All providers
CoreWeave
GPU cloud$0.010/GB egress
GPU-native cloud with InfiniBand fabrics. Fast multi-node scheduling at well under hyperscaler rates.
Instance offers
42
28 available right now
Regions
3
EU West, US East, US West
Reliability
84/100
Uptime, support and SLA maturity
Cheapest anywhere
0/7
GPU models where this is the market low
Entry price per GPU model
| GPU | Region | vCPU / RAM | Price | Spot | vs market low | Status |
|---|---|---|---|---|---|---|
| A10G 24GB | US East | 8 / 32 GB | $0.734/hr | — | +164% | Available |
| RTX A6000 48GB | US East | 14 / 64 GB | $1.008/hr | — | +143% | Available |
| L40S 48GB | US East | 16 / 96 GB | $1.270/hr | — | +117% | Limited |
| A100 PCIe 40GB | US East | 12 / 117 GB | $1.787/hr | — | +164% | Available |
| A100 SXM 80GB | US East | 12 / 145 GB | $2.194/hr | — | +161% | Limited |
| H100 SXM 80GB | US East | 26 / 200 GB | $2.375/hr | — | +48% | Limited |
| H200 SXM 141GB | US East | 28 / 226 GB | $3.260/hr | — | +49% | Available |
Single-GPU instances only, cheapest region per model. Multi-GPU node shapes are in the full marketplace.