platform-eng
AI Cloud Kubernetes for Managed GPU Offerings
Deliver managed Kubernetes as an AI cloud service. vMetal provisions GPU servers, and vCluster provides independent tenant environments with dedicated worker capacity through Private Nodes.
Trusted by the fastest-growing AI cloud providers
Problem
The AI Cloud Kubernetes Gap
GPU providers building managed Kubernetes on their own hardware run into three problems.
DIY Takes Too Long
Building tenant provisioning, isolation, upgrades and a portal in-house is a large project, and most teams underestimate how many pieces it involves.
Isolation or Efficiency
Plain Kubernetes makes you pick: namespaces are too weak for paying customers, and a separate cluster per customer is too expensive.
Customers Expect the Cloud
AI teams expect self-service Kubernetes and familiar tooling. Manual onboarding and unclear tenant boundaries make the infrastructure harder to use.
Solution
The Full Stack for AI Cloud Kubernetes
vMetal provisions GPU infrastructure, and vCluster orchestrates independent Kubernetes tenant clusters with dedicated Private Nodes. Platform templates and access policies provide the foundation for your customer-facing service. Lintasarta runs 170+ tenant clusters in production.
Built for AI Cloud Kubernetes at Scale
What an AI cloud needs for managed Kubernetes, from server provisioning to tenant isolation to ready-to-use AI environments.
Tenant Clusters
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 AI cloud onboarding; 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
Bare Metal
Zero-Touch Bare Metal GPU Provisioning
vMetal exposes one Machine API above bare metal and VM provisioning drivers. Registered inventory, images and credentials enable repeatable machine lifecycle workflows for AI cloud onboarding. Network automation uses supported integrations such as Netris.
- Stable Machine API
- Bare metal and VM provisioning
- Driver-specific lifecycle automation
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 AI cloud onboarding; inventory, driver readiness and release rules determine actual capacity.
- Karpenter-based node management
- Supported infrastructure provider integration
- Inventory-aware capacity policies
AI Workload Integration
Supported AI Workload Templates
Use supported workload templates for AI cloud onboarding, including the NVIDIA Run:AI partner integration. Configure GPU, network and storage prerequisites, and validate the selected template for Private Nodes before offering it to production customers.
- Native NVIDIA Run:AI catalog
- Validate Private Node configuration
- Release-specific integration prerequisites
Tenant Experience
Template Based Experience for Your Customers
Give customers a self-service portal to create and manage their own tenant clusters. They get the cloud experience they expect, and you skip building a custom UI.
- Approved templates and access policies
- Dedicated production tenant workers
- Readiness depends on available capacity
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 standard Kubernetes distributions for AI cloud providers?
A distribution gives you one cluster. To serve many customers you either build a cluster per customer, which is slow and expensive, or split one cluster with namespaces, which is too weak for paying tenants. vCluster runs a virtualized control plane per customer as a pod and puts each customer's workloads on Private Nodes, so adding a customer costs little and keeps the boundary hard.
How long does it take to launch a managed Kubernetes service 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.
Does vCluster work directly on bare metal without an existing Kubernetes cluster?
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.
What level of tenant isolation does vCluster provide for GPU workloads?
Each customer has its own control plane. Private Nodes are the recommended production model, giving each customer dedicated worker nodes with its own CNI and CSI. For your own dev, test and CI work, internal teams can use Shared Nodes. vNode can add a workload boundary using Linux user namespaces and seccomp, without a hypervisor.
Is vCluster's Kubernetes cluster certified and compatible with standard tooling?
vCluster is a CNCF-certified Kubernetes distribution and exposes standard Kubernetes APIs in tenant environments. This supports familiar clients, Helm charts and cluster-scoped resources such as CRDs. Certification covers Kubernetes conformance rather than an entire AI platform or customer deployment. Validate the selected version, networking, storage, GPU drivers and runtime integrations for the workloads you plan to support.
What AI platform integrations are available for AI cloud Kubernetes deployments?
NVIDIA Run:AI is available through the native Certified Stacks integration, with release-specific prerequisites and templates. Its bundled templates do not automatically supply dedicated Private Nodes. Configure and validate the production worker model, GPU access, networking and storage before offering it to customers. Ray and Jupyter can use compatible Kubernetes deployment tooling and require their own supported configuration.
Launch Your AI Cloud Kubernetes Platform
Explore managed Kubernetes on your GPU hardware with dedicated customer workers.
