virtual machines
Shared Infrastructure, Dedicated Control
Governed, hardware-isolated GPU instances on your dedicated compute fabric. Partition one cluster across teams and tenants - with full control of isolation, policy, residency, and visibility.
You hold the controls
Visibility
Every tenant, workload, and quota across the environment
Identity
Instance access tied to approved users, roles, and teams, policy enforced automatically
Networking
Private VPCs and custom security groups on dedicated GPU compute
Instances isolated with no shared-host software - each tenant has direct access to the GPU compute fabric with separate network and security domains.
Right-sized capacity
Fractional GPUs
Consume only the GPU capacity a workload needs
Dynamic node pools
Scale with demand
Suspend and resume
Pause and restart GPU resources to remove idle spend
Programmatic lifecycle
Manage the full instance lifecycle by API, CLI, Terraform, and SDK. Deploy from approved instance profiles preconfigured with drivers, frameworks, storage paths, network policies, and access controls.
