ai-cloud
Kubernetes for AI Workloads at Cloud Scale
Give AI teams room to train and serve models. vMetal provisions infrastructure, and vCluster runs tenant workloads on dedicated Private Nodes with independent control planes.
Trusted by the fastest-growing AI cloud providers
Problem
The Hard Tradeoffs in AI Kubernetes
Scaling Kubernetes for AI workloads forces painful choices most platforms cannot resolve.
Isolation vs. Efficiency
Namespace isolation is too weak for GPU tenants. Separate physical clusters are too expensive and slow to provision at scale.
DIY Takes Too Long
Building a GPU cloud platform in-house requires significant engineering investment.
Customers Expect Cloud UX
AI teams have used AWS and GCP. They expect self-service environments and managed Kubernetes, and they will leave if you cannot deliver it.
Solution
The Full Stack for Kubernetes AI Workloads
vMetal provisions infrastructure for training and inference. vCluster gives each team a Kubernetes tenant cluster with its own API and dedicated Private Nodes, while teams install their chosen workload tooling. vCluster Labs software powers 100K+ GPUs.
Built for Kubernetes AI Workload Scale
Every layer of the stack is built for running AI workloads on GPU infrastructure with strong tenant isolation, without giving up GPU performance.
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 training and inference; 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
AI Workload Integration
Supported AI Workload Templates
Use supported workload templates for training and inference, 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
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 training and inference. Network automation uses supported integrations such as Netris.
- Stable Machine API
- Bare metal and VM provisioning
- Driver-specific lifecycle automation
Workload Security
Kernel-Native AI Workload Isolation
vNode uses Linux user namespaces and seccomp to restrict workload privileges and system calls. It adds runtime hardening for training and inference 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
Dynamic Scaling
Auto-Provision GPU Nodes on Demand
Karpenter-based Auto Nodes requests worker capacity according to configured policies. With the vMetal provider, Machine requests provision available infrastructure for training and inference; inventory, driver readiness and release rules determine actual capacity.
- Karpenter-based node management
- Supported infrastructure provider integration
- Inventory-aware capacity policies
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
How does vCluster isolate Kubernetes AI workloads between tenants?
vCluster gives each tenant an independent Kubernetes API, data store and RBAC scope. Private Nodes dedicate worker capacity to that tenant, including its own CNI and storage configuration. vMetal orchestrates the machines underneath. This separates tenant administration and worker placement; shared infrastructure, application permissions and network policies still require an explicit security design.
What makes vCluster different from running standard Kubernetes for AI workloads?
Namespaces divide resources within a Kubernetes cluster, but share its API and cluster-wide administrative scope. RBAC, quotas and network policies remain useful controls. vCluster adds a separate tenant API and data store, with Private Nodes dedicating production workers. This reduces dependencies between tenant environments while platform administrators still manage the underlying infrastructure and its failure domains.
Can vCluster run on bare metal GPU servers without an existing Kubernetes cluster?
Yes. vCluster Standalone runs as a single binary directly on bare metal Linux with no dependency on k3s, kubeadm, or any external Kubernetes control plane binary. vMetal handles zero-touch bare metal provisioning covering PXE boot, OS installation, network automation, and full GPU server lifecycle management. This means you can go from a rack of GPU servers to production-ready Kubernetes for AI workloads without standing up an intermediate infrastructure layer.
Which AI platforms and frameworks work with vCluster tenant clusters?
NVIDIA Run:AI is available through the native Certified Stacks integration, with release-specific prerequisites and templates. Its bundled templates do not automatically supply dedicated Private Nodes. Configure and validate the production worker model, GPU access, networking and storage before offering it to customers. Ray and Jupyter can use compatible Kubernetes deployment tooling and require their own supported configuration.
How quickly can an AI cloud provider go live with managed Kubernetes using vCluster?
Boost Run went from decision to production launch in under 45 days with vCluster, and Lintasarta runs 170+ tenant clusters in production. Building a GPU cloud platform in-house takes significant time and engineering. vCluster covers the path from bare metal provisioning to tenant cluster orchestration to workload isolation in one stack.
Is vCluster's Kubernetes control plane certified and compatible with standard tooling?
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.
Launch Your Kubernetes AI Cloud Faster
See how vCluster powers AI workloads across 50+ GPU clouds & Fortune 500s.
