NVIDIA Enterprise Reference Architectures

Type: Reference Architecture Program Tags: NVIDIA, Enterprise Reference Architecture, AI factory, certified systems, Spectrum-X, AI Enterprise, Kubernetes Related: NVIDIA-Enterprise-AI-Factory, NVIDIA-RTX-PRO-AI-Factory, NVIDIA-HGX-AI-Factory, NVIDIA-NVL72-AI-Factory, NVIDIA-DGX-BasePOD-B200-H200-H100-RA, NVIDIA-DGX-SuperPOD-B200-RA, NVIDIA-DGX-SuperPOD-GB200-RA, NVIDIA-DGX-SuperPOD-B300-Spectrum-4-Ethernet-RA, NVIDIA-DGX-SuperPOD-B300-Quantum-X800-InfiniBand-RA, NVIDIA-AI-Enterprise-Software-Reference-Architecture, NVIDIA-Enterprise-RA-Observability-Guide, NVIDIA-AI-Factory-for-Government, Red-Hat-AI-Factory-with-NVIDIA, NVIDIA-AI-Enterprise, NVIDIA-Certified-Systems, NVIDIA-Certified-Storage, NVIDIA-Spectrum-X, NVIDIA-Spectrum-X-Validated-Solution-Stack, NVIDIA-Quantum-X800-InfiniBand, NVIDIA-BlueField-DPU, NVIDIA-Base-Command-Manager, NVIDIA-Run-ai, NVIDIA-NetQ Sources: https://docs.nvidia.com/enterprise-reference-architectures/index.html, https://docs.nvidia.com/enterprise-reference-architectures/white-paper/latest/index.html, https://docs.nvidia.com/enterprise-reference-architectures/white-paper/latest/introduction.html, https://docs.nvidia.com/ai-enterprise/deployment/red-hat-ai-factory/latest/index.html, https://docs.nvidia.com/dgx-basepod/reference-architecture-infrastructure-foundation-enterprise-ai/latest/index.html, https://docs.nvidia.com/dgx-superpod/reference-architecture-scalable-infrastructure-b200/latest/index.html, https://docs.nvidia.com/dgx-superpod/reference-architecture-scalable-infrastructure-gb200/latest/index.html, https://docs.nvidia.com/dgx-superpod/reference-architecture/scalable-infrastructure-b300/latest/index.html, https://docs.nvidia.com/dgx-superpod/reference-architecture/scalable-infrastructure-b300-xdr/latest/index.html Last Updated: 2026-05-09

Summary

NVIDIA Enterprise Reference Architectures are NVIDIA-authored design patterns for building enterprise AI factories from validated compute, networking, storage, software, and operations components. The current docs hub groups the program into hardware reference architectures, software reference architecture guidance, observability, and deployment guides.

Detail

Purpose

Enterprise RAs are meant to reduce the risk of building AI infrastructure from scratch. They give partners and enterprise customers prescriptive patterns for GPU node configurations, scalable units, network fabrics, storage expectations, Kubernetes-oriented software, and operational tooling.

Current RA family

NVIDIA context

This page is the canonical program-level page. It should not absorb every deployment recipe, partner-endorsed design, PDF appendix, or build.nvidia example. Use the specific RA pages for durable NVIDIA-authored architecture documents, and use NVIDIA-Enterprise-AI-Factory for the broader strategy and planning concept.

Connections

Source Excerpts

  • NVIDIA’s docs hub says Enterprise RAs are for building AI factories that scale and groups the content into overview, hardware, software, observability, and deployment areas.
  • The Enterprise RA overview positions the program as a way to simplify deployment, reduce complexity, and accelerate time to value for enterprise-class AI factory deployments.

Resources