NVIDIA donates OpenShell agent security runtime to CNCF and releases DSX Kubernetes tools in public preview

At KubeCon on November 10, 2026, NVIDIA contributed OpenShell to the Cloud Native Computing Foundation as a Sandbox project. OpenShell is an open-source runtime that applies deterministic controls to what autonomous agents can do: which tools they can use, which inference services they can reach and which workloads they can run, with hooks for a team's own compute, credential storage and monitoring. NVIDIA says hundreds of organizations already use OpenShell and the project has received tens of thousands of contributions in less than a year since launch.

NVIDIA is also contributing to Agent Substrate, the Google-led open-source framework for running long-running agent workloads on Kubernetes, which has also entered the CNCF Sandbox. Agent Substrate preserves an agent's environment between tasks while releasing compute when the agent is idle, and NVIDIA is wiring it to its Dynamo inference platform and KAI Scheduler, plus an integration so OpenShell can govern agents running on Agent Substrate across clouds and on-premises systems.

For the infrastructure underneath, NVIDIA released two components of its DSX AI Factory Platform in public preview: NVIDIA Kubernetes Engine (NVKE) and NVIDIA Kubernetes Recipes (NVKR), open-source building blocks for managed Kubernetes services on GPU infrastructure with signed, version-pinned images. NVIDIA Snapshot, also open source, restores checkpointed inference replicas in seconds without reloading models: design partner Crusoe cut the time for a GLM-5.3 workload to become ready from roughly eight minutes to 31 seconds, and Saturn Cloud cut vLLM startup times from between two and 8.5 minutes to between six and 34 seconds.