NVIDIA AI Enterprise Infrastructure Support Matrix
Type: Support Matrix Tags: NVIDIA, AI Enterprise, support matrix, compatibility, infrastructure, GPUs, networking, Kubernetes, cloud, virtualization Related: NVIDIA-AI-Enterprise, NVIDIA-AI-Enterprise-Software, NVIDIA-AI-Enterprise-Lifecycle-Policy, NVIDIA-AI-Enterprise-Quick-Start-Guide, NVIDIA-Certified-Systems, NVIDIA-DGX, NVIDIA-HGX, NVIDIA-GB300-NVL72, NVIDIA-Data-Center-GPU-Drivers, NVIDIA-DOCA, NVIDIA-BlueField-DPU, NVIDIA-vGPU, NVIDIA-vGPU-for-Compute, NVIDIA-MIG, NVIDIA-Container-Toolkit, NVIDIA-Run-ai, NVIDIA-GPU-Operator, NVIDIA-Network-Operator, NVIDIA-NIM-Operator, NVIDIA-Base-Command-Manager, NVIDIA-AI-Enterprise-Bare-Metal-Deployment, NVIDIA-AI-Enterprise-VMware-Deployment, NVIDIA-AI-Enterprise-Cloud-Deployment Sources: https://docs.nvidia.com/ai-enterprise/software/latest/infrastructure-software.html, https://docs.nvidia.com/ai-enterprise/release-8/latest/support-matrix/, https://docs.nvidia.com/ai-enterprise/release-8/latest/support/support-matrix.html, https://docs.nvidia.com/ai-enterprise/release-8/latest/support/support-matrix-8/8.0.html, https://docs.nvidia.com/ai-enterprise/release-8/latest/overview/release-notes-8/8.0.html Last Updated: 2026-04-29
Summary
NVIDIA AI Enterprise Infrastructure Support Matrix is the current compatibility reference for AI Enterprise infrastructure software, NVIDIA GPUs and networking, platforms, operating systems, hypervisors, orchestration options, and public-cloud deployment configurations. The interactive matrix compares releases 7.0 through 8.0, while the static 8.0 support matrix gives stable citations for the latest infrastructure release.
Detail
Purpose
Use this page when a question asks whether an AI Enterprise infrastructure stack is supported, which GPU/OS/hypervisor/orchestration combinations are valid, or which infrastructure software versions belong to the current release. Use NVIDIA-AI-Enterprise-Lifecycle-Policy for branch support timelines and NVIDIA-AI-Enterprise-Software for the software catalog.
Matrix scope
- Deployment configurations: bare metal, virtualized on-premises, and public cloud.
- Infrastructure software: GPU drivers, DOCA drivers and microservices, Fabric Manager, vGPU for Compute, Container Toolkit, Run:ai, DPU/GPU/Network/NIM Operators, and Base Command Manager.
- Hardware coverage: NVIDIA GPUs, accelerated platforms, ConnectX NICs, BlueField SuperNICs, and BlueField DPUs.
- Platform coverage: DGX, HGX, IGX, GB200/GB300 NVL, and related NVIDIA accelerated systems.
- Software environment coverage: operating systems, Kubernetes distributions, OpenShift, managed Kubernetes services, hypervisors, standalone containers, and cloud instance types.
- Interactive matrix features: release comparison, deployment-type filters, OS/hypervisor/orchestration filters, footnote tooltips, and per-configuration release badges.
Current 8.0 infrastructure software versions
- NVIDIA Data Center GPU Driver: 595.58.03.
- NVIDIA DOCA Driver for Networking and DOCA Microservices: 3.3.0.
- NVIDIA Fabric Manager: integrated into AI Enterprise drivers at 595.58.03.
- NVIDIA vGPU for Compute Virtual GPU Manager and Guest Drivers: 20.0.
- NVIDIA Container Toolkit: 1.19.0.
- NVIDIA Run:ai: 2.25 in the AI Enterprise Infra 8.0 support matrix.
- NVIDIA DPU Operator (DPF): 25.10.1.
- NVIDIA GPU Operator: 26.3.0.
- NVIDIA Network Operator: 26.1.0.
- NVIDIA NIM Operator: 3.1.0.
- NVIDIA Base Command Manager: 11.32.1.
NVIDIA context
Use the support matrix before deploying or upgrading. The matrix is especially important for Blackwell and Grace Blackwell systems, vGPU for Compute, Kubernetes/OpenShift deployments, and cloud deployment paths where a supported instance, OS, driver, operator, and orchestration combination matters.
Connections
- NVIDIA-AI-Enterprise - parent product whose infrastructure compatibility is being validated.
- NVIDIA-AI-Enterprise-Software - software catalog that points to the support matrix for deployment compatibility.
- NVIDIA-AI-Enterprise-Lifecycle-Policy - branch timelines and compatibility explorer complement this matrix.
- NVIDIA-AI-Enterprise-Quick-Start-Guide - quick start sends users to support-matrix checks after first setup.
- NVIDIA-Certified-Systems - many on-premises configurations require certified systems; GB200/GB300 NVL systems use NVIDIA-qualified status.
- NVIDIA-DGX, NVIDIA-HGX, and NVIDIA-GB300-NVL72 - platform families represented in the support matrix.
- NVIDIA-Data-Center-GPU-Drivers, NVIDIA-DOCA, NVIDIA-vGPU, NVIDIA-vGPU-for-Compute, NVIDIA-MIG, and NVIDIA-Container-Toolkit - infrastructure foundations checked by the matrix.
- NVIDIA-Run-ai - self-hosted workload orchestration component listed in Infra 8.0 compatibility.
- NVIDIA-GPU-Operator, NVIDIA-Network-Operator, and NVIDIA-NIM-Operator - Kubernetes operators with release-specific support entries.
- NVIDIA-AI-Enterprise-Bare-Metal-Deployment, NVIDIA-AI-Enterprise-VMware-Deployment, and NVIDIA-AI-Enterprise-Cloud-Deployment - deployment guides whose prerequisites should be checked against the matrix.
Source Excerpts
- The infrastructure software docs say the matrix verifies supported AI Enterprise configurations across public cloud, data centers, workstations, DGX, and edge.
- The interactive matrix compares releases 7.0 through 8.0 and filters by deployment type, OS, hypervisor, and orchestration.
- The static 8.0 matrix lists supported infrastructure software versions including R595 driver, vGPU 20.0, Container Toolkit 1.19.0, Run:ai 2.25, and the current Kubernetes operators.