NVIDIA AI Enterprise Infra 6.2 Release Notes#

Release Highlights#

  • Blackwell Platform Support - R570 Data Center Driver and GPU Operator add support for NVIDIA HGX GB200 NVL72 and NVIDIA HGX B200 platforms

  • GPU Infrastructure - NVIDIA GPU Data Center Driver provides enhanced GPU acceleration and management capabilities for Hopper, Ada Lovelace, and Ampere architectures

  • Kubernetes Orchestration - GPU Operator and Network Operator enable streamlined deployment and lifecycle management for GPU workloads

  • Networking Performance - DOCA-OFED Driver for high-performance InfiniBand and Ethernet networking

  • Enterprise Management - Base Command Manager for cluster provisioning and workload orchestration

  • Platform Lifecycle Notice - Final release supporting vGPU C-Series drivers for VMware vSphere 7.x platform. VMware vSphere 7.x reaches End of Service (EoS) on October 2, 2025. For continued support, upgrade to a newer vSphere version. See VMware EoS announcement

What’s Included in NVIDIA AI Enterprise Infra 6.2#

Complete list of infrastructure components with versions and documentation links:

Table 6 Supported Infrastructure Software#

Product

Description

Version

Core Infrastructure Drivers

NVIDIA GPU Data Center Driver

GPU driver with Hopper, Ada Lovelace, and Ampere architecture support, including HGX GB200 NVL72 and HGX B200

570.133.20

NVIDIA DOCA-OFED Driver for Networking

High-performance networking for InfiniBand and Ethernet

2.10.0

Virtualization

NVIDIA vGPU for Compute (Virtual GPU Manager and Guest Drivers)

Enterprise GPU virtualization with enhanced performance and management capabilities

18.1

Container Platform

NVIDIA Container Toolkit

GPU-accelerated container runtime with enhanced security

1.17.5

Kubernetes Operators

NVIDIA GPU Operator

Automated GPU software lifecycle management for Kubernetes

25.3.0

NVIDIA Network Operator

Streamlined networking for GPU workloads in Kubernetes

25.1.0

NVIDIA NIM Operator

Simplified deployment and scaling of NVIDIA Inference Microservices

1.0.1

Cluster Management & Orchestration

NVIDIA Base Command Manager (BCM)

Enterprise cluster provisioning, workload orchestration, and lifecycle management

10.25.03

Compatibility and Support#

Refer to the NVIDIA AI Enterprise Infrastructure Support Matrix for:

  • Supported GPU architectures

  • Operating system compatibility

  • Hypervisor and orchestration platform versions

  • Cloud provider instance types