All providers
Google Cloud
Hyperscalerspot pricing$0.085/GB egress
Strong networking and TPU alternatives. Committed-use discounts reward predictable workloads.
Instance offers
48
43 available right now
Regions
4
Asia Pacific, EU West, US East, US West
Reliability
94/100
Uptime, support and SLA maturity
Cheapest anywhere
0/6
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 | $0.916/hr | $0.366/hr | +229% | Available |
| L40S 48GB | US East | 16 / 96 GB | $1.785/hr | $0.714/hr | +205% | Limited |
| A100 PCIe 40GB | US East | 12 / 117 GB | $2.483/hr | $0.993/hr | +266% | Available |
| A100 SXM 80GB | US East | 12 / 145 GB | $2.853/hr | $1.141/hr | +239% | Available |
| H100 SXM 80GB | US East | 26 / 200 GB | $3.906/hr | $1.562/hr | +144% | Available |
| H200 SXM 141GB | US East | 28 / 226 GB | $4.979/hr | $1.992/hr | +127% | Available |
Single-GPU instances only, cheapest region per model. Multi-GPU node shapes are in the full marketplace.