Overview#
The NVIDIA AI Enterprise lifecycle policy defines how NVIDIA releases, maintains, and supports AI software. Starting with NVIDIA AI Enterprise 4.0, AI frameworks and SDKs are decoupled from infrastructure management software and distributed through distinct release branches-each offering a different balance of innovation, stability, and support duration.
Attention
All branches are covered by NVIDIA AI Enterprise Support Services. Supported software is listed in the NVIDIA NGC Catalog and is identified by the NVIDIA AI Enterprise Supported icon.
NVIDIA supports individual NIMs and software components in Blueprints, AI Workflows, and Reference Releases, but not glue code, third-party applications, or full deployment integration. Support duration depends on the release branch type.
NVIDIA AI Enterprise software is organized into two layers, each with its own set of release branches:
Application Layer - AI frameworks, SDKs, NIMs, and pre-trained models released through Feature Branch (FB), Production Branch (PB), and Long-Term Support Branch (LTSB). Some products also offer Government Ready Containers.
Infrastructure Layer - NVIDIA GPU drivers, NVIDIA GPU Operator, NVIDIA Network Operator, and NVIDIA Run:ai self-hosted released through the Infrastructure Branch. NVIDIA Run:ai SaaS is also included in NVIDIA AI Enterprise but follows a separate, NVIDIA-managed cloud-service lifecycle - see NVIDIA Run:ai SaaS Lifecycle.
For help selecting the appropriate branch for your use case, refer to Choosing the Right Release Branch.