The senior official said: "It will be an enormous pain in the ass to disentangle, and we are going to make sure they pay a price for forcing our hand like this."
Why it matters: That kind of penalty is usually reserved for foreign adversaries.
Chief Pentagon spokesman Sean Parnell told Axios: "The Department of War's relationship with Anthropic is being reviewed. Our nation requires that our partners be willing to help our warfighters win in any fight. Ultimately, this is about our troops and the safety of the American people."
The big picture: Anthropic's Claude is the only AI model currently available in the military's classified systems, and is the world leader for many business applications. Pentagon officials heartily praise Claude's capabilities.
As a sign of how embedded the software already is within the military, Claude was used during the Maduro raid in January, as Axios reported on Friday.
Breaking it down: Anthropic and the Pentagon have held months of contentious negotiations over the terms under which the military can use Claude.
Anthropic CEO Dario Amodei takes these issues very seriously, but is a pragmatist.
Anthropic is prepared to loosen its current terms of use, but wants to ensure its tools aren't used to spy on Americans en masse, or to develop weapons that fire with no human involvement.
The Pentagon claims that's unduly restrictive, and that there are all sorts of gray areas that would make it unworkable to operate on such terms. Pentagon officials are insisting in negotiations with Anthropic and three other big AI labs — OpenAI, Google and xAI — that the military be able to use their tools for "all lawful purposes."
A source familiar with the dynamics said senior defense officials have been frustrated with Anthropic for some time, and embraced the opportunity to pick a public fight.
The other side: Existing mass surveillance law doesn't contemplate AI. The Pentagon can already collect troves of people's information, from social media posts to concealed carry permits, and there are privacy concerns AI can supercharge that authority to target civilians.
An Anthropic spokesperson said: "We are having productive conversations, in good faith, with DoW on how to continue that work and get these new and complex issues right."
The spokesperson reiterated the company's commitment to using frontier AI for national security, noting Claude was the first to be used on classified networks.
The stakes: Designating Anthropic a supply chain risk would require the plethora of companies that do business with the Pentagon to certify that they don't use Claude in their own workflows.
Some of them almost certainly do, given the wide reach of Anthropic, which recently said eight of the 10 biggest U.S. companies use Claude.
The contract the Pentagon is threatening to cancel is valued at up to $200 million, a small fraction of Anthropic's $14 billion in annual revenue.
Friction point: A senior administration official said that competing models "are just behind" when it comes to specialized government applications, complicating an abrupt switch.
The intrigue: The Pentagon's hardball with Anthropic sets the tone for its negotiations with OpenAI, Google and xAI, all of which have agreed to remove their safeguards for use in the military's unclassified systems, but are not yet used for more sensitive classified work.
A senior administration official said the Pentagon is confident the other three will agree to the "all lawful use" standard. But a source familiar with those discussions said much is still undecided.
