🇺🇸 GPU Cloud in United States
37 cloud providers offer GPU instances in United States. 1894 live instances, prices from $0.007/hr.
Last updated April 25, 2026 · Data refreshed every 6 hours
Providers
37
GPU Instances
1894
Cheapest
$0.007/hr
All Providers in United States
| Provider | GPU Models | Instances | From |
|---|---|---|---|
| Vast.ai | 13 | 78 | $0.007/hr |
| Vultr | 4 | 35 | $0.059/hr |
| Azure | 6 | 154 | $0.069/hr |
| RunPod | 15 | 282 | $0.110/hr |
| GCP | 16 | 679 | $0.124/hr |
| AWS | 8 | 223 | $0.138/hr |
| Salad Cloud | 6 | 6 | $0.200/hr |
| FluidStack | 4 | 6 | $0.290/hr |
| Beam | 4 | 4 | $0.310/hr |
| TensorDock | 6 | 8 | $0.340/hr |
| Shadeform | 23 | 143 | $0.350/hr |
| Crusoe | 6 | 9 | $0.400/hr |
| Lightning.ai | 6 | 8 | $0.410/hr |
| Novita | 4 | 5 | $0.440/hr |
| Jarvis Labs | 9 | 9 | $0.440/hr |
| Thunder Compute | 3 | 3 | $0.490/hr |
| Linode | 1 | 9 | $0.520/hr |
| Cerebrium | 4 | 4 | $0.522/hr |
| Paperspace | 4 | 4 | $0.560/hr |
| Modal | 6 | 8 | $0.590/hr |
| OCI | 6 | 85 | $0.637/hr |
| Replicate | 4 | 4 | $0.810/hr |
| Contabo | 1 | 1 | $0.890/hr |
| Lambda Labs | 6 | 75 | $1.29/hr |
| Hetzner | 2 | 2 | $1.35/hr |
| Nebius | 2 | 8 | $1.45/hr |
| Cirrascale | 4 | 5 | $1.50/hr |
| Fly.io | 3 | 3 | $1.50/hr |
| Latitude.sh | 3 | 3 | $1.68/hr |
| Together AI | 2 | 4 | $1.80/hr |
| Oblivus | 3 | 5 | $1.89/hr |
| Voltage Park | 1 | 1 | $1.99/hr |
| TensorWave | 2 | 4 | $2.49/hr |
| Gcore | 2 | 2 | $2.49/hr |
| Fal.ai | 3 | 4 | $2.70/hr |
| DigitalOcean | 1 | 2 | $4.46/hr |
| CoreWeave | 9 | 9 | $6.50/hr |
Why Choose a GPU Cloud in United States?
If you're building AI workloads serving users in United States or nearby regions, picking a GPU cloud provider with local datacenters reduces latency and may help with data residency requirements (GDPR for EU countries, sovereign cloud requirements for government workloads).
Common GPU types available locally: H100, A100, L40S, RTX 4090, T4. Prices vary 5-10x between hyperscalers (AWS, GCP, Azure) and specialized providers.
Explore More