Is this page helpful?

NVIDIA AI Enterprise Documentation#

This documentation covers the NVIDIA AI Enterprise Infrastructure Layer software - GPU and network drivers, Kubernetes operators, NVIDIA vGPU for Compute, and NVIDIA Run:ai for AI workload management.

For application-layer software (NIM, NeMo, Omniverse, domain SDKs), refer to Application Software. For enterprise support services, refer to Support.

Quick Start#

🆕 What Is New in NVIDIA AI Enterprise Infra 8.2#

Latest Release Highlights

  • NVIDIA Run:ai 2.26 - Updated from 2.25 in 8.1. The same 2.26 release applies to both NVIDIA Run:ai self-hosted and NVIDIA Run:ai SaaS. Refer to the NVIDIA Run:ai release notes for scheduling, GPU-utilization, and platform updates in this version.

  • NVIDIA Data Center GPU Driver 595.91.07 - Maintenance update within the R595 production driver branch (from 595.71.05 in 8.1). NVIDIA Fabric Manager updates to 595.91.07 in lockstep (integrated into the NVIDIA AI Enterprise drivers). Refer to the 595.91.07 release notes for fixes and platform-support details.

  • NVIDIA vGPU for Compute Updates - NVIDIA Virtual GPU Manager (595.91.04, from 595.71.03 in 8.1) and the NVIDIA vGPU for Compute Guest Driver (Linux 595.91.07, from 595.71.05 in 8.1; Windows 596.86, from 596.36 in 8.1) refresh the full vGPU stack in a coordinated release.

    New in 8.2 for vGPU for Compute:

    • Newly supported hypervisor:

      • Red Hat Enterprise Linux 10.2

      • Red Hat Enterprise Linux 9.8

      • VMware vSphere 9.1

    • Newly supported guest operating systems:

      • Red Hat Enterprise Linux 10.2

      • Red Hat Enterprise Linux 9.8

    • MIG-backed Multi-vGPU on VMware vSphere - Supported with VMware Cloud Foundation (VCF) 9.1 or later. Refer to Multi-vGPU and P2P.

    • NVIDIA HGX B200 and HGX B300 on VMware vSphere - NVIDIA vGPU for Compute on HGX B200 and HGX B300 now includes VMware vSphere, in addition to Linux KVM. On VMware vSphere, only 1:1 vGPU VMs and multi-vGPU VMs are supported. Fractional vGPU VMs are not supported. Refer to the Support Matrix. On these platforms, NVSwitch support is delivered through Fabric Manager CRX. Refer to Installing Fabric Manager CRX.

  • Kubernetes Operator Updates - NVIDIA GPU Operator 26.3.3 (unchanged from 8.1), NVIDIA Network Operator 26.4.1 (from 26.4.0 in 8.1), NVIDIA DPU Operator (DPF) 26.4.0 (unchanged from 8.1), and NVIDIA NIM Operator 3.1.2 (from 3.1.1 in 8.1). NVIDIA Container Toolkit remains at 1.19.1 (unchanged from 8.1).

  • DOCA Ecosystem Updates - NVIDIA DOCA Driver for Networking 3.4.0 and NVIDIA DOCA Microservices 3.4.0 continue from 8.1.

  • Enterprise Management - NVIDIA Base Command Manager (BCM) 11.33.1 continues from 8.1 for cluster provisioning and workload orchestration.

View Full 8.2 Release Notes

Previous Releases#

📋 Release 8.1 Highlights

  • NVIDIA Run:ai SaaS Now Included - In addition to NVIDIA Run:ai self-hosted, the NVIDIA-managed Run:ai SaaS offering is now included in the NVIDIA AI Enterprise license under the same enterprise SLA. Refer to the NVIDIA Run:ai SaaS Documentation for the NVIDIA-managed cloud-service option, or choose the deployment that fits your environment.

  • NVIDIA Run:ai 2.25 - Updated from 2.24 in 8.0. The same 2.25 release applies to both NVIDIA Run:ai self-hosted and NVIDIA Run:ai SaaS. Refer to the NVIDIA Run:ai release notes for scheduling, GPU-utilization, and platform updates in this version.

  • NVIDIA Data Center GPU Driver 595.71.05 - Maintenance update within the R595 production driver branch (from 595.58.03 in 8.0). Refer to the 595.71.05 release notes for fixes and platform-support details.

  • NVIDIA vGPU Software - NVIDIA Virtual GPU Manager (595.71.03, from 595.58.02 in 8.0) and the NVIDIA vGPU for Compute Guest Driver (Linux 595.71.05, from 595.58.03 in 8.0; Windows 596.36, from 595.97 in 8.0) refresh the full vGPU stack in a coordinated release. New in 8.1 for vGPU for Compute:

    • Newly supported hypervisor: Ubuntu 26.04 LTS

    • Newly supported guest operating systems:

      • SUSE Linux Enterprise Server 15 SP6, 15 SP7, and 16

      • Ubuntu 26.04 LTS

  • Kubernetes Operator Updates - NVIDIA GPU Operator 26.3.3 (from 26.3.0 in 8.0), NVIDIA Network Operator 26.4.0 (from 26.1.0 in 8.0), NVIDIA DPU Operator (DPF) 26.4.0 (from 25.10.1 in 8.0), and NVIDIA NIM Operator 3.1.1 (from 3.1.0 in 8.0). NVIDIA Container Toolkit advances to 1.19.1 (from 1.19.0 in 8.0).

  • DOCA Ecosystem Updates - NVIDIA DOCA Driver for Networking 3.4.0 and NVIDIA DOCA Microservices 3.4.0 (both from 3.3.0 in 8.0) provide enhanced networking performance and infrastructure acceleration.

  • Enterprise Management - NVIDIA Base Command Manager (BCM) 11.33.1 (from 11.32.1 in 8.0) for cluster provisioning and workload orchestration.

  • Partner-Led Support - NVIDIA AI Enterprise is also available through SUSE AI Factory with NVIDIA, where SUSE delivers technical support directly for deployments on SLES. Refer to Partner-Led Support in the Support Matrix for details.

View Full 8.1 Release Notes

📦 Archived Releases (8.0)

Earlier NVIDIA AI Enterprise 8.x releases with key highlights:

  • 8.0 Release Notes - New GPU Driver Branch (R595), NVIDIA B300 NVL8 Support, NVIDIA RTX PRO 4500 Support, vGPU for Compute Updates, Updated Kubernetes Operators, DOCA Ecosystem Updates, Container Toolkit, Enterprise Management, Fabric Manager Integration