(SeaPRwire) –   By: Lucas Caldwell

Everyone’s sharing clips of Alex Karp’s unhinged CNBC rant like he’s a rogue truth-teller calling out Silicon Valley AI grift. Don’t buy the act. Karp is one of the most well-connected defense tech operators in the industry, and this entire 20-minute tirade was a carefully staged sales pitch. Every offhand quip, every angry jab at OpenAI and Anthropic, was calibrated to hit exact fear buttons for US and Western government procurement teams.

Karp was originally booked on CNBC to announce Palantir’s new infrastructure partnership with Nvidia. The two companies will sell an all-in-one data analytics package, pairing Palantir’s operating systems with Nvidia’s AI hardware and pre-trained models. He spent most of the segment bashing commercial LLM providers for their standard pay-as-you-go token model. He claimed clients hand over cash and sensitive data without getting any control over model weights, the core variables that shape AI decision-making.

He went on to accuse OpenAI and Anthropic of selling their technology to US adversaries while refusing to work with the US military over self-imposed ethical guidelines. He rambled tangentially about wealth taxes, far left and far right fears of “warlocks roaming the street building technology,” and his personal relationship with Israel. When anchor Becky Quick noted he sounded angry, he claimed he was channeling the voice of American business as a “neurodivergent crazy person.” He then plugged Palantir’s software as a middle layer that lets clients retain full control of their data and model weights.

The US government is currently shopping for scalable, secure AI tools to integrate into defense and intelligence workflows. It has been hesitant to deploy commercial LLMs at scale because of documented data leakage risks and lack of control over model outputs for sensitive use cases. Karp is positioning Palantir as the only trusted alternative, leaning on the company’s 20-year history of working closely with the CIA and US defense agencies. He’s also directly undermining his biggest commercial competitors in the government AI space with one high-profile media hit.

But Karp’s pitch has a massive, unacknowledged flaw that limits his total addressable market drastically. His close public alignment with US military supremacy goals has already turned off major Western US allies. France and Germany banned their domestic intelligence agencies from renewing Palantir contracts years ago, and Spain forbids all state-backed companies from using its products entirely. Those countries don’t trust commercial LLM providers either, but they trust Palantir even less to not pass their sensitive national security data straight to US intelligence.

The US government will formalize a stake in Palantir within the next 18 months as part of its defense AI procurement strategy.

Author bio: Lucas Caldwell, a tech opinion leader with 2.8 million followers on X/Twitter covering enterprise AI and defense tech market shifts.