from Claude to another AI system. That decision removed more than $100 million from Anthropic’s expected sales pipeline. Negotiations with financial institutions worth about $180 million have also stalled.
Anthropic’s Legal Battle Over Federal Oversight
The legal conflict expanded further on Monday. Anthropic filed a second case in a federal appeals court in Washington, D.C. This lawsuit challenges a broader supply-chain risk label that could extend across the entire civilian government.
If that review moves forward, many federal agencies could be forced to stop using the company’s AI tools.
Support for Anthropic has come from within the ranks of the AI research world. A number of engineers and scientists from prominent AI labs have filed a brief in support of the company.
The argument is that the government’s pressure could stifle debate about the safety of AI.
The fear is that the government’s punishment of the company could silence others from talking about the risks related to artificial intelligence.
The outcome could determine the future relationship between the government and AI firms in the U.S.
It could determine the extent to which the government can control these firms and how they can control the use of the technology in warfare and surveillance. For now, the case is a sign of how far artificial intelligence has progressed from the research phase to the forefront of security issues.
