platform-eng
Neocloud Kubernetes for Managed GPU Offerings
Build your AI cloud Kubernetes service on bare metal. vCluster deploys fully isolated tenant clusters in seconds, without full cluster overhead or months of engineering.
Trusted by the fastest-growing AI cloud providers
Problem
The Neocloud Kubernetes Gap
GPU cloud providers face three hard problems building managed Kubernetes on bare metal.
DIY Takes Years
Based on typical platform engineering experience, building a GPU cloud platform yourself takes 6 to 10 engineers, 6 to 12 months, and over $1M. Most teams are still building two years in.
Isolation or Efficiency
Standard Kubernetes forces you to choose between tenant isolation and operational efficiency. Namespace isolation is too weak. Separate clusters are too expensive.
Customers Expect the Cloud
AI teams have used AWS and GCP. They expect self-service environments, managed Kubernetes, and cloud-native tooling — and they will go back to a hyperscaler if you cannot deliver it.
Solution
The Full Stack for Neocloud Kubernetes
vCluster delivers the complete path from bare metal GPU racks to managed Kubernetes. Fully isolated, CNCF-certified tenant clusters spin up in seconds as lightweight processes. Boost Run launched their managed Kubernetes offering in less than 45 days. Lintasarta launched Indonesia's leading GPU cloud in 90 days.
Built for AI Cloud Kubernetes at Scale
Every layer your neocloud Kubernetes platform needs, from bare metal provisioning to tenant isolation to AI-ready environments.
Tenant Clusters
Isolated Tenant Clusters in Seconds
Each tenant gets a fully isolated CNCF-certified Kubernetes cluster with its own API server, etcd, and RBAC — running as a lightweight pod. No separate physical clusters, no minutes-long provisioning delays.
- Own API server per tenant
- Spins up in seconds
- No physical cluster overhead
Bare Metal
Zero-Touch Bare Metal GPU Provisioning
vMetal handles PXE boot, OS installation, machine registration, and network automation for GPU servers. Go from rack to production-ready Kubernetes node with zero manual intervention.
- PXE boot and OS install
- Full GPU server lifecycle management
- Network automation built in
Dynamic Scaling
Auto-Provision GPU Nodes on Demand
Automatically provisions bare metal GPU nodes via Terraform when tenants schedule workloads. Dynamic scaling for physical servers without manual intervention, maximizing GPU utilization across your fleet.
- Bare metal Karpenter for GPU
- Terraform-driven node provisioning
- Maximizes GPU fleet utilization
AI Environments
Pre-Validated AI Platform Environments
Turn a bare Kubernetes cluster into a production AI platform in minutes with pre-validated integrations for Run:AI, Ray, and Jupyter. Tested and certified against vCluster tenant isolation.
- Run:AI, Ray, Jupyter certified
- Cluster to AI platform in minutes
- Works with tenant isolation
Tenant Experience
EKS-Like Experience for Your Customers
Give tenants a self-service portal to provision and manage their own Kubernetes environments. Deliver the cloud-native experience AI teams expect without building a custom control plane UI.
- EKS-like self-service portal
- Tenant-controlled environments
- No custom UI engineering
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 makes vCluster different from standard Kubernetes distributions for neocloud providers?
vCluster virtualizes the Kubernetes control plane itself — each tenant gets a real API server, etcd, and RBAC running as a lightweight process inside a host cluster. Standard distributions require provisioning separate physical clusters per tenant, which is expensive and slow, or rely on namespace-level isolation, which is too weak for GPU cloud customers. vCluster gives you strong tenant isolation with near-zero marginal cost per tenant, proven at 100K+ GPU nodes in production.
How long does it take to launch a managed Kubernetes service on bare metal?
Boost Run launched their managed Kubernetes offering in less than 45 days using vCluster. Lintasarta launched Indonesia's leading GPU cloud in 90 days with 170+ tenant clusters. The full stack covers bare metal provisioning through tenant cluster orchestration to AI-ready environments, so you are not assembling disparate tools or writing custom orchestration code from scratch.
Does vCluster work directly on bare metal without an existing Kubernetes cluster?
Yes. vCluster Standalone runs as a single binary directly on bare metal Linux — no k3s, kubeadm, or k0s required as a base layer. The vMetal component handles zero-touch provisioning from PXE boot through OS installation and network configuration. This gives you a complete path from physical GPU servers to tenant Kubernetes clusters without intermediate dependencies.
What level of tenant isolation does vCluster provide for GPU workloads?
vCluster supports a flexible isolation spectrum. Shared Nodes give cost-efficient density with namespace boundaries. Private Nodes give each tenant fully dedicated physical hardware with their own CNI and CSI. Dedicated Nodes eliminate noisy-neighbor GPU contention with guaranteed performance. vNode (currently in private beta) adds kernel-native workload isolation using seccomp, cgroups, namespaces, and AppArmor without hypervisor overhead, preserving bare metal GPU performance.
Is vCluster's Kubernetes distribution certified and compatible with standard tooling?
Yes. Every tenant cluster created by vCluster is a CNCF-certified Kubernetes distribution with 100% API compatibility. Tenants get full cluster-admin access, can install their own CRDs, configure RBAC, and use any standard Kubernetes tooling. vCluster is also named in the NVIDIA DGX SuperPOD reference architecture and in SemiAnalysis ClusterMax evaluation criteria.
What AI platform integrations are available for neocloud Kubernetes deployments?
vCluster's Certified Stacks include pre-validated integrations for Run:AI, Ray, and Jupyter, turning a bare Kubernetes cluster into a production AI platform in minutes. Slurm-on-Kubernetes support is available via the Slinky integration. All integrations are tested and certified to work with vCluster tenant isolation, so AI platforms run in isolated environments without custom configuration work.
Launch Your Neocloud Kubernetes Platform
See how GPU cloud providers deploy managed Kubernetes on bare metal in weeks.
