NVIDIA AI Enterprise: A Foundation for Secure and Compliant AI

NVIDIA AI Enterprise is a software suite of APIs, libraries, and runtimes that accelerates and simplifies the development and deployment of AI applications. It is engineered to provide a stable, secure, and performant foundation, allowing development and operations teams to focus on innovation and mission outcomes rather than low-level infrastructure management. The suite comprises two distinct but integrated layers:

• Infrastructure Software: This layer provides the essential, hardware-aware components to support AI workloads on accelerated systems. It includes NVIDIA drivers and Kubernetes operators for GPUs and networking. These components have dedicated support and maintenance lifecycles to ensure robust production stability at scale.

• Application Software: This layer includes NVIDIA NIM microservices for optimized model inference, frameworks like NVIDIA NeMo for building and customizing generative AI models, and essential libraries such as Dynamo-Triton Inference, TensorRT, PyTorch, and RAPIDS. Customers receive exclusive access to Production Branch software, where each release is maintained for nine months with monthly security updates and guaranteed API stability, providing a predictable and reliable development lifecycle.

To further enhance security across the platform, both layers include a comprehensive Software Bill of Materials (SBOM) and cryptographically signed container images to ensure integrity. Where vulnerabilities cannot be immediately remediated, NVIDIA provides detailed context and exploitability information in Vulnerability Exploitability eXchange (VEX) documents, enabling a more intelligent and risk-informed security posture.