Here's a tricky situation: the Pentagon wants to stop using a certain AI company's technology, but one of its key contractors is still running that tech. That's the spot Palantir Technologies Inc. (PLTR) finds itself in, according to CEO Alex Karp. In a recent interview, Karp said Palantir's defense platforms are still humming along with models from AI startup Anthropic, even after the Defense Department labeled Anthropic a "supply-chain risk" and started moving to blacklist it.

At the same time, a famous investor who's been betting against Palantir isn't letting up. Michael Burry is out there warning of more pain for the stock, which has already taken a tumble. So you've got a government contract headache on one side and a vocal short-seller on the other. Not the most relaxing week for Palantir shareholders.

The AI That's Still in the Room

Karp made his comments at a company event, clarifying the status of Anthropic's Claude model within Palantir's systems. "The Department of War is planning to phase out Anthropic; currently, it's not phased out," he told CNBC. He added that Palantir expects to plug in other large language models alongside Claude over time. So for now, the tech is still live, but the writing is on the wall.

This all stems from a White House directive. President Donald Trump has ordered federal agencies to "immediately cease" using Anthropic's technology. However, the order gives the Pentagon about six months to unwind existing deployments. That's not a simple flick of a switch, especially when the AI is woven into operational systems.

Unwinding a Classified AI Partnership

Claude holds a notable distinction: it was the first frontier AI model deployed on U.S. classified networks, thanks to partnerships with Palantir and other contractors. It's reportedly even in use supporting U.S. military operations in Iran right now. So the phase-out clock is ticking on technology that's actively being used in sensitive contexts.

Other agencies, like State, Treasury, and Health and Human Services, have already started shifting from Claude to rivals like OpenAI. But for the Pentagon and its contractors, it's more complicated. Palantir's Maven Smart Systems framework, used for intelligence and targeting, has prompts and workflows that were originally built on top of Claude. Ripping that out isn't just a software update; it's potentially re-engineering parts of a critical defense system.

Michael Burry's Relentless Short Campaign

While Palantir navigates this government contract shift, it's also facing sustained criticism from a high-profile investor. Michael Burry has been using his Substack and X account to renew his bearish call on Palantir. He's pointing to technical chart patterns, like a head-and-shoulders formation, and predicting more than 40% downside for the shares.

Burry isn't just talking charts. He's also resurfacing past reports and controversial comments from CEO Alex Karp—including a viral quip about "fentanyl-laced urine drones" aimed at hostile analysts—to question the company's tactics and the durability of its AI-driven growth narrative. It's a full-court press of public skepticism.

Where This Leaves Palantir Stock

Burry first disclosed a sizable short position in Palantir back in November. Since then, the stock has dropped roughly 30%, caught up in the broader volatility hitting software and high-multiple AI stocks. Bulls haven't given up, though. They argue that recent approvals from the Defense Information Systems Agency and deeper integration with core Pentagon systems actually strengthen Palantir's competitive moat.

But between the looming Anthropic phase-out and Burry's very public campaign, there's no shortage of headline risk hanging over the company. It's a classic clash of narratives: the long-term government contractor versus the skeptical investor seeing a bubble.

According to market data, Palantir shares were up 1.05% at $153.19 at the time of the original report.