en.Wedoany.com Reported - Palantir Technologies has partnered with U.S. government agencies to launch an AI intelligence engine capable of operating in air-gapped environments, built on NVIDIA's Nemotron open model.
This deployment brings NVIDIA's open models to computing devices physically isolated from unsecured networks. Agencies can run customized models on their own infrastructure, train models using their own data, and retain ownership of the resulting models and their weights.
The product is centered on Palantir's Sovereign AI operating system, which integrates the company's AIP, Ontology, Foundry, and Apollo offerings. Palantir plans to leverage NVIDIA's Nemotron open models to build customized models for U.S. government work, addressing agency needs in areas such as commerce, energy, healthcare, agriculture, education, and transportation. The software stack manages deployment, data authorization, isolation, and auditability in sensitive government environments.
NVIDIA views this collaboration as part of its efforts to promote open-source and open-model software in the United States. The company believes that open models enable government agencies and enterprises to more thoroughly inspect, adapt, and deploy AI in sensitive environments, helping users maintain control over proprietary data, deployment environments, and model development. In Palantir's approach, agencies can continuously refine models within their own environments using new data and real-world operational feedback, keeping the model improvement process within the customer's system.
The emphasis on air-gapped deployment is intended to serve national security and other sensitive public sector uses. This means models can run on infrastructure with no connection to external unsecured networks. NVIDIA stated that deployment can be supported through its AI Enterprise software suite, providing customers with a way to manage and run software around open models in operational environments. NVIDIA noted that independent review helps researchers identify vulnerabilities, biases, and unexpected behaviors that a single organization might overlook. Open models can also be modified and fine-tuned for specific uses, particularly in regulated industries where closed models may raise data security or privacy concerns. NVIDIA cited industry research showing that approximately two-thirds of enterprises have adopted open models and achieved cost benefits from them.
