Pay for Infrastructure, Not for Markup
Transparent per-hour rates, up to 70% below hyperscaler list prices — with zero egress fees and no hidden charges.
Published Rate Card
The same transparent rates for every customer. Metered per hour, billed monthly in arrears.
Cloud Compute
| Instance Family | Configuration | Per Hour | Per Month |
|---|---|---|---|
| General Purpose | 2 vCPU / 4 GB RAM | $0.0343/hr | ~$25 |
| General Purpose | 4 vCPU / 8 GB RAM | $0.0557/hr | ~$41 |
| Compute Optimised | 8 vCPU / 16 GB RAM | $0.0987/hr | ~$72 |
| Memory Optimised | 8 vCPU / 64 GB RAM | $0.2578/hr | ~$188 |
Larger and custom shapes available. Kubernetes clusters and managed apps are billed at these compute rates plus a management fee.
GPU Nodes
| GPU | Node Configuration | GravityStack | Typical Hyperscaler* | You Save |
|---|---|---|---|---|
| NVIDIA H100 (80GB) | 26 vCPU / 256 GB / 80 GB HBM3 | $2.85/hr | $12.29/hr | −77% |
| NVIDIA A100 (80GB) | 12 vCPU / 96 GB / 80 GB HBM2e | $1.65/hr | $4.10/hr | −60% |
| NVIDIA L40S (48GB) | 8 vCPU / 64 GB / 48 GB GDDR6 | $0.95/hr | $1.86/hr | −49% |
*Representative on-demand list prices for comparable single-GPU capacity at major US hyperscalers, July 2026. Multi-node clusters with InfiniBand available on request.
Storage & Data Transfer
| Service | Unit | Rate |
|---|---|---|
| Block Storage (NVMe SSD) | per GB / month | $0.03 |
| Object Storage (S3-compatible) | per GB / month | $0.008 |
| Snapshots & Backups | per GB / month | $0.02 |
| Data Egress | per GB | Free — always |
Commercial Models That Scale With You
Start on-demand. Commit when your usage is predictable. Negotiate enterprise terms when you're ready.
On-Demand
Pay for what you use
List rates
- Per-hour metered billing
- No commitment, cancel anytime
- All regions and services
- Scale up or down instantly
Reserved Capacity
1- or 3-year committed use
Save up to 45%
- Guaranteed capacity, priority placement
- Fixed pricing for budget certainty
- Applies to compute, GPU, and storage
- Flexible payment schedules
Enterprise Agreement
Custom terms at scale
Volume pricing
- Custom SLA up to 99.99%
- Invoicing, PO, and NET-30/60 terms
- Dedicated account manager & TAM
- Migration and onboarding assistance
Procurement & Billing FAQ
The answers your finance and platform teams will ask for.
How does billing work?
All services are metered per hour and billed monthly in arrears. You pay only for what you use — there are no minimum commitments on the on-demand tier, and no charges for stopped resources beyond storage.
Do you support purchase orders and invoicing?
Yes. Enterprise customers can pay by invoice with NET-30 or NET-60 terms against a purchase order. Credit card and bank transfer are available for all accounts.
Is data egress really free?
Yes — zero egress fees on all tiers, with no fair-use cap on standard workloads. Moving terabytes out of the platform costs nothing, which alone often cuts hyperscaler bills by 20-30%.
What SLAs do you offer?
Standard infrastructure carries a 99.95% availability SLA with service credits. Enterprise agreements can include custom SLAs up to 99.99%, defined RTO/RPO targets, and dedicated support response times.
Where is my data hosted?
Your infrastructure runs in the region you choose — US, UK, EU, Middle East, or India — and stays under that jurisdiction. A sovereign alternative to the hyperscalers: your data, prompts, and workloads never leave infrastructure you control.
Can you help us migrate from AWS, Azure, or VMware?
Yes. Our repatriation tooling supports one-click migration from vSphere and guided migration from AWS, GCP, and Azure — with engineering assistance included on Reserved and Enterprise tiers.
Get a Custom Quote for Your Workload
Tell us your current cloud spend and we'll model the savings — most enterprises cut infrastructure costs by 50-70% including egress.