Anthropic Warns Pentagon Dispute May Result in Billions in Losses

Anthropic executives claim that existing and potential customers have been pressing for new terms and even withdrawing from negotiations since the US Department of Defense identified the AI startup as a supply-chain risk late last month, according to court documents that also disclosed new financial insights about the company.

The company risks losing hundreds of millions of dollars in anticipated revenue this year from projects associated with the Pentagon, chief financial officer Krishna Rao noted in a court filing on Monday. If the government enforces its stance and pressures a host of companies to cease any dealings with the AI startup, regardless of military ties, Anthropic could ultimately see billions in lost sales, he stated. The total sales since the commercialization of its technology in 2023 have surpassed $5 billion, Rao mentioned.

Anthropic’s revenue surged as its Claude models began surpassing competitors and demonstrating advanced capabilities in areas like software code generation. However, the company invests heavily in computing infrastructure and remains significantly unprofitable. Rao indicated that Anthropic has spent over $10 billion on training and deploying its models.

Anthropic’s chief commercial officer Paul Smith shared multiple specific instances where partners have recently expressed worries to the AI startup. He stated that a financial services client has paused negotiations on a $15 million deal due to the supply-chain classification, while two prominent financial firms have refused to finalize deals amounting to $80 million unless they secure the right to unilaterally cancel their contracts for any reason. Additionally, a grocery chain canceled a sales meeting, citing the supply-chain-risk designation, Smith revealed.

“All have made moves that signal deep distrust and a growing concern about associating with Anthropic,” Smith wrote.

These remarks from the executives are part of statements from six leaders at Anthropic supporting a preliminary order that would permit the San Francisco firm to maintain its business engagements with the Department of Defense until the lawsuits regarding the supply-chain-risk matter are resolved.

Anthropic has initiated legal action against the Trump administration in two courts. A lawsuit filed in the federal court in San Francisco on Monday claims the government infringed on the company’s free speech rights. A second case, filed the same day in the federal appeals court in Washington, DC, alleges the Defense Department has unfairly discriminated against and retaliated against Anthropic.

The company seeks a hearing as soon as Friday in San Francisco for a temporary reprieve. This legal conflict and the resulting sales impact follow a weeks-long disagreement between Anthropic and the Pentagon about the possible use of AI technologies for mass domestic surveillance and autonomous weapons. Anthropic asserts that AI is not yet capable of safely performing these tasks, while the Pentagon desires the authority to make that determination independently.

Legally, the supply-chain designation restricts a limited number of companies working with the Pentagon from integrating Anthropic into their systems. However, Defense Secretary Pete Hegseth has broadened the scope. He announced on X late last month that “effective immediately, no contractor, supplier, or partner that does business with the United States military may engage in any commercial activities with Anthropic.”

Rao mentioned that the Pentagon reinforced this message by contacting several startups regarding their use of Claude, information he obtained through conversations with a shared investor of Anthropic and the smaller companies. They “have become anxious and uncertain about their ability to utilize Claude,” Rao wrote.

The Pentagon did not provide comments on the lawsuits and did not promptly respond to requests for comments regarding Rao’s claims about the outreach.