ai-cloud
Bare Metal Kubernetes for AI Cloud Providers
Deploy Kubernetes directly on GPU servers. vMetal orchestrates machine provisioning, Standalone supplies the Kubernetes binary, and vCluster dedicates tenant workers through Private Nodes.
Trusted by the fastest-growing AI cloud providers
Problem
Why Bare Metal Kubernetes Is Hard
Traditional approaches force costly tradeoffs between performance, isolation, and time to market.
VM Overhead Kills GPU Performance
Bare metal removes a hypervisor layer, but GPU drivers, storage and networking still need configuration and workload validation.
Namespace Isolation Is Too Weak
Tenants can see platform internals they should not: cluster-wide agents, other tenants' nodes and pods.
Building a Platform Requires Investment
Building a GPU cloud platform in-house requires significant engineering investment.
Solution
One Stack from Bare Metal to Tenant Clusters
vCluster Standalone supplies Kubernetes as a binary on Linux. vMetal orchestrates machine provisioning from an established platform installation, and tenant clusters use Private Nodes for dedicated workers. Lintasarta runs 170+ tenant clusters in production.
Built for Bare Metal Kubernetes at GPU Scale
Every layer of the stack is purpose-built for AI cloud providers running workloads directly on GPU hardware.
Bare Metal
Zero-Touch Bare Metal Provisioning
vMetal exposes one Machine API above bare metal and VM provisioning drivers. Registered inventory, images and credentials enable repeatable machine lifecycle workflows for bare metal Kubernetes. Network automation uses supported integrations such as Netris.
- Stable Machine API
- Bare metal and VM provisioning
- Driver-specific lifecycle automation
K8s Control Plane
Kubernetes as a Single Binary
vCluster Standalone runs Kubernetes as a CNCF-certified binary on Linux, without an existing Kubernetes cluster. It supplies a Kubernetes foundation for bare metal Kubernetes; vMetal separately orchestrates machine and OS provisioning.
- CNCF-certified Kubernetes binary
- Runs directly on Linux
- No existing Kubernetes cluster required
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 bare metal Kubernetes; 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
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 bare metal Kubernetes; inventory, driver readiness and release rules determine actual capacity.
- Karpenter-based node management
- Supported infrastructure provider integration
- Inventory-aware capacity policies
Workload Security
Kernel-Native Workload Isolation
vNode uses Linux user namespaces and seccomp to restrict workload privileges and system calls. It adds runtime hardening for bare metal Kubernetes 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
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 using k3s or kubeadm on bare metal?
vCluster Standalone runs as a Kubernetes binary directly on Linux without requiring another Kubernetes cluster, k3s or kubeadm. It can provide an initial Kubernetes foundation. vMetal is a separate infrastructure orchestration layer that requires vCluster Platform and a connected Kubernetes cluster for its Metal3/Ironic path. Hardware registration, images, credentials and driver setup remain prerequisites for machine provisioning.
How does tenant isolation work on bare metal Kubernetes with vCluster?
vCluster virtualizes the Kubernetes control plane. Each tenant receives a dedicated API server, data store, scheduler and RBAC running as a lightweight pod inside the control plane cluster, and tenants cannot see other tenants' nodes, pods or platform internals. For production, Private Nodes give each tenant dedicated worker nodes with its own CNI and CSI. vNode then helps prevent container breakout and limits blast radius at the workload level without hypervisor overhead.
Does vCluster support GPU workloads directly on bare metal?
Yes. The stack (vMetal for provisioning, vCluster Standalone as the CNCF-certified control plane binary, and vCluster for tenant cluster orchestration) powers 100K+ GPUs in production, and vCluster is validated in NVIDIA's DGX reference architecture. Auto Nodes adds GPU machines through the vMetal Machine API as tenant workloads are scheduled, and vNode keeps workloads isolated without putting a hypervisor between them and the GPU.
What determines readiness for managed Kubernetes on bare metal?
Boost Run went from decision to production launch in under 45 days for its managed Kubernetes service. That is a customer example rather than a promised implementation time. Your launch depends on machine inventory, driver configuration, networking, storage, templates and service integrations. vMetal and vCluster connect machine orchestration with tenant cluster delivery, reducing the infrastructure workflows you need to build yourself.
Is vCluster certified Kubernetes on bare metal?
Yes. Every tenant cluster created by vCluster runs a CNCF-certified Kubernetes control plane with standard Kubernetes APIs. Tenants talk to a real API server, install their own CRDs and have cluster-admin access inside their own tenant cluster. That holds whether control planes run as lightweight pods or as dedicated VMs for stricter OS-level separation.
Can bare metal Kubernetes with vCluster run in air-gapped or compliance environments?
Yes. vCluster supports air-gapped and FIPS deployments for environments with strict compliance or network isolation requirements. With vMetal provisioning on-premises machines and vNode isolating workloads, the stack runs in disconnected environments without depending on public cloud APIs or public registries. That suits regulated enterprises and GPU cloud providers with data sovereignty requirements.
Launch Bare Metal Kubernetes on Your GPUs
See how vCluster powers GPU clouds from rack to isolated tenant clusters.
