The standard data center GPU for large-scale AI training and inference.
We're tracking 52 cloud providers for the H100, and the pricing spread is significant. While the average sits at $4.05/hr, the lowest price is currently $1.73/hr per GPU (on-demand). Spot instances start lower, at $0.30/hr per GPU.
USA 14 configs 1x-8x
USA 20 configs 1x-8x
USA 3 configs 1x-4x
USA 1 config 1x
USA 5 configs 1x-8x
USA 20 configs 1x-8x
USA 16 configs 1x-8x
USA 6 configs 1x-8x
France 4 configs 1x-8x
France 7 configs 1x-8x
USA 24 configs 1x-8x
Finland 8 configs 1x-8x
Germany 9 configs 1x-8x
France 5 configs 1x-8x
Finland 35 configs 1x-8x
Heads up: We do our best to keep these specs & prices accurate. However, cloud costs may fluctuate based on region, usage, and other factors not listed here. These are estimates based on common setups and are for informational purposes only. Always verify current rates & exact specs with the provider before provisioning.
80GB HBM3 with 900 GB/s NVLink for multi-node training. FP8 Transformer Engine for improved training throughput over A100. The standard choice for large-scale AI training and high-throughput inference.
May be more than needed for single-GPU inference on models under 30B parameters. Consider the L40S or A10 for cost-effective inference workloads.
The median on-demand price across providers has risen about 12% since August 2025, from $3.02 to $3.38/hr per GPU. This reflects the market-wide median, which moves both when providers change prices and when lower or higher priced offerings enter the market.
With 80GB of VRAM, the H100 is well suited to 30B-class models in FP16, and 70B-class models in 4-bit or 8-bit quantized form.
The H100 has 80GB of VRAM. Multi-GPU setups increase total memory, but that memory is not automatically pooled across GPUs.
The H100 has 3,350 GB/s of memory bandwidth. Higher bandwidth helps with faster data transfer between GPU memory and compute cores.
The H100 supports 7 precision formats. Training: BF16, FP16, TF32, FP32. Inference: FP8, INT8. Scientific: FP64.
Yes. The H100 supports NVLink with 900 GB/s of bidirectional bandwidth. This helps accelerate multi-GPU communication.
H100 pricing currently ranges from $0.30/hr to $14.90/hr per GPU, depending on the provider, instance type, and billing model.
At 720 hours per month, one H100 can cost between $215.50 to $10,730.24 per month, depending on the provider. Reserved and spot pricing can lower that further.
The H100 is available from 52 cloud providers, including Verda, Runpod, Spheron. Pricing and availability vary by region and billing model.
Yes. We currently track 331 H100 listings across 52 cloud providers:
| Billing type | Listings | Avg $/GPU/hr |
|---|---|---|
| On-demand | 151 | $3.87/hr |
| Reserved | 121 | $3.42/hr |
| Spot | 53 | $1.75/hr |
| Custom contract | 6 | $1.89/hr |
| H100 SXM | H100 PCIe | H100 NVL | |
|---|---|---|---|
| FP64 | 34 TFLOPS | 26 TFLOPS | 68 TFLOPS |
| FP64 Tensor Core | 67 TFLOPS | 51 TFLOPS | 134 TFLOPS |
| FP32 | 67 TFLOPS | 51 TFLOPS | 134 TFLOPS |
| TF32 Tensor Core | 989 TFLOPS | 756 TFLOPS | 1,979 TFLOPS |
| BFLOAT16 Tensor Core | 1,979 TFLOPS | 1,513 TFLOPS | 3,958 TFLOPS |
| FP16 Tensor Core | 1,979 TFLOPS | 1,513 TFLOPS | 3,958 TFLOPS |
| FP8 Tensor Core | 3,958 TFLOPS | 3,026 TFLOPS | 7,916 TFLOPS |
| INT8 Tensor Core | 3,958 TOPS | 3,026 TOPS | 7,916 TOPS |
| GPU Memory | 80GB | 80GB | 188GB |
| GPU Memory Bandwidth | 3.35TB/s | 2TB/s | 7.8TB/s |
| Decoders | 7 NVDEC, 7 JPEG | 7 NVDEC, 7 JPEG | 14 NVDEC, 14 JPEG |
| Max Thermal Design Power (TDP) | Up to 700W | 300-350W | 2x 350-400W |
| Multi-Instance GPUs | Up to 7 MIGs @ 10GB each | Up to 7 MIGs @ 10GB each | Up to 14 MIGs @ 12GB each |
| Form Factor | SXM | PCIe, dual-slot, air-cooled | 2x PCIe, dual-slot, air-cooled |
Source: official Nvidia H100 datasheet.
Next generation with 76% more memory. From $0.93/hr per GPU across 36 providers.
Previous generation, more affordable. From $0.20/hr per GPU across 41 providers.
AMD alternative with 2.4x more memory. From $1.45/hr per GPU across 10 providers.
Consumer gaming GPU. From $0.11/hr per GPU across 14 providers.
Previous generation Volta GPU. From $0.08/hr per GPU across 21 providers.
Last updated
