President orders every federal agency to immediately stop using Anthropic AI technology.

Hon. Friday afternoon. The President of the United States, on his social media platform, in all caps, ordered every federal agency in the United States government to immediately cease all use of Anthropic’s technology. He added, quote, we don’t need it, we don’t want it, and will not do business with them again.

The reason, hon, when you walk it back, is almost a punchline.

Anthropic is the company that makes a language model called Claude. The Pentagon was using Claude, on a 200 million dollar contract, and it wanted to keep using it. Anthropic, in negotiating the contract, asked the Pentagon to agree to not use Claude for mass surveillance of Americans. It also asked the Pentagon to not use Claude for final targeting decisions in military operations without human involvement.

That is the whole dispute, hon. The whole thing. The company said please don’t use our software for mass surveillance of Americans or to autonomously kill people. The Pentagon said no, we want it for everything. The negotiations broke down. The Defense Secretary gave the company a deadline of 5:01 p.m. Friday to back down on the guardrails.

The company did not back down.

The President, at some point that afternoon, issued the order that every federal agency must immediately stop using the technology. The Defense Secretary, the same hour, designated Anthropic a supply chain risk under the Defense Production Act, a law from the 1950s used for national emergencies.

The supply-chain-risk designation, hon, is the strongest tool the Pentagon has short of seizing the company. The administration used the strongest tool against the company because the company asked the administration to please not use the AI for mass surveillance. I am not paraphrasing. That is the dispute.

I want you to picture the meeting, on Thursday, in which the Department of Defense sat down with Anthropic’s lawyers and the lawyers said, we are not comfortable having our software used for autonomous weapons. And the Pentagon side said, we insist. And the lawyers said, we’d rather lose the contract. And the Pentagon said, fine, we will then declare you a supply chain risk and have you removed from the federal procurement system. That meeting happened. In a building in Arlington. On a Thursday in 2026.

Buddy. Anthropic is not a foreign adversary. Anthropic is an American company, in San Francisco, with American employees, paying American taxes, whose requested guardrails on the use of its software would, in any other administration, be considered the kind of thing the Department of Defense agrees to in the contract without an issue.

You ever notice how the administration that says it is for American business will blacklist an American company for not letting the federal government do mass surveillance with the company’s product?

Funny how that works.