Cloud providers we track
63 providers. “From” = lowest published on-demand price for the category, in US$.
All 63Hyperscaler 9GPU cloud 13GPU marketplace 4Developer cloud 7Regional cloud 9India cloud 11India MSP / data centre 10
| Provider | Type | HQ | GPU from | VPS from | Pricing |
|---|---|---|---|---|---|
| CoreWeave 'The Essential Cloud for AI' - Kubernetes-native GPU cloud for frontier labs | GPU cloud | United States | — | — | — |
| Crusoe Energy-first AI cloud and campus builder (Abilene 900 MW) | GPU cloud | United States | $1.50/GPU-hr | — | — |
| FluidStack Single-tenant AI supercomputer builder for frontier labs (Anthropic anchor) | GPU cloud | United States | — | — | — |
| Hyperstack | GPU cloud | United Kingdom | $1.00/GPU-hr | — | — |
| Jarvislabs | GPU cloud | India | $0.44/GPU-hr | — | — |
| Lambda | GPU cloud | United States | $2.79/GPU-hr | — | — |
| Nebius | GPU cloud | Netherlands | $1.55/GPU-hr | — | — |
| Neysa AI Acceleration Cloud System - fast training & inference on GPUs | GPU cloud | India | — | — | Publishes some prices |
| RunPod 'The AI Developer Cloud' - Pods, Serverless, Clusters, Hub; 1M+ developers claimed | GPU cloud | United States | $4.79/GPU-hr | — | — |
| TensorWave AMD-only (Instinct) AI cloud; 8,192 MI325X online | GPU cloud | United States | — | — | — |
| Together AI (GPU Clusters) AI-native cloud: inference APIs plus H100-GB300 GPU clusters | GPU cloud | United States | — | — | — |
| Verda (formerly DataCrunch) | GPU cloud | Finland | $1.41/GPU-hr | — | — |
| Voltage Park H100/B200/GB300 cloud, now merged with Lightning AI's dev platform | GPU cloud | United States | $1.99/GPU-hr | — | — |