ai-cloud
GPU Cloud Platform with Hard Tenant Isolation
Build a GPU cloud customers can use independently. vMetal manages the machine layer, and vCluster combines separate tenant control planes with dedicated Private Nodes.
Trusted by the fastest-growing AI cloud providers
Problem
The GPU Cloud Platform Gap
GPU cloud providers face hard tradeoffs that slow revenue and erode margins.
Bare Metal Alone Falls Short
Selling raw compute is a race to the bottom. Customers want the cloud experience on top of the GPU specs.
DIY Platform Takes Too Long
Building a GPU cloud platform in-house requires significant engineering investment, making it slow to stand up and costly to maintain.
Isolation vs. Efficiency Tradeoff
Namespace isolation is too weak. Separate physical clusters per tenant are too expensive to operate at scale.
Solution
One Stack from Bare Metal to Tenant Clusters
vMetal manages the infrastructure layer and vCluster orchestrates Kubernetes tenant clusters. Private Nodes dedicate production workers, templates support onboarding, and vNode can add workload runtime isolation. Lintasarta runs 170+ tenant clusters in production.
Built for GPU Cloud Providers at Scale
Every layer your GPU cloud platform needs, from bare metal provisioning to workload isolation, in one integrated stack.
Tenant Isolation
Lightweight Tenant Kubernetes Control Planes
Each tenant has an independent Kubernetes API, data store and RBAC boundary. Lightweight control-plane hosting reduces dedicated server requirements for GPU cloud operations; Private Nodes keep production workers tenant-specific.
- Independent tenant API and data store
- Tenant-scoped scheduling on Private Nodes
- No dedicated control-plane servers per tenant
Bare Metal
Zero-Touch GPU Server Provisioning
vMetal exposes one Machine API above bare metal and VM provisioning drivers. Registered inventory, images and credentials enable repeatable machine lifecycle workflows for GPU cloud operations. Network automation uses supported integrations such as Netris.
- Stable Machine API
- Bare metal and VM provisioning
- Driver-specific lifecycle automation
Workload Security
Kernel-Level Isolation Without a Hypervisor
vNode uses Linux user namespaces and seccomp to restrict workload privileges and system calls. It adds runtime hardening for GPU cloud operations without a guest kernel or hypervisor, complementing dedicated workers rather than allocating GPUs.
- Linux user namespaces and seccomp
- No guest kernel or hypervisor
- Helps limit workload breakout
AI Workload Integration
Supported AI Workload Templates
Use supported workload templates for GPU cloud operations, 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.
- Native NVIDIA Run:AI catalog
- Validate Private Node configuration
- Release-specific integration prerequisites
Customer Experience
Template Based Portal for Your Customers
Give your GPU cloud customers a self-service portal to provision and manage their own environments. Deliver the managed Kubernetes experience AI teams expect without building a custom frontend.
- Approved templates and access policies
- Dedicated production tenant workers
- Readiness depends on available capacity
Why vCluster
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.
100K+
GPUs Powered
50+
GPU Clouds & F500s
<45
Days to Launch
30K
GitHub Stars
Get Started in 3 Steps
1
Schedule a Demo
Talk to our team about your stack
2
Deploy vCluster
Deploy vCluster on your infra in minutes
3
Onboard Your Tenants
Go live with a hyperscaler-grade tenant experience in days
FAQs
What makes vCluster different from building a GPU cloud platform in-house?
Evaluate vCluster against the service you need to deliver: separate tenant Kubernetes APIs, dedicated production workers and infrastructure lifecycle operations. Private Nodes provide the recommended customer worker model, while vMetal orchestrates machines and Platform coordinates templates and access. Compare documented versions and configurations directly; an embedded runtime alone does not describe every capability of another vendor's platform.
How does vCluster provide tenant isolation on GPU hardware?
vCluster virtualizes the Kubernetes control plane. Each tenant gets its own API server, data store, scheduler, and RBAC running as lightweight pods inside a control plane cluster, without the cost of a separate physical cluster per customer. Private Nodes give each tenant dedicated worker nodes with per-tenant CNI and storage. For your own internal dev, test, and CI/CD, Shared Nodes are an option.
Is vCluster certified for production Kubernetes workloads?
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.
Can vCluster run directly on bare metal GPU servers?
Yes. vCluster Standalone runs as a single binary directly on Linux bare metal, with no external Kubernetes dependency and no k3s, kubeadm, or k0s base layer. vMetal adds zero-touch provisioning underneath, handling PXE boot, OS installation, machine registration, and network automation from rack to production.
What GPU cloud providers are using vCluster in production?
vCluster Labs software powers 100K+ GPUs and 1M+ CPUs, serving 50+ GPU clouds & Fortune 500s combined. In production, Lintasarta operates 170+ tenant clusters, while Boost Run completed its managed Kubernetes launch in under 45 days from the decision. These named deployments provide scale and launch examples for teams evaluating the platform for their own infrastructure.
What isolation models does vCluster support for a GPU cloud platform?
Every customer-facing tier runs on Private Nodes, the recommended production model, where each tenant gets dedicated hardware with its own CNI and CSI. vNode can add workload isolation with Linux user namespaces and seccomp for stricter tiers, without VM overhead. GPU partitioning such as MIG or time-slicing sits at the GPU layer, and vCluster puts the tenant boundary around it.
Launch Your GPU Cloud Platform Faster
Talk to us about turning your GPU racks into a self-service cloud for your customers.
