Anthropic faces new executive order threat as lawsuit over Pentagon ban unfolds

The Trump administration is drafting an executive order against Anthropic, the San Francisco AI startup behind Claude, even as a federal lawsuit over an existing government ban on the company plays out in court.

Sources familiar with the matter say White House officials are preparing the order. The move is notable because it comes while Anthropic’s legal challenge to earlier government restrictions is still pending. The company sued after it said it was shut out of competing for defense contracts.

When asked whether the administration planned further steps beyond the executive order, White House official did not say no. He told Axios any policy announcement will come directly from” Trump and the “discussion about potential executive orders is speculation.”

The dispute traces back to Defense Secretary Pete Hegseth, who last month branded Anthropic a “supply chain risk” after contract talks with the Pentagon broke down. On February 27, Hegseth wrote on X, “Effective immediately, no contractor, supplier, or partner that does business with the United States military may conduct any commercial activity with Anthropic.”

Anthropic says in court that the ban is retaliation for CEO Dario Amodei’s refusal to allow Claude to be used for mass surveillance or autonomous weapons.

Government pressuring Anthropic customers to switch, lawyer says

The harm, the company’s lawyers say, is already showing up. At a court status conference last Tuesday, attorney Michael Mongan said the blacklisting is doing “real and irreparable harm” to Anthropic every single day.

Customers are pulling back, he said, and the government has been reaching out to them directly to encourage them to switch to other AI companies. “We’ve had university systems and business-to-business companies that have switched to competing AI companies,” Mongan told the court. “Defendants have been affirmatively reaching out to our customers and pressuring them to stop working with Anthropic and switch to other AI companies.”

Court filings put numbers on the damage. Anthropic has generated over $5 billion in total commercial revenue to date and spent $10 billion on model training and computing. The company warned that at minimum, hundreds of millions of dollars in 2026 revenue is now at risk. In the worst case, it could be several billion.

Dozens of companies have called Anthropic asking what the ban means for them, in some cases asking whether they can walk away from their contracts. Amazon, Alphabet, and Microsoft have all said they will continue offering Claude to customers for now.

Why the rest of Big Tech is watching closely

The fight matters well beyond Anthropic. The entire AI supply chain, including some of the biggest names in tech, is tangled up in the fortunes of a handful of companies.

Microsoft said last month that 45 percent of its $625 billion backlog of demand is tied to OpenAI. OpenAI itself has a $300 billion agreement with Oracle. Those two deals alone represent about two-thirds of the roughly $800 billion that HSBC estimates OpenAI will pour into chips and data centers. The rest goes to companies like Nvidia, AMD, Amazon, and CoreWeave.

Anthropic’s spending plans run into the billions as well, and the Pentagon standoff now raises questions about whether those plans can move forward.

The timing is particularly awkward because Anthropic is reportedly planning to go public this year. A prolonged government fight is not the backdrop any company wants heading into an IPO. For all the concern, though, if things went badly wrong for Anthropic, someone would probably buy the company before it collapsed.

Anthropic reshuffles from inside

On Wednesday, Anthropic also said it is creating a new in-house research group called the Anthropic Institute, formed by merging three existing teams. The group will look at how advanced AI affects jobs and economies, whether it makes the world safer or more dangerous, how the values baked into AI systems might rub off on people, and whether humans can keep meaningful control as the technology grows more powerful.

The news came with a reshuffling at the top.  Jack Clark, one of the company’s co-founders, is stepping down as head of public policy after more than five years and will now lead the new institute under the title of head of public benefit. His replacement on the policy side is Sarah Heck, who was previously head of external affairs.

The policy team, which grew three times larger in 2025, will keep working on national security, AI infrastructure, energy, and what Anthropic describes as democratic leadership in AI.

The company also plans to open a Washington, D.C. office. Clark said the shift came down to the pace of change in AI. When he looked back at his work last year, he realized he had spent more time on legislation like California’s SB 53 than on the AI research questions he cared most about. He said he thinks powerful AI, the kind the industry often calls AGI, will be here by the end of 2025 or early 2027.

When asked if it made sense to build out long-term research while the company’s short-term income is under threat, Clark said he had no worries. “People tend to buy trust,” he said. “Long-term, Anthropic has always viewed its investment in safety, and studying and reporting on the safety of its systems, as being not a cost center but a profit center.”

If you want a calmer entry point into DeFi crypto without the usual hype, start with this free video.