ai-cloud
Bare Metal GPU Provisioning for AI Clouds
Meet growing AI workload demand with zero-touch bare metal GPU provisioning. vMetal goes from rack to production Kubernetes without complex virtualization overhead.
Trusted by the fastest-growing AI cloud providers
Problem
GPU Provisioning Slows Revenue
AI cloud providers lose competitive ground when hardware sits idle and tenants wait.
Racks Sit Idle Too Long
Manual provisioning workflows delay time to first workload, costing revenue on every GPU hour your cluster is not serving tenants.
No Path from Metal to Kubernetes
Without an integrated stack, teams stitch together k3s, kubeadm, and custom scripts — and most are still building two years in.
Bare Metal Alone Is a Dead End
Selling bare metal GPUs alone is a race to the bottom. Customers want the cloud experience, not raw compute.
Solution
Zero Touch from GPU Rack to Production
vMetal handles PXE boot, OS installation, machine registration, and network automation so your team never touches a server manually. Combine with vCluster to deliver isolated tenant clusters on top — giving every customer a dedicated Kubernetes environment on your bare metal GPU infrastructure.
The Complete Bare Metal GPU Stack
From server registration to isolated tenant Kubernetes environments, every layer is covered in one integrated platform.
Bare Metal
Zero Touch GPU Server Provisioning
PXE boot, OS install, machine registration, and full GPU server lifecycle management in one platform. Go from powered-on rack to production-ready infrastructure without manual intervention.
- PXE boot and OS automation
- Full GPU lifecycle management
- Rack to production, zero touch
Kubernetes Layer
K8s Running Directly on Bare Metal
vCluster Standalone runs as a single binary on bare metal with no external Kubernetes dependency. No k3s, no kubeadm, no intermediate layers between your hardware and production workloads.
- Single binary on bare metal
- No k3s or kubeadm required
- No external K8s dependency
Dynamic Scaling
Automatic GPU Node Provisioning
Auto Nodes provisions bare metal GPU servers via Terraform the moment tenants schedule workloads. Your infrastructure scales to demand without manual intervention on every new request.
- Demand-driven node provisioning
- Terraform-based automation
- No manual scaling steps
Tenant Isolation
Isolated Tenant Clusters in Seconds
Each tenant gets a fully isolated Kubernetes control plane — own API server, etcd, and RBAC — running as a lightweight process on your shared GPU infrastructure. Spin up hundreds of environments without new hardware.
- Own API server per tenant
- Seconds to provision
- Shared hardware, strongly isolated tenant boundaries
AI Environments
Production AI Platforms in Minutes
Pre-validated stacks with Run:AI, Ray, and Jupyter turn a bare metal Kubernetes cluster into a production AI platform in minutes. Certified to work with tenant isolation out of the box.
- Run:AI, Ray, Jupyter included
- Cluster to AI platform in minutes
- Certified with tenant isolation
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
What is bare metal GPU provisioning and why does it matter for AI clouds?
Bare metal GPU provisioning is the process of configuring physical GPU servers — from PXE boot and OS installation to network setup and Kubernetes registration — without intermediate virtualization layers. For AI cloud providers, it matters because every hour a GPU rack sits unprovisioned is revenue lost. Zero-touch bare metal provisioning eliminates manual steps in the deployment path, letting teams go from powered-on hardware to production-ready GPU compute without custom scripts or operator intervention.
How does vMetal handle GPU server lifecycle management?
vMetal manages the complete GPU server lifecycle: PXE boot, OS installation, machine registration, network automation including VLANs and VXLANs, and decommissioning. It integrates with Netris for hardware-level network policy enforcement. The result is a zero-touch provisioning pipeline where new GPU servers enter your fleet automatically without manual configuration at each stage.
Do I need an existing Kubernetes cluster to use vMetal?
No. vCluster Standalone runs as a single binary directly on bare metal Linux with no external Kubernetes dependency. You do not need k3s, kubeadm, or k0s as a base layer. vMetal handles the hardware provisioning and vCluster Standalone provides the Kubernetes distribution on top — giving you a complete path from raw server to production Kubernetes without assembling separate components.
How quickly can an AI cloud provider go from bare metal to serving tenants?
Lintasarta launched Indonesia's leading GPU cloud in 90 days using vMetal and vCluster, with over 170 tenant environments deployed. Boost Run launched a managed Kubernetes offering in under 45 days with zero new platform engineering hires. These timelines include bare metal GPU provisioning, Kubernetes distribution, and tenant isolation — not just the hardware layer.
How does bare metal GPU provisioning connect to tenant isolation?
vMetal handles the hardware provisioning layer while vCluster delivers isolated tenant Kubernetes environments on top of the same bare metal infrastructure. Each tenant gets a dedicated control plane — their own API server, etcd, and RBAC — running as a lightweight process. This means you can provision GPU racks once and serve many tenants with full isolation, without purchasing separate physical clusters per customer.
Is vMetal compatible with NVIDIA GPU infrastructure?
Yes. vMetal and the broader vCluster platform are named in the NVIDIA DGX SuperPOD reference architecture. The platform supports bare metal GPU provisioning for DGX-class hardware and is production-proven across 100K+ GPU nodes in deployment with more than 50 GPU clouds and Fortune 500 customers including CoreWeave and Nscale.
Start Provisioning GPU Bare Metal Faster
See how vMetal goes from rack to production Kubernetes without the manual overhead.
