NVIDIA Adds Open Source Kubernetes Tools for AI Factories
NVIDIA unveiled three new open source components for its DSX AI Factory Platform at KubeCon + CloudNativeCon North America 2026.
Updated on Nov. 10, 2026 in Artificial Intelligence
NVIDIA announced three new open source components of its NVIDIA DSX AI Factory Platform: NVIDIA Kubernetes Engine (NVKE), NVIDIA Kubernetes Recipes (NVKR) and NVIDIA Snapshot. NVKE and NVKR are in public preview and provide building blocks for managed Kubernetes services on GPU infrastructure.
Why it matters
AI factories need to offer customers managed Kubernetes services while retaining control of their infrastructure. The announcement comes alongside plans from Mirantis, Netris, OpenNebula Systems, Rafay, Spectro Cloud, vCluster, Firmus and SK Telecom to bring NVKE and NVKR capabilities into their offerings.
Crusoe tested GLM-5.3 on Crusoe Cloud and cut workload readiness from roughly eight minutes to 31 seconds using NVIDIA Snapshot, while Saturn Cloud reduced vLLM startup times from between two and 8.5 minutes to between six and 34 seconds. The source does not detail the test methodology or whether these results apply beyond the two companies' setups.
The players
NVIDIA
NVIDIA is a American technology company known for its graphics processing units, which have become central to training and running artificial intelligence models.
Google is a multinational technology company that originated the Agent Substrate project now being contributed to by NVIDIA.
Mirantis
Mirantis is a cloud infrastructure software company that plans to bring NVKE and NVKR capabilities into its offerings.
The details
NVIDIA Snapshot restores checkpointed replicas in seconds without model reload or warmup, supporting faster startup for inference and agentic workloads on Kubernetes. NVIDIA is also contributing to Agent Substrate, a Google-originated CNCF Sandbox project, and is collaborating with Google on integrating Agent Substrate with OpenShell.
Timeline
KubeCon + CloudNativeCon North America 2026: NVIDIA announced the three new open source components.
The Tech Race
NVIDIA's contribution to Agent Substrate, a Google-originated CNCF Sandbox project, places it inside an open-source effort that Google helped start, rather than a purely proprietary stack. The move positions NVIDIA's AI factory tooling alongside Kubernetes-based infrastructure that cloud providers are already building on.
Cloud providers and enterprises running AI workloads may see faster model startup times if they adopt NVKE, NVKR or NVIDIA Snapshot, which are in public preview. Customers of partner platforms such as Mirantis, Rafay and Spectro Cloud could gain managed Kubernetes services built on these components.
The takeaway
Faster model startup and managed Kubernetes built on open source components could lower the operational burden of running AI workloads at scale. Teams evaluating AI infrastructure should watch how these preview tools perform outside the early adopters named in the announcement.
Further reading
Explore more Artificial Intelligence coverage for context on how infrastructure tools are shaping AI deployment.
