Palantir and NVIDIA Introduce Sovereign AI Operating System Reference Architecture

Palantir Technologies has announced a Sovereign AI Operating System Reference Architecture developed in collaboration with NVIDIA, according to a press release. The system provides a complete AI datacenter solution, covering hardware procurement, setup, and deployment of enterprise AI applications.

The Palantir AI OS Reference Architecture is based on NVIDIA’s Enterprise Reference Architectures and supports Palantir’s software suite, including AIP, Foundry, Apollo, Rubix, and AIP Hub. It runs on NVIDIA Blackwell Ultra systems equipped with eight Blackwell Ultra GPUs and NVIDIA Spectrum-X Ethernet networking for AI training and inference.

The architecture integrates Palantir’s hardened Kubernetes compute infrastructure with a unified management plane combining Rubix and Apollo. It also includes NVIDIA’s full-stack software acceleration, such as CUDA-X Libraries, AI Enterprise, and Nemotron open models.

Designed for organizations with data sovereignty and latency-sensitive requirements, the Sovereign AI OS enables control over data, models, and applications while supporting on-premise, edge, and sovereign cloud deployments.