Nvidia and Palantir Launch Sovereign AI Stack with Nvidia’s Supply Chain as First Deployment
Nvidia and Palantir have introduced a joint product for running AI over corporate supply chains without sharing underlying data, with Nvidia becoming the first customer to implement it on its own infrastructure. This partnership combines Palantir’s Foundry (AI Platform), Ontology, and open Nemotron models with Nvidia’s cuOpt optimization software, packaged as the Sovereign AI Operating System.
The system can be deployed in various settings: on-premises, in the cloud, or at a colocation facility, with Dell, Cisco, Rackspace, and Nebius serving as partners. This offering is particularly relevant for managing complex systems like Nvidia’s Vera Rubin racks, which consist of 1.3 million components, showcasing a genuine optimization challenge.
Jensen Huang, founder and CEO of NVIDIA, emphasized the significance: "Supply chains are the operating system of the physical economy, and AI factories are among the most intricate systems ever constructed."
The term "sovereign" in this context refers to data ownership and control remaining with the customer, a crucial distinction. However, it does not extend to controlling the underlying technology stack, including the ontology, models, optimization software, or the American companies providing these components.
This product launch resonates with recent European discussions regarding data sovereignty. Nvidia has secured agreements with Australian data center operators for local AI capacity, while also facing scrutiny from Germany, which ruled out Palantir for military support and is exploring alternative German suppliers. France’s intelligence service is also transitioning away from Palantir in favor of a domestic rival.
These developments suggest that the market for sovereign AI solutions is gaining traction, with companies recognizing the importance of data and technology control in a rapidly evolving digital landscape.