# NVIDIA Opens NVKE and NVKR in Public Preview for GPU Clusters By Simon Yoon Canonical URL: https://www.tokenpost.com/news/technology/54177 Published: 2026-11-11T15:54:18.000Z Updated: 2026-11-11T15:54:18.000Z Section: Technology > NVKE provisions workload clusters, while NVKR builds reproducible images and packages. The Apache 2.0-licensed tools are not recommended for production use during the preview. NVIDIA has opened two Apache 2.0-licensed tools in public preview to help AI factory operators build and manage Kubernetes clusters on GPU infrastructure, with separate tools for cluster provisioning and software-image builds. NVIDIA announced the projects Tuesday, Nov. 10, as part of its DSX AI Factory Platform. NVIDIA Kubernetes Engine (NVKE) provides APIs and a management plane for provisioning and managing clusters. The setup is designed to let operators run a managed Kubernetes service while retaining control over their software, operations and customer experience. Operators use the nvkectl tool to bootstrap an NVKE management cluster and provision workload clusters. Each workload cluster has its own Kubernetes control plane, etcd and node pools. NVKE uses NVIDIA Infra Controller (NICo) for bare-metal lifecycle management. Monitoring uses an OpenTelemetry-based telemetry pipeline. NVIDIA Kubernetes Recipes (NVKR) builds reproducible operating system images, container images and software packages tailored to GPU generations and Kubernetes versions. Its builds include checksum-pinned Kubernetes 1.36 packages and GPU-specific runtime components, and record their inputs. The tools can be used together or independently: NVKE does not depend on NVKR at runtime, and NVKR can produce artifacts for other Kubernetes management systems. AI Cluster Runtime and NVSentinel are among the additional components that can support GPU-accelerated runtime and automation within clusters. During the public preview, APIs, configurations and behavior may change between releases. The tools are not recommended for production workloads. The launch adds cluster provisioning and image-building tools to NVIDIA’s Kubernetes work. TokenPost previously covered NVIDIA’s [three-stage DSX OS workflow for GPU Kubernetes clusters](). ## Links in this article - [three-stage DSX OS workflow for GPU Kubernetes clusters](https://www.tokenpost.com/news/technology/50597)