Blog

Oracle Expands AI Infrastructure Options for U.S. Government Customers

By Peter Guerra, GVP, Data & AI, Government, Defense and Intelligence (GDI)—Mar 31, 2026

We are pleased to announce our plans to strengthen AI infrastructure and increase model deployment options for our U.S. government customers.

To help U.S. government customers scale demanding AI inference and accelerated computing workloads while maintaining mission-critical security and compliance, NVIDIA B300 GPUs will be available in Oracle Cloud Infrastructure (OCI) government regions. Powered by NVIDIA Blackwell Ultra GPUs, DGX B300 helps accelerate large language model (LLM) inference and training while delivering energy efficiency and performance per watt.

Additionally, OCI plans to offer xAI Grok and NVIDIA Nemotron models to give agencies greater choice and flexibility for mission-focused AI deployments, such as situational awareness and cybersecurity. xAI’s recent Grok 4 and Grok 4 Fast models excel at data extraction, code generation, and text summarization. NVIDIA Nemotron models’ high throughput enables agents to think faster and generate higher-accuracy responses while lowering inference cost.

Together, these planned expansions is designed to bring the same pace of AI innovation available in commercial cloud to the agencies and regulated, mission-focused environments that need them most.

 

Addressing Unique Government AI Requirements

Federal agencies need secure and flexible ways to adopt AI in environments with inherently complex mission, regulatory, and operational requirements. Such requirements include navigating stringent data privacy laws, integrating with legacy systems, and facilitating resilience against cyber threats.

By expanding AI infrastructure and model choice, Oracle is empowering government customers to deploy AI with greater flexibility, speed, and control. In effect, this allows users to  implement AI in ways that precisely align with their unique security protocols, compliance mandates, and critical performance needs across various agencies.

 

Building on OCI’s Existing AI Capabilities

Today’s updates build on existing AI infrastructure, model deployment options, and accelerated computing capabilities available in OCI today. Government cloud customers can currently build and deploy production-grade AI applications using NVIDIA AI Enterprise on OCI, including NVIDIA NIM microservices for generative AI model deployment and NVIDIA NeMo tools for training and inference workflows.

With NVIDIA AI Enterprise software on OCI, developers can use a full-stack accelerated computing platform to develop proprietary, secure AI applications that help improve agency operations and the lives of citizens they serve.

The planned AI infrastructure and model options are designed to support key government AI workloads, including agentic AI, retrieval-augmented generation (RAG), secure inference, and mission-focused analytics. These capabilities are subject to service readiness and regional infrastructure availability.

 

Additional Resources