Palantir has introduced a new intelligent engine powered by NVIDIA Nemotron open models, enabling US government agencies and critical infrastructure operators to deploy customised AI in secure, air-gapped environments while retaining full ownership of their data and models.
Palantir has introduced a new intelligent engine built on NVIDIA Nemotron open models to support the needs of US government agencies – reinforcing the role of open source innovation in advancing American AI leadership.
The announcement highlights the continued importance of open technologies in the evolution of computing. It traces the origins of the modern Internet to 1969, when DARPA connected four university computers at UCLA, Stanford, UCSB and the University of Utah, creating the infrastructure that ultimately became the Internet.
Early US leadership in open source software also included UNIX in 1969 and the C programming language at Bell Labs in 1972. These technologies laid the foundations for later innovations including the Linux Kernel in 1991, GitHub in 2008 and Docker in 2013.
Today, open AI models are making frontier-level artificial intelligence more widely accessible by giving organisations greater control over customisation, transparency and trust. Enterprises and government agencies can inspect, adapt and deploy AI in sensitive environments, making open models increasingly important for national security, industrial innovation and corporate sustainability.
Palantir will use NVIDIA Nemotron open models to build customised frontier-quality AI models for the US government. Many government operations mirror those of private-sector organisations, including commerce, energy, healthcare, agriculture, education and transportation. With approximately three million civilian employees, the US government represents one of the world’s largest enterprises.
The new intelligent engine enables agencies and critical infrastructure operators to run customised Nemotron models on their own infrastructure, train them using their own data and retain complete ownership of the resulting models, including the weights that encode operational knowledge.
Palantir’s Sovereign AI Operating System, built on AIP, Ontology, Foundry and Apollo, provides the operational and data authorisation layer required for deployment in highly sensitive environments. The platform incorporates explicit data authorisation, architecturally enforced isolation and full auditability as core capabilities.
As agencies continue using the customised models in production, they can refine them using new operational data and user feedback. This creates a continuous data flywheel that improves model performance while ensuring that data, models and auditability remain under customer control.
While NVIDIA Nemotron open models provide the AI model layer within Palantir’s Sovereign AI Operating System, enterprise deployments can also be supported through the NVIDIA AI Enterprise software suite.
Open models support trust, access, control and lower costs
Together, NVIDIA Nemotron open models and Palantir’s critical infrastructure products are designed to deliver greater trust, accessibility, control and lower costs.
Trust through transparency
Independent review of open models enables researchers to identify vulnerabilities, biases and unintended behaviours that might otherwise go unnoticed. This transparency allows models to be refined and improved, strengthening safety and trust.
Customization and control
Companies, governments and developers can modify and fine-tune open models to suit specific operational requirements. Open models can also be deployed in regulated environments, including financial services, where closed models may present data security or privacy challenges.
Lower costs fuel economic development
Open models are already widely adopted, with around two-thirds of organisations reporting cost efficiencies as they scale AI deployments. These savings are viewed as an important factor in enabling AI adoption at enterprise scale.
The combination of NVIDIA Nemotron open models with Palantir’s critical infrastructure platforms – AIP, Foundry, Ontology and Apollo – strengthens US technology leadership for government agencies and commercial organisations alike. Running the models on air-gapped, NVIDIA-powered infrastructure helps keep data and AI models secure while supporting mission-critical operations.
