ai-cloud
Provisioning Orchestration for AI Clouds
Coordinate provisioning across bare metal and VM infrastructure. vMetal provides a stable Machine API, and vCluster consumes that capacity as dedicated tenant Private Nodes.
Trusted by the fastest-growing AI cloud providers
Problem
Why Manual Provisioning Breaks at Scale
AI cloud infrastructure becomes difficult to automate when every provisioning stack exposes a different API and lifecycle.
Slow, Fragmented Provisioning Stacks
Disconnected provisioning systems create separate workflows for bare metal, VMs, networking, and cluster attachment.
No Stable API Across Drivers
Direct dependencies on driver-specific APIs make infrastructure changes ripple into customer-facing automation.
Manual Lifecycle Work Adds Handoffs
Manual Machine lifecycle work delays provisioning, release, cleaning, and capacity reuse.
Solution
One API to Orchestrate Every Machine Layer
vMetal exposes a stable Machine API above infrastructure-specific provisioning drivers. vCluster and Auto Nodes consume supported capacity as Private Nodes, connecting infrastructure lifecycle operations to tenant cluster requirements. Lintasarta runs 170+ tenant clusters in production.
Built for GPU Rack to Production Orchestration
vMetal unifies physical and virtual Machine provisioning behind one API, while vCluster turns that capacity into tenant cluster products.
Bare Metal Orchestration
Automated Machine Provisioning
vMetal exposes one Machine API above bare metal and VM provisioning drivers. Registered inventory, images and credentials enable repeatable machine lifecycle workflows for provisioning coordination. Network automation uses supported integrations such as Netris.
- Stable Machine API
- Bare metal and VM provisioning
- Driver-specific lifecycle automation
Dynamic Scaling
Capacity Requests Through Auto Nodes
Karpenter-based Auto Nodes requests worker capacity according to configured policies. With the vMetal provider, Machine requests provision available infrastructure for provisioning coordination; inventory, driver readiness and release rules determine actual capacity.
- Karpenter-based node management
- Supported infrastructure provider integration
- Inventory-aware capacity policies
Network Automation
Network Configuration by Driver
Tenant-owned CNI configuration and supported Netris integrations provide separate network controls for provisioning coordination. Ethernet segmentation can use VLANs, VRFs and ACLs; InfiniBand isolation uses fabric partitions rather than Ethernet constructs.
- Tenant-owned worker network configuration
- Netris integration for supported fabrics
- Explicit segmentation and access policies
Tenant Cluster Delivery
Independent Tenant Kubernetes Control Planes
Each tenant gets its own Kubernetes API, backing store and RBAC scope as a lightweight control-plane workload. vMetal supplies machines, while Private Nodes attach dedicated workers to the tenant environment.
- Independent tenant API and data store
- Tenant-scoped scheduling on Private Nodes
- No dedicated control-plane servers per tenant
Tenant Isolation
Private Nodes for Production Tenants
Private Nodes dedicate worker capacity to one tenant cluster, with its own CNI and storage configuration. For provisioning coordination, select physical servers or VMs according to the required infrastructure boundary.
- Dedicated worker nodes per tenant
- Tenant-owned CNI and storage configuration
- Physical servers or supported VMs
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 vMetal a provisioning orchestrator rather than a provisioning tool?
vMetal is the Infrastructure Orchestrator. It coordinates the machine layer, including provisioning, compute, network, storage and virtualization, through one stable Machine API. Existing infrastructure drivers perform the underlying provisioning work. Operators configure inventory, credentials and images, then vCluster uses the resulting machines for tenant environments with dedicated Private Nodes.
What does machine provisioning orchestration cover end to end?
vMetal requires vCluster Platform and a connected Kubernetes cluster for its Metal3/Ironic provisioning components. If you are starting without Kubernetes, vCluster Standalone can bootstrap that foundation as a binary on Linux. Once infrastructure drivers, hardware inventory, images and credentials are configured, vMetal can automate supported provisioning workflows and supply machines for tenant Private Nodes.
How does vMetal integrate with Kubernetes cluster orchestration?
vMetal is the Infrastructure Orchestrator for the machine layer; vCluster is the Cluster Orchestrator for tenant environments. A tenant cluster is a compute environment for one tenant, with Kubernetes as one cluster type among several. For Kubernetes deployments, vCluster provides independent tenant control planes and uses provisioned machines as dedicated Private Nodes. Product availability and deployment prerequisites depend on the selected cluster type.
Can vMetal work with our existing provisioning infrastructure?
vMetal uses a driver-based architecture, so integration depends on your provisioning system and its supported driver. Identify the required machine lifecycle operations, images, credentials and networking before connecting capacity. The stable Machine API gives higher layers a consistent request model; it does not imply every infrastructure backend supports identical provisioning, cleaning or cluster attachment workflows.
How does Auto Nodes relate to machine provisioning orchestration?
Auto Nodes uses Karpenter-based node management, with vCluster Platform handling infrastructure provisioning through a configured node provider. Providers include vMetal and other supported integrations. On the vMetal path, Machine requests allocate and provision available inventory as tenant Private Nodes. Scaling depends on configured policies and usable capacity; it does not create physical inventory or automatically diagnose hardware failures.
What evidence supports vCluster Labs software at GPU 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.
Turn Your GPU Racks Into a Cloud Product
See how vMetal orchestrates your full machine provisioning lifecycle end to end.
