Palantir launched a new intelligent engine integrating NVIDIA's open-source Nemotron models into secure federal environments for U.S. government agencies, announced today. The deployment combines open models with government computing security protocols, allowing agencies to run state-of-the-art AI while maintaining strict data controls. According to NVIDIA's announcement, the collaboration showcases "the importance of open source innovation in American AI." Open models enable government agencies to inspect code, customize behavior, and run everything on-premise, addressing the challenge of operating AI when data cannot touch the public internet. Palantir's platform has security certifications across defense and intelligence agencies, channeling that access into AI infrastructure. NVIDIA's Nemotron family represents the company's push into enterprise-grade open models that can compete with proprietary alternatives. Nemotron's open architecture lets Palantir package models into air-gapped environments where agencies maintain complete control. The timing aligns with broader government AI initiatives as federal agencies face pressure to modernize AI capabilities. The announcement nods to DARPA's 1969 network connecting UCLA, Stanford, and UCSB computers, the precursor to the internet.