Palantir has launched a new "intelligent engine" that runs Nvidia's Nemotron open AI models inside the locked-down computing environments used by U.S. government agencies, the two companies announced.
The pitch is captured in Nvidia's own framing: "Open Models, Closed Environments." Nemotron is a family of open AI models, but federal agencies can't simply send sensitive data to a public chatbot. According to the Nvidia blog, Palantir's engine lets agencies deploy these open models in what the companies call secure, "sovereign" environments — meaning the data and the AI stay inside government-controlled systems rather than the open internet.
Nvidia frames the move as part of a long American tradition of open-source innovation, noting in its blog that open software has underpinned U.S. technology leadership since DARPA connected the first ARPANET nodes in 1969.
The news also lands as a business story. According to Seeking Alpha, Palantir is pushing to commercialize new software — including expanding its SurfOS product for private aviation — while the Nvidia tie-up lets it run AI models securely in sovereign U.S. settings. Business Wire and Yahoo Finance carried the launch announcement under the headline that Palantir is "deploying Nvidia Nemotron open models in sovereign environments."
Investors took notice: Blockonomi reported that Palantir (PLTR) stock surged on the strategic partnership for federal agencies.
Why it matters: as governments race to adopt AI, the hard part isn't the model — it's running powerful AI on classified or sensitive data without it leaking, and this deal is a bet that secure, on-premises deployment is how AI actually reaches Washington.