After offending the Ministry of Defense, the collaboration between Anthropic and Palantir is also about to collapse

A ban issued by the Pentagon not only caused AI unicorn Anthropic to lose a major military contract, but also triggered a domino effect that is forcing government-backed data giant Palantir to consider a complete break with Anthropic.

According to the latest report from The Information, for more than a year, Anthropic has been providing services to the U.S. government through Palantir. Palantir’s clients use Anthropic’s models to identify patterns in massive amounts of classified data to aid decision-making. However, as the Department of Defense lists Anthropic as a “supply chain risk,” this partnership is now in jeopardy.

As a favored government contractor, about 42% of Palantir’s nearly $4.5 billion revenue last year came from U.S. government contracts. If Palantir stops using Anthropic for federal work, it will cut off a source of revenue for the AI startup. Although this portion of sales is not expected to contribute much to Anthropic’s forecasted revenue of up to $18 billion this year, losing Palantir as an important partner is undoubtedly a blow to Anthropic’s expansion in the major client market.

Palantir May Need to Modify Software, CEO Criticizes Anthropic

Citing people familiar with the matter, some of Palantir’s custom software and AI applications are specifically designed to work with Anthropic’s Claude model. If Palantir abandons Claude, it will have to modify its services, which may take several weeks. However, Palantir can pivot to other AI model providers and derive roughly the same income from those contracts.

At a defense technology summit in Washington on Tuesday, Palantir CEO Alex Karp appeared to criticize Anthropic without naming it. He rebuked Silicon Valley for opposing the U.S. military and warned that AI companies risk angering both liberals and conservatives.

“If Silicon Valley thinks we’re going to take away everyone’s white-collar jobs... and you’re also going to screw over the military, if you don’t think that leads to our technology being nationalized, then you’re an idiot,” Karp said. “This is the direction things are headed.”

Anthropic Plans Legal Action, Attempts to Reassure Clients

Last Friday, after a dispute with Anthropic over military restrictions on its AI, Secretary of Defense Pete Hegseth stated that he ordered the government to designate Anthropic as a “supply chain risk.”

The Pentagon required Anthropic to license Claude for “all lawful uses.” Anthropic refused to compromise, citing concerns over “mass surveillance” and “fully autonomous weapons.”

Last Friday, U.S. President Trump posted on Truth Social that for agencies such as the Department of Defense using Anthropic’s products, there would be a “six-month phase-out period.”

Anthropic CEO Dario Amodei argued that the designation “can only extend to the use of Claude as part of a War Department (Trump administration’s preferred term for the Department of Defense) contract.” Legal experts state Anthropic could fully contest the designation, partly because the law generally applies to non-U.S. companies.

According to sources cited by the media, Anthropic also stated last Friday that it would take legal action in court against the supply chain risk designation. As the designation has not yet taken effect, Anthropic has not filed a lawsuit.

Anthropic is attempting to reassure clients and business partners, saying that the Defense Department’s decision (if implemented) would only restrict government contractors from using Anthropic’s technology in U.S. military work, not any company cooperating with the federal government.

Federal Contractors Await Government Guidance

According to sources familiar with Palantir’s products, Palantir currently allows clients to choose which AI model to use from providers including Anthropic, OpenAI, and Google when analyzing data collected by its platform. The Pentagon has been using Anthropic models hosted by Amazon Web Services (AWS) in conjunction with Palantir’s software.

Sources familiar with how the Pentagon uses these products revealed that in addition to Anthropic’s models, the Pentagon has also used models from OpenAI, Google, and other providers through its Palantir software.

According to tech company employees engaged in government business, as of this week, federal contractors are waiting for government guidance on how and whether they need to cut ties with Anthropic to retain their federal contracts.