ai-cloud
Bare Metal GPU Platform for AI Cloud Providers
Make bare metal GPUs available as tenant infrastructure. vMetal automates machine provisioning, while vCluster gives each tenant separate control planes and dedicated Private Nodes.
Trusted by the fastest-growing AI cloud providers
Problem
The GPU Infrastructure Gap
Raw hardware alone cannot power the AI cloud your customers expect.
Race to the Bottom
Selling bare metal GPUs alone commoditizes your offering. Customers want a cloud experience they can self-serve.
DIY Takes Too Long
Building a GPU cloud platform in-house requires significant engineering investment.
Isolation Without Overhead
Standard Kubernetes forces you to choose between tenant isolation and operational efficiency on bare metal GPU infrastructure.
Solution
One Stack from Bare Metal to Tenant Clusters
vMetal turns registered bare metal inventory into provisioned machines. vCluster assigns dedicated Private Nodes to tenant clusters with independent control planes, and vNode can harden supported workload runtimes. vCluster Labs software powers 100K+ GPUs.
Built for Bare Metal GPU at Scale
Every layer of the stack is purpose-built for GPU cloud providers deploying bare metal infrastructure at production scale.
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 bare metal GPU workloads. Network automation uses supported integrations such as Netris.
- Stable Machine API
- Bare metal and VM provisioning
- Driver-specific lifecycle automation
Control Plane Binary
Kubernetes Directly on Bare Metal
vCluster Standalone runs Kubernetes as a CNCF-certified binary on Linux, without an existing Kubernetes cluster. It supplies a Kubernetes foundation for bare metal GPU workloads; 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 GPU workloads; 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 GPU workloads; inventory, driver readiness and release rules determine actual capacity.
- Karpenter-based node management
- Supported infrastructure provider integration
- Inventory-aware capacity policies
Workload Security
Workload Isolation Without VM Tax
vNode uses Linux user namespaces and seccomp to restrict workload privileges and system calls. It adds runtime hardening for bare metal GPU workloads 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 is bare metal GPU provisioning and why does it matter for AI clouds?
Bare metal GPU provisioning prepares physical GPU servers for workloads through OS installation, networking and driver configuration. vMetal orchestrates supported provisioning drivers after inventory and infrastructure setup. vCluster then provides tenant environments with dedicated Private Nodes. Running directly on servers removes a hypervisor layer, while actual training and inference performance still depends on the complete workload and infrastructure configuration.
How does vCluster deliver tenant isolation on bare metal GPU hardware?
vCluster runs each tenant's Kubernetes control plane as a lightweight process inside the control plane cluster, so every tenant has its own API server, data store, scheduler, and RBAC without a separate physical server for the control plane. Private Nodes give each tenant dedicated worker nodes with per-tenant CNI and storage. With vMetal handling the bare metal lifecycle and vNode adding workload isolation, each additional tenant costs little in control plane overhead.
Does vCluster require an existing Kubernetes cluster to run on bare metal?
No. vCluster Standalone runs as a single binary directly on bare metal Linux. There is no dependency on k3s, kubeadm, or any external Kubernetes distribution as a base layer. vMetal handles the bare metal provisioning lifecycle (PXE boot, OS installation, network configuration) and vCluster Standalone provides the CNCF-certified control plane binary that runs directly on that hardware.
How quickly can a GPU cloud provider launch a managed Kubernetes service?
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 used in production GPU environments at scale?
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 network isolation capabilities are available for bare metal GPU deployments?
vMetal integrates with Netris for hardware-enforced network isolation per tenant: VLANs, VXLANs, VRFs, ACLs, and DPU policies give each tenant its own network boundary in hardware. That matters when untrusted or competing workloads run on the same bare metal fabric, where namespace-level network policies alone fall short.
Launch Your Bare Metal GPU Cloud Faster
See how vCluster powers production AI clouds on bare metal GPU infrastructure.
