NVIDIA has unveiled the NVIDIA AI Factory for Government reference design at GTC Washington, D.C., in collaboration with leading technology companies to help federal agencies and regulated industries build secure, scalable AI platforms tailored to mission-critical needs. The initiative addresses the growing gap between the demands of modern AI and outdated government infrastructure, offering a comprehensive blueprint for deploying AI systems that meet stringent security, performance, and reliability standards. Built on the NVIDIA Blackwell architecture, the reference design leverages cutting-edge hardware including RTX PRO Servers, HGX B200 systems, Spectrum-X Ethernet, and the BlueField platform. It integrates NVIDIA AI Enterprise software, now optimized for FedRAMP-authorized clouds and high-assurance environments, with enhanced security features such as advanced code scanning, vulnerability management, and continuous monitoring. The design also incorporates NVIDIA Nemotron open models and the NeMo Agent Toolkit to support the development of intelligent, autonomous AI agents. Key partners are aligning their technologies with the AI Factory for Government. Palantir is integrating its Ontology framework with NVIDIA’s data processing and optimization tools to accelerate AI deployment across government and enterprise environments. ServiceNow is embedding the latest NVIDIA AI Enterprise software into its AI Platform for U.S. federal customers, including the new Apriel 2.0 model, which delivers advanced reasoning and multimodal capabilities in a lightweight, cost-efficient package. Cybersecurity is a central focus. CrowdStrike is expanding its Agentic Security Platform to support the AI Factory, enabling real-time threat detection and response using NVIDIA Nemotron models and the NeMo Agent Toolkit through Charlotte AI AgentWorks. Astris AI, a Lockheed Martin company, is integrating NVIDIA AI Enterprise software into its commercially available Astris AI Factory to deliver secure, high-performance AI for classified operations, building on over five years of internal use. Northrop Grumman is deploying an AI factory powered by NVIDIA RTX PRO Servers, Spectrum-X networking, and government-ready AI software to enhance workforce productivity and advance AI capabilities across its 100,000 employees. The ecosystem of partners is expanding the reach and functionality of the AI Factory. Developers can use platforms from Dataiku, DataRobot, Domino Data Lab, and H20.ai to build and scale AI workflows. Vector databases from Elastic and EnterpriseDB support efficient data retrieval. Security and observability tools from Dynatrace, Fiddler, JFrog, Trend Micro, and Weights & Biases help monitor and secure AI operations. NVIDIA Confidential Computing, supported by partners like Fortanix, enables data protection in hybrid and on-premises environments. Cloud providers including CoreWeave and Oracle Cloud Infrastructure are offering government-ready AI factory solutions in secure cloud environments. Server manufacturers such as Cisco, Dell Technologies, HPE, Lenovo, and Supermicro are delivering full-stack AI factory systems based on the reference design, accelerating deployment across the public sector and regulated industries. The NVIDIA AI Factory for Government marks a major step toward modernizing government AI infrastructure, enabling faster, more secure, and more reliable deployment of intelligent systems critical to national security and public service.