Choose GPU orchestration tools for infrastructure and tenants. vMetal handles machines, and vCluster provides independent tenant clusters with dedicated Private Nodes beneath workload schedulers.
Generic tools were not built for the demands of AI cloud infrastructure at scale.
Selling raw GPU compute is a race to the bottom. Customers want managed Kubernetes and cloud-native tooling, not just hardware.
Building a GPU cloud platform in-house requires significant engineering investment.
Namespace isolation is too weak for real tenants. Separate physical clusters per tenant are too expensive and slow to provision.
vMetal orchestrates machines and vCluster orchestrates tenant Kubernetes environments with dedicated Private Nodes. Job schedulers and AI frameworks run within those environments; they retain responsibility for workload placement. Lintasarta runs 170+ tenant clusters in production.
Every layer of the stack from bare metal provisioning to tenant isolation is purpose-built for GPU cloud operators.
Each tenant has an independent Kubernetes API, data store and RBAC boundary. Lightweight control-plane hosting reduces dedicated server requirements for infrastructure orchestration; Private Nodes keep production workers tenant-specific.
vMetal exposes one Machine API above bare metal and VM provisioning drivers. Registered inventory, images and credentials enable repeatable machine lifecycle workflows for infrastructure orchestration. Network automation uses supported integrations such as Netris.
Karpenter-based Auto Nodes requests worker capacity according to configured policies. With the vMetal provider, Machine requests provision available infrastructure for infrastructure orchestration; inventory, driver readiness and release rules determine actual capacity.
Use supported workload templates for infrastructure orchestration, including the NVIDIA Run:AI partner integration. Configure GPU, network and storage prerequisites, and validate the selected template for Private Nodes before offering it to production customers.
vNode uses Linux user namespaces and seccomp to restrict workload privileges and system calls. It adds runtime hardening for infrastructure orchestration without a guest kernel or hypervisor, complementing dedicated workers rather than allocating GPUs.
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
vCluster virtualizes the Kubernetes control plane, running each tenant's cluster as a lightweight process instead of a separate physical cluster. That gives you hard tenant isolation at a low control plane cost per tenant. The stack covers bare metal provisioning (vMetal), tenant cluster orchestration (vCluster) and workload isolation (vNode), and in our analysis few vendors deliver all three as one integrated stack.
Yes. vCluster Standalone runs as a single binary directly on bare metal Linux with no external Kubernetes dependency. There is no need for k3s, kubeadm or k0s as an underlying base layer. Combined with vMetal for zero-touch provisioning, you can go from a GPU rack to a production Kubernetes environment without intermediate infrastructure.
Boost Run went from decision to production launch of a managed Kubernetes service in under 45 days, and Lintasarta runs 170+ tenant clusters in production on vCluster. Compare that with an in-house build, which needs a dedicated platform team and a long runway before the first tenant goes live.
Private Nodes are the recommended production model, giving each tenant dedicated worker nodes with its own CNI and CSI, which removes noisy-neighbor contention from other tenants. Shared Nodes are an option for trusted internal teams running dev, test and CI/CD. vNode adds workload isolation with user namespaces and seccomp, with no hypervisor overhead.
vCluster is a CNCF-certified Kubernetes distribution and exposes standard Kubernetes APIs in tenant environments. This supports familiar clients, Helm charts and cluster-scoped resources such as CRDs. Certification covers Kubernetes conformance rather than an entire AI platform or customer deployment. Validate the selected version, networking, storage, GPU drivers and runtime integrations for the workloads you plan to support.
vCluster Platform can coordinate tenant environments across supported clouds and data centers. Each Kubernetes tenant has its own control plane; a central fleet interface does not mean one tenant API or one failure domain. Plan worker placement, routing, storage and data residency for each location. An optional encrypted VPN can connect control planes and workers where appropriate for the deployment.
See how vCluster powers cloud GPU orchestration for AI cloud providers at scale.
