Nvidia, Palantir team up to build ‘Sovereign AI’ for supply chains

Nvidia and Palantir Technologies announced a new AI collaboration on Wednesday, which they call “sovereign AI” for critical supply chains.

The system will first run inside Nvidia’s own manufacturing and logistics network, which other industries will follow later.

The system combines Nvidia’s Nemotron open models with Palantir’s Foundry and AIP platforms, and it is grounded in Palantir’s Ontology system.

The goal is real-time visibility into supply chain constraints, which will also help codify operational knowledge and speed up decisions.

Companies keep control of their own data, and they don’t need to route it through a shared commercial model.

Built first for Nvidia’s own chip supply chain

Nvidia runs one of the most complex supply chains in the industry, spanning millions of parts, thousands of suppliers, and a global network of manufacturing partners.

Each Nvidia Vera Rubin server rack alone needs 1.3 million parts.

The new AI system should help Nvidia’s teams spot material shortages earlier.

It should also help them weigh alternatives faster, and it should guide allocation decisions based on their full impact on production.

“Nvidia has arguably the most valuable, intricate, and complex supply chain in the world,” said Alex Karp, Palantir’s co-founder and CEO, in the companies’ joint statement.

Nvidia founder and CEO Jensen Huang called supply chains “the operating system of the physical economy.”

He said the companies are turning that network into what he called sovereign intelligence.

Customers can build their own version

Other companies can also use the system, including firms in agriculture, manufacturing, pharmaceuticals, retail, technology and government.

They can post-train Nemotron models with their own data inside Palantir’s Foundry and AIP, and Nvidia’s cuOpt software adds scenario planning on top.

It lets teams model supply constraints and weigh tradeoffs, and the AI recommends actions and flags risks, but final decisions stay with human experts.

The deployment runs on a jointly built reference architecture called the Palantir Sovereign AI Operating System.

Dell Technologies and Cisco support on-premises deployments, and Rackspace and Nebius support cloud and co-location options.

Part of a deepening partnership

This is the latest step in a relationship the two companies have been building for close to a year.

Nvidia and Palantir first showed a joint AI framework at Nvidia’s GTC conference in October 2025, followed by the “Chain Reaction” domestic AI infrastructure initiative in December, and a sovereign AI partnership for US government and critical infrastructure customers announced in June.

The tie-up also gives both companies a high-profile reference case: Nvidia gets to show its own supply chain running on the system before pitching it to outside customers, while Palantir gains another marquee use case tied to the world’s most closely watched AI hardware maker.

Details of the collaboration will be presented in full at Palantir’s AIPCon 11 conference.