Ahead of Nvidia's flagship GTC event next week, data mining giant Palantir Technologies revealed a partnership with the chip leader centered around sovereign AI.

Palantir said its AI OS Reference Architecture (AIOS-RA), based on Nvidia Enterprise Reference Architectures, is aimed at clients with data sovereignty requirements, existing GPU architecture, latency-sensitive workflows, and wide geographic distribution.

The platform runs on eight Nvidia Blackwell Ultra GPUs, alongside Nvidia Spectrum-X Ethernet networking for AI inference and training. The stack also boasts Nvidia AI Enterprise, Nvidia CUDA-X Libraries, Nvidia Nemotron open models, and Nvidia Magnum IO.

On the Palantir side, AIOS-RA touts a unified management planet comprising the Rubix zero-trust Kubernetes platform, the Apollo autonomous deployment and lifecycle management service, and Palantir's enterprise AI-centric AIP suite. The overall compute structure includes Foundry services for features like catalog and build.

For Nvidia, the Palantir union adds another sovereign jewel in its crown, coming on top with similar deals with Orange Business, India's NxtGen, and the British government. Palantir, meanwhile, co-signed a deal with Accenture to help U.K.-based infrastructure provider Sovereign AI deliver a sovereign foundation for commercial and government sectors across Europe.

In a recent interview, Accenture Head of Technology Research for Europe Surya Mukherjee defined AI sovereignty as comprising the entire tech stack.

"Where does the AI decide, what data does it use and, what does it throw up? So those are two layers," Mukherjee said. "Then there's security on a granular, fine level looking at the model itself: which country produced it, where was it produced?"

Mukherjee pointed to Stanford AI Index research highlighting 70% of the leading large language models (LLMs) were made in the U.S., with 25% in China.

"So that's 95% of the world's models not being produced here in Europe," Mukherjee said. "It's something that you need to think very deeply about as a country as well as a company."

On the AIOS-RA deal, Justin Boitano, VP for enterprise AI platforms at Nvidia commented: “AI is redefining the infrastructure stack – demanding, latency-sensitive and data-sovereign environments require a full-stack architecture – built from silicon to systems to software. By combining Palantir’s sovereign AI OS reference architecture with Nvidia AI infrastructure, industries and nations can turn data into intelligence with speed, efficiency, and trust.”

“From our first deployment with the United States government and in every deployment since, our software has had to meet the moment in the most complex and sensitive environments where customers must maintain control,” added Palantir Chief Architect Akshay Krishnaswamy. “Together with Nvidia – and building on many customers’ existing investments – we are proud to deliver a fully integrated AI operating system that is optimized for Nvidia accelerated compute infrastructure and enables customers to realize the promise of on-premise, edge, and sovereign cloud deployments.”