vCluster Labs Turns Open-Source NVIDIA Run:ai Into a Managed Cluster Product

NVIDIA is open sourcing the full NVIDIA Run:ai platform. vCluster Labs gives AI cloud providers and AI factories an isolated, ready-to-run NVIDIA Run:ai environment for every tenant.

SAN FRANCISCO--(BUSINESS WIRE)--vCluster Labs, the platform operators use to build and run their own AI clouds, today announced managed NVIDIA Run:ai clusters built on the newly open-source NVIDIA Run:ai platform. NVIDIA Run:ai is the AI workload and GPU orchestration platform within NVIDIA DSX OS, the open, modular software portfolio of the NVIDIA DSX AI Factory Platform.

The open-source release enables vCluster Labs to integrate NVIDIA Run:ai more directly into the vCluster platform, giving customers a simple path to access, deploy, and operate NVIDIA Run:ai as part of a complete solution. No commercial NVIDIA Run:ai software license is required.

vCluster provides infrastructure-level tenant isolation, while NVIDIA Run:ai schedules workloads and allocates GPU capacity across teams and projects within each tenant. Together, they give AI clouds a repeatable way to deliver managed NVIDIA Run:ai clusters across many customers while maximizing GPU utilization within each environment. vCluster automatically deploys and configures NVIDIA Run:ai, and maintains consistent operations across the fleet as the tenant count grows. The result is a repeatable NVIDIA Run:ai cluster product that operators can deploy and sell consistently.

"Open sourcing NVIDIA Run:ai puts it within reach of every team that wants it, which is a real change for this market," said Lukas Gentele, Co-Founder and CEO of vCluster Labs, a member of NVIDIA Inception. "What we add is scale. Standing NVIDIA Run:ai up for one team is a project. Doing it for fifty customers, each in their own isolated environment, is something else entirely. That is what lets an operator sell NVIDIA Run:ai as a product."

What vCluster Labs adds to NVIDIA Run:ai

vCluster Labs extends the value of NVIDIA Run:ai with per-tenant isolation, pre-validated deployment templates, and ongoing operations, helping customers deliver NVIDIA Run:ai to many teams and customers at once.

Customer-ready isolation. Each NVIDIA Run:ai tenant gets its own cluster with its own API and its own Private Nodes, isolated at the network, storage, and compute layers. That is the isolation an operator needs to hand an NVIDIA Run:ai cluster to an external customer.

One-step deployment. Four pre-validated NVIDIA Run:ai Certified Stacks ship with the platform, covering the patterns operators actually run, from a single self-hosted cluster to a shared control plane serving many tenants. Standing up NVIDIA Run:ai becomes a deployment rather than an integration.

Operations without added headcount. Health, logs, upgrades, and clean tenant removal are handled by the platform, so an operator can add customers without adding people to keep each environment running.

"The teams building AI infrastructure today are not competing on GPUs alone. They are competing on how quickly a researcher or a customer gets to a working environment," said Omri Geller, Vice President at NVIDIA. "When a partner like vCluster Labs can deploy NVIDIA Run:ai as a repeatable, governed environment for every tenant, that speed reaches a far wider set of customers than it otherwise would."

Supporting resources

• Blog: Stronger tenant isolation for NVIDIA Run:ai
• Documentation: NVIDIA Run:ai Certified Stacks
• Reference Architecture: vCluster on NVIDIA DGX

About vCluster Labs

vCluster Labs is the platform operators use to build and run their own AI cloud, like a hyperscaler. vCluster turns raw GPUs into cluster products they can sell, from bare metal provisioning up through managed Kubernetes, Slurm, and inference clusters, with tenant isolation at the infrastructure layer. The result is more margin per GPU, less dependence on a few big customers, and new managed services live in days instead of months. Trusted by fast-growing AI cloud providers including Nebius, Groq, Firmus, and Corvex, and by enterprises including Adobe, Samsung, and Deloitte, across 100K+ GPUs.