Palantir Brings NVIDIA Nemotron Open Models to Air-Gapped US Government Environments

Palantir has unveiled a new intelligent engine that runs NVIDIA Nemotron open models inside air-gapped, secure infrastructure for U.S. government agencies. The integration gives agencies full control over data, model weights, and deployment — while enabling frontier-quality AI capabilities in the most sensitive environments.

Palantir has announced a new AI engine built on NVIDIA Nemotron open models, designed specifically to serve U.S. government agencies operating in secure, isolated environments. The system runs on NVIDIA accelerated computing inside air-gapped setups — completely cut off from unsecured networks — giving agencies frontier-level AI without sacrificing data sovereignty.

Why Open Models Matter for Government AI

Open source innovation has underpinned U.S. technology leadership for decades — from DARPA's ARPANET in 1969 to Linux in 1991, GitHub in 2008, and Docker in 2013. Today, open models extend that legacy into AI.

For government and enterprise users, open models offer three core advantages:

Trust through transparency — Independent researchers can audit for vulnerabilities, biases, and unintended behaviors
Customization and control — Agencies can fine-tune models for specific use cases and deploy them in regulated, sensitive environments
Lower costs — Roughly two-thirds of companies already use open models and cite cost efficiency as a key driver of adoption

How the Palantir–NVIDIA Integration Works

Palantir's Sovereign AI Operating System — built on AIP, Ontology, Foundry, and Apollo — handles data authorization, architecturally enforced isolation, and full auditability. NVIDIA Nemotron open models slot in as the customizable, continuously learning model layer.

Agencies and operators can:

Run customized Nemotron models on their own infrastructure
Train on their own data
Retain full ownership of the resulting models, including the weights

As models are used in production, agencies can improve them using new data and feedback — creating a data flywheel that optimizes performance while keeping everything under customer control.

Government as Enterprise

With roughly 3 million civilian employees, the U.S. government operates at the scale of one of the world's largest enterprises. Its operations span commerce, energy, healthcare, agriculture, education, and transportation.

Open models can be deployed in regulated environments where closed models might breach data security or other privacy laws.

AI can meaningfully streamline this complexity — from food safety inspections to interstate infrastructure management — mirroring the productivity gains already seen in the private sector.

Enterprise-Grade Support

For organizations that need additional infrastructure backing, NVIDIA AI Enterprise software can support full enterprise-grade deployments on top of this stack. The combined architecture of Nemotron open models and Palantir's sovereign platform is positioned to advance U.S. AI leadership across both government agencies and commercial businesses.