Palantir and Nvidia Build Sovereign AI Platform for U.S. Government

Joint platform targets classified government workloads

Palantir Technologies and Nvidia have unveiled a collaborative platform designed to bring large language model capabilities to U.S. government agencies and critical infrastructure operators while maintaining strict data sovereignty and security controls.

The initiative integrates Nvidia's AI compute infrastructure, software ecosystem, and Nemotron open models with Palantir's suite of products—AIP, Ontology, Foundry, and Apollo. The combined offering allows government entities to train AI models on their own sensitive data, retain complete ownership of those models, and operate them in classified and air-gapped environments that are physically isolated from external networks.

Security architecture for sensitive operations

The platform incorporates several security features tailored to government requirements. These include explicit data authorization controls, customer-specific isolation enforced at the architectural level, secure perimeter enforcement, data portability, erasure rights, and comprehensive auditability.

Agencies can establish continuous improvement cycles by feeding new operational data and user feedback back into their custom models, refining performance over time without exposing proprietary information to external systems.

"Combining Palantir infrastructure with Nvidia's AI and Nemotron models will allow the U.S. government to unleash the full power of LLMs while removing the underlying security risks and rational concerns around proprietary insights migrating into the weights of closed models," Palantir co-founder and CEO Alex Karp stated.

Open source foundation for national security

Nvidia founder and CEO Jensen Huang emphasized that open source AI serves as a cornerstone for national security, public safety, and U.S. technology leadership. The partnership provides government agencies with what he described as "a secure, customizable and fully controlled foundation to build mission-critical AI systems."

The offering extends a previously announced Sovereign AI Operating System Reference Architecture between the two companies. Enterprise deployments can leverage Nvidia's AI Enterprise software suite, including Nvidia NIM microservices for streamlined model deployment.

Why it matters

Government adoption of AI has been constrained by legitimate concerns about data leakage, model ownership, and the security implications of using commercial cloud-based AI services. This partnership addresses a critical gap: enabling federal agencies to harness advanced language models without sending sensitive data to third-party providers or risking that classified information becomes embedded in shared model weights. As AI becomes central to defense, intelligence, and infrastructure operations, sovereign platforms that keep data and models under direct government control will likely become the standard architecture for public sector AI deployment.

Market response

Palantir shares rose 3.3% to $116.69 following the announcement, building on a 5.3% gain from the previous Friday that ended a seven-day decline. Despite the recent recovery, the stock remains down 34% year-to-date, according to details first reported by Barron's and Quartz.