ai-cloud
EKS on Bare Metal Foundation for Your AI Cloud
Launch an EKS-equivalent managed Kubernetes service on your GPU hardware. vCluster delivers fully isolated tenant clusters with a self-service experience on bare metal.
Trusted by the fastest-growing AI cloud providers
Problem
Why Bare Metal Kubernetes Falls Short
Building an EKS-like experience on bare metal exposes three hard infrastructure gaps.
Raw Compute Is Not Enough
Customers want cloud-native tooling and self-service environments, not raw GPUs. Selling bare metal alone is a race to the bottom.
DIY Takes Years and Millions
Building a GPU cloud platform yourself takes 6 to 10 engineers, 6 to 12 months, and over a million dollars. Most teams are still building two years in.
Isolation Without a Managed Layer
Namespace isolation is too weak. Separate physical clusters per tenant are too expensive. EKS on bare metal demands a better middle path.
Solution
vCluster Delivers EKS on Bare Metal in Days
vCluster virtualizes the Kubernetes control plane so every tenant gets a real, CNCF-certified K8s cluster on your bare metal hardware. From zero-touch GPU provisioning to a self-service portal, the full stack is production-proven across 100K+ GPU nodes and 50+ GPU clouds.
Everything You Need for EKS on Bare Metal
From bare metal GPU provisioning to tenant isolation and self-service portals, vCluster covers the full stack.
Bare Metal Layer
Zero-Touch GPU Server Provisioning
vMetal handles PXE boot, OS installation, machine registration, and network configuration automatically. Go from GPU rack to production-ready Kubernetes nodes without manual intervention or external dependencies.
- PXE boot and OS install
- Full machine lifecycle management
- Network automation included
K8s Distribution
Lightweight K8s Directly on Bare Metal
vCluster Standalone runs as a single binary directly on your bare metal servers. No k3s, kubeadm, or external Kubernetes dependency required as a base layer.
- Single binary on bare metal
- No k3s or kubeadm needed
- Replaces legacy K8s base layers
Tenant Isolation
Isolated Tenant Clusters in Seconds
Each tenant gets their own Kubernetes API server, etcd, scheduler, and RBAC running as lightweight pods on your host cluster. Spin up hundreds of isolated tenant environments with near-zero marginal cost.
- Own API server per tenant
- Spins up in seconds
- No separate physical clusters needed
Cloud Experience
EKS-Like Self-Service for Tenants
Deliver the cloud experience your customers expect. Tenants provision and manage their own Kubernetes environments through a self-service portal that mirrors the EKS and GKE experience.
- EKS-like tenant portal
- Self-service cluster provisioning
- SSO, quotas, and templates built in
Dynamic Scaling
Auto-Provision GPU Nodes on Demand
Auto Nodes acts as a bare metal Karpenter, automatically provisioning GPU servers via Terraform when tenants schedule workloads. Scale GPU capacity dynamically without manual intervention.
- Automatic GPU node provisioning
- Terraform-driven bare metal scaling
- Triggered by tenant workload demand
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+
GPU Nodes 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
Can vCluster replace EKS for workloads running on bare metal?
Yes. vCluster delivers a CNCF-certified, fully conformant Kubernetes experience on bare metal hardware without requiring a cloud provider. Each tenant gets their own API server, etcd, and RBAC, making it a direct equivalent to EKS for teams that need managed Kubernetes on their own GPU infrastructure. The vCluster Standalone binary runs directly on Linux with no external Kubernetes dependency, so you are not layering k3s or kubeadm underneath.
How long does it take to launch a managed Kubernetes service on bare metal?
Boost Run launched a managed Kubernetes service in under 45 days using vCluster, with no new platform engineering hires. Lintasarta launched Indonesia's leading GPU cloud in 90 days. The full stack, from bare metal provisioning with vMetal through tenant cluster orchestration, is designed to get you to production in days rather than quarters.
What isolation model does vCluster use for EKS on bare metal?
vCluster offers a flexible isolation spectrum from shared nodes with namespace boundaries to private nodes where each tenant gets dedicated physical hardware with their own CNI and CSI. For workload-level security, vNode (currently in private beta) adds kernel-native isolation using seccomp, cgroups, and AppArmor without hypervisor overhead. This means you can match isolation strength to each tenant's requirements without sacrificing GPU performance.
Does vCluster support GPU workloads and AI platforms on bare metal?
Yes. vCluster is production-proven across 100K+ GPU nodes and is named in the NVIDIA DGX SuperPOD reference architecture. Certified Stacks provide pre-validated AI environments for Run:AI, Ray, and Jupyter, so teams can go from a bare Kubernetes cluster to a production AI platform in minutes rather than weeks.
What happens to tenant isolation when running EKS on bare metal with vCluster?
Every tenant receives a dedicated Kubernetes control plane, cluster-admin access within their own cluster, and full CRD and RBAC freedom without affecting neighboring tenants. Hardware-enforced tenant networking via VLANs, VXLANs, and VRFs ensures network boundaries are maintained. Tenants are strongly isolated from platform internals, other tenants' nodes, and pods outside their own environment.
Is vCluster open source and how is it deployed?
vCluster has an open-source core licensed under Apache 2.0 with 29.8K GitHub stars. The platform can be deployed on your own bare metal infrastructure in an air-gapped or FIPS-compliant environment. It supports GitOps and IaC workflows via Terraform, Argo CD, and CI/CD pipelines, so your existing automation tooling integrates without re-architecture.
Launch EKS on Bare Metal with vCluster
See how GPU cloud providers go from bare metal to managed Kubernetes in days.
