NEW YORK, March 10, 2026, 09:38 EDT
- Anthropic’s lawsuit filed March 9 threw Palantir’s Pentagon AI contracts back into question, following a Reuters story revealing that Palantir’s Maven system incorporates workflows built on Claude. 1
- The Pentagon has slapped a supply-chain risk label on Anthropic, effectively blocking contractors from tapping the company for military jobs. Anthropic argues the order is limited in scope and is fighting it in court. 2
- Palantir dipped roughly 0.4% early Tuesday. Last month, the company reported a 66% jump in U.S. government revenue for the fourth quarter. 3
Palantir Technologies (PLTR.O) is dealing with fresh questions around its role in a major Pentagon AI initiative, as Anthropic moved to challenge a U.S. blacklist that barred contractors from using its tech. Early Tuesday, Palantir stock slipped roughly 0.4%. 1
This is suddenly a big deal: Palantir’s Maven Smart Systems, which helps militaries sift through intelligence and targeting data, is running on prompts and workflows that use Anthropic’s Claude. Last week, Reuters put Palantir’s Maven-related contracts at over $1 billion, all riding on that platform. 4
Anthropic, in a pair of lawsuits filed Monday, argued the Pentagon’s “supply-chain risk” label is illegal, claiming the move penalizes it for blocking Claude’s use in autonomous weapons or mass surveillance targeting Americans. The designation prevents contractors from bringing Anthropic tech into Pentagon projects. CEO Dario Amodei last week described the restriction as “a narrow scope,” adding that Anthropic will help manage a transition for as long as it remains allowed. 1
Palantir now finds itself squeezed between ongoing legal battles and the tough facts of government procurement. Agencies like State, Treasury, and Health and Human Services have started stepping back from Anthropic tools. State swapped its internal chatbot for OpenAI, while HHS directed employees to stick with OpenAI’s ChatGPT or Google Gemini. 5
OpenAI recently reached an agreement to run its models on the Defense Department’s classified network, and later clarified it would restrict language to prohibit intentional domestic surveillance of U.S. persons. The Pentagon has inked deals with Anthropic, OpenAI, and Google, each potentially worth as much as $200 million—highlighting just how quickly Palantir’s defense-software turf is getting crowded, not clearing out. 5
Wedbush’s Dan Ives flagged that the Anthropic dispute has the potential to prompt certain customers to halt—“pencils down”—on Claude rollouts as the court process unfolds. For Palantir, it’s a signal that the risk extends beyond engineering alone; trust is at stake. Defense clients are weighing just how much third-party AI they’re willing to let into their critical mission tools. 1
There’s a lot riding on this. In its latest update, Palantir reported a 66% surge in U.S. government revenue for the fourth quarter, hitting $570 million. That pushed total sales to $1.41 billion. Even after all that, Tuesday’s close left Palantir trading near a $433 billion market cap. 3
The upcoming phase looks anything but simple. If the blacklist remains, Palantir will likely need to pull Claude from sections of Maven and rework those processes. Should a court scale back or delay the order, contractors could be in for another set of adjustments as the legal battle drags on. 4
Last week, Chief Executive Alex Karp didn’t hold back: Palantir’s boss warned that Silicon Valley outfits who “screw the military” could see “the nationalization of our technology.” With the Anthropic dispute still playing out, Palantir now finds itself caught in the larger tug-of-war—who draws the line on military AI, Washington or the model makers? 4
