Reserved GPU capacity only means something if the reserved nodes belong to one customer. vCluster assigns Private Nodes to that customer's tenant cluster and applies capacity policy centrally, and vMetal can add bare metal Machines when needed.
Reserved GPU services need clear ownership of worker capacity and a consistent way to deliver it to tenant clusters.
A separate Kubernetes management stack for every tenant increases infrastructure and operational work.
Namespace-only designs do not dedicate worker capacity or provide a separate tenant control plane.
Custom capacity workflows create manual handoffs between tenant requests, Machine provisioning, and cluster attachment.
vCluster Platform assigns Private Nodes to production tenant clusters and applies tenant access and capacity policy centrally. vMetal can provision physical Machines when additional bare metal capacity is required.
Reserved GPU capacity combines dedicated Private Nodes, tenant control planes, capacity policy, and optional Machine provisioning.
Private Nodes dedicate worker capacity, networking, and storage to one production tenant cluster at a time.
Each tenant receives its own virtualized Kubernetes control plane, API server, and RBAC boundary.
Auto Nodes can create and delete Machine requests to maintain configured Private Node capacity based on supported resource requirements.
vMetal provisions physical Machines behind a stable API and can join completed servers to the target tenant cluster automatically.
Manage tenants, access, templates, quotas, capacity policy, and fleet observability through one UI, CLI, and API.
This isn’t a side project. Behind every vCluster deployment is 5+ years of deep K8s engineering, security hardening, and battle-tested infrastructure work at massive scale.
Talk to our team about your stack
Deploy vCluster on your infra in minutes
Go live with a hyperscaler-grade tenant experience in days
Private Nodes assign dedicated worker capacity to one production tenant cluster at a time. Each tenant also gets its own virtualized control plane and RBAC boundary, while capacity policy controls which node types the tenant can request.
No separate Kubernetes control-plane servers are required for each tenant. Private Nodes provide dedicated worker capacity, and the tenant control plane runs as isolated pods on the control plane cluster.
Reserved capacity can back managed Kubernetes, Slurm, Ray, inference and other cluster products, including NVIDIA Run:ai through our partnership. vCluster provides the tenant boundary, while Slurm and NVIDIA Run:ai keep responsibility for their own scheduling.
Tenant clusters come up in minutes, not hours, once the reserved nodes exist. If capacity must be provisioned first, the selected Machine driver determines infrastructure time. Boost Run went from decision to production launch in under 45 days.
Private Nodes provide tenant-scoped networking. When Metal3 and Netris are configured, separate network environments can also provide hardware-backed L2 isolation.
vCluster Labs software powers 100K+ GPUs across 50+ GPU clouds and Fortune 500s and is validated in NVIDIA's DGX reference architecture.
See how vCluster Platform turns dedicated GPU nodes into reserved capacity you can sell per tenant.
