Anthropic vs. White House puts $60 billion at risk
Add Axios as your preferred source to
see more of our stories on Google.
Photo illustration: Allie Carl/Axios. Photo: Kevin Dietsch/Getty Images & PhotoQuest via Getty Images.
Anthropic has raised more than $60 billion from over 200 "venture capital" investors, half of which came just last month.
Why it matters: That massive investment is now at risk due to a contract dispute with the Pentagon that's become as much about ego as AI.
Catch up quick: Anthropic and the Defense Department on Friday failed to reach agreement on a long-term deal for the military to continue licensing Anthropic's AI models, as it has done in both the Venezuela and Iran operations.
- Axios first reported on the licensing dispute several weeks ago, saying that it centered on disagreements over fully autonomous weapons and mass surveillance.
- Anthropic CEO Dario Amodei last Thursday reiterated these concerns in a blog post. For weapons, he argued that AI wasn't yet up to the task. For surveillance, he argued that the law hadn't yet caught up to the tech.
Driving the news: President Trump responded on Truth Social by directing every federal agency to cease using Anthropic technology, and Defense Secretary Pete Hegseth followed up by tweeting that he would define Anthropic as a supply chain risk to national security.
- Hegseth added: "Effective immediately, no contractor, supplier, or partner that does business with the United States military may conduct any commercial activity with Anthropic."
- Hours later, OpenAI signed the deal that Anthropic wouldn't.
Behind the scenes: Amodei's blog post is said to have infuriated Defense Dept officials, who believed he was trying to virtue signal to: (a) Anthropic employees upset about the Venezuela revelations, and (b) AI engineers at rival companies who might share similar concerns.
- The latter, they felt, could poison the well for future engagement with AI companies. Had Amodei just wanted to take his ball and go home, they felt, he should have done so more quietly.
The big picture: Designating a Silicon Valley company as a supply chain risk is unprecedented and a very blunt force tool, but arguably the best one DoD had to prevent other military contractors from deploying Claude in their applications.
- But what Hegseth tweeted went several steps beyond what that designation would require.
- For example, Nvidia does business with the U.S. military and also has commercial activities with Anthropic. Namely, it sells chips to Anthropic that are vital to the company's survival. Hegseth's language, taken at face value, would require that relationship be severed.
- DoD hasn't yet provided Anthropic with formal notice of the supply chain risk designation, which Anthropic has pledged to challenge in court, so the actual language could be much more limited than what Hegseth tweeted.
Zoom out: Venture capitalists historically avoided investing in tech startups that interacted deeply with government, outside of biotech, believing that the political and procurement variables were too volatile. In the past few years, however, many have thrown that caution to the wind.
- This situation may be chickens coming home to roost.
- Amodei stuck a finger in Trump's eye, and he responded with a series of uppercuts. It's not how U.S. presidents typically respond to pointed public criticism from a private company, but it's nonethess the current normal.
The bottom line: Anthropic has been at the heart of America's business narrative this year, introducing tools that could upend industries from cybersecurity to law and everything in between.
- But now it's under siege from its own government, and the fallout risks becoming an existential moment for all of American AI and its investors.
