How It Works
Getting started with Runcrate Private Cloud is straightforward. Here’s what to expect from first contact to a running cluster.Process
Tell us what you need
Reach out to support@runcrate.ai with your requirements:
- GPU type and quantity
- Preferred region or location
- Contract duration (12 or 24 months)
- Any specific networking, storage, or compliance needs
Receive a proposal
We’ll put together a configuration and pricing proposal based on your requirements. This includes:
- Cluster specifications (nodes, GPUs, networking, storage)
- Monthly pricing with no hidden fees
- Available regions and delivery timeline
- Contract terms
Sign the agreement
Once you’re happy with the proposal, we finalize the service agreement. Simple terms, no surprises.
Cluster provisioning
Your dedicated cluster is provisioned and configured. This typically takes 1–2 weeks depending on GPU availability and location.
Contract Terms
| Term | Details |
|---|---|
| Duration | 12 or 24 months |
| Billing | Fixed monthly rate, invoiced monthly |
| Pricing | Per-GPU-hour rate locked for the contract duration |
| Scaling | Add nodes mid-contract (subject to availability) |
| Support | Dedicated support with SLA |
What You Manage vs. What We Manage
| You | Runcrate | |
|---|---|---|
| Software stack | Your choice — Kubernetes, Slurm, Docker, bare metal | Assistance available |
| Workload scheduling | Full control | — |
| Hardware | — | Provisioning, maintenance, replacements |
| Network | Application-level | Infrastructure, InfiniBand fabric |
| Monitoring | Application metrics | Hardware health, uptime |
| Support | — | 24/7 infrastructure support |
Regions
Private Cloud clusters are available in multiple regions globally. Current availability includes data centers in:- North America — US West, US East, Canada
- Europe — Germany, Netherlands, UK
- Asia Pacific — South Korea, Japan, Singapore, India, Vietnam
Region availability depends on GPU type and cluster size. Contact sales for current availability in your preferred location.
Scaling
Already have a cluster and need more capacity?- Add nodes to your existing cluster (subject to availability)
- Deploy additional clusters in the same or different regions
- Upgrade GPU type when transitioning between contract periods