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

• 🆕 New to NVIDIA AI Enterprise? → Start with the Quick Start Guide to deploy your first AI workload

• ⬆️ Upgrading from 8.x or earlier? → Follow the Upgrade from 8.0 to 8.1 checklist; for highlights, refer to What Is New in 8.1 in the following section

• 🔧 Planning a deployment? → Refer to Planning & Deployment on the NVIDIA AI Enterprise Docs Hub for reference architectures and sizing guidance

• 🖥️ Setting up vGPU? → Refer to Installing NVIDIA vGPU for Compute and NVIDIA vGPU for Compute Licensing

🆕 What Is New in NVIDIA AI Enterprise Infra 8.1

Latest Release Highlights

• NVIDIA Run:ai SaaS Now Included — In addition to NVIDIA Run:ai self-hosted, the NVIDIA-managed NVIDIA 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 20.1 — NVIDIA Virtual GPU Manager and the NVIDIA vGPU for Compute Guest Driver are both updated from 20.0 to 20.1, refreshing 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 Patch Updates — NVIDIA GPU Operator 26.3.1 (from 26.3.0 in 8.0) and NVIDIA Network Operator 26.1.1 (from 26.1.0 in 8.0). NVIDIA DPU Operator (DPF) 25.10.1, NVIDIA NIM Operator 3.1.0, and NVIDIA Container Toolkit 1.19.0 carry forward from 8.0.

View Full 8.1 Release Notes

Previous Releases

📋 Release 8.0 Highlights

• New GPU Driver Branch (R595) — NVIDIA Data Center GPU Driver 595.58.03 introduces the R595 driver branch with support for new Blackwell platforms.

• NVIDIA B300 NVL8 Support (HGX Blackwell) — NVSwitch-connected 8-GPU topology with NVLink multicast support. Supports MIG-backed and time-sliced vGPU configurations.

• NVIDIA RTX PRO 4500 Support — New hardware SKU supported across bare metal and virtualized deployments.

• vGPU for Compute Updates — NVIDIA vGPU Manager and Guest Driver 20.0 with B300 NVL8 support, MIG-backed and time-sliced vGPU configurations, and NVSwitch multicast. This release introduces vGPU for Compute support for HGX B300 (Linux KVM) and RTX PRO 4500 (Linux KVM and vSphere).

• Updated Kubernetes Operators — GPU Operator 26.3.0, Network Operator 26.1.0, DPU Operator 25.10.1, and NIM Operator 3.1.0 for GPU workload lifecycle and deployment in Kubernetes

• DOCA Ecosystem Updates — DOCA Driver 3.3.0 and DOCA Microservices 3.3.0 provide enhanced networking performance and infrastructure acceleration

• Container Toolkit — NVIDIA Container Toolkit 1.19.0 with updated runtime components for GPU-accelerated containers

• Enterprise Management — Base Command Manager 11.32.1 offers refined cluster provisioning and workload orchestration for large-scale AI infrastructure

• Fabric Manager Integration — Fabric Manager and Fabric Manager development binaries are now integrated into the NVIDIA AI Enterprise drivers, eliminating the need for separate installation. NVIDIA NVLink System Monitor (NVLSM) continues as a standalone utility.

• NVIDIA RTX PRO 6000 Blackwell Server Edition — Available in air‑cooled and liquid‑cooled form factors. Both variants support the same vGPU for Compute profiles. Refer to Blackwell Architecture vGPU Types for details.

• vGPU for Compute Licensing Alignment — Starting with vGPU 20.0, the NVIDIA Licensing Service (NLS) license checkout and entitlement workflow now reflects the vGPU for Compute product name that was introduced with NVIDIA AI Enterprise Infra release 7.0 documentation. No action is required for existing deployments; licenses are checked out under the updated product name automatically.

View Full 8.0 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