NVIDIA · Hopper · 2022
NVIDIA H100 PCIe
Hopper in a standard PCIe slot: ~75% of SXM compute at half the power, often meaningfully cheaper.
80 GB
HBM2e
2 TB/s
Mem bandwidth
756 TF
FP16 dense
1,513 TF
FP8 dense
350 W
TDP
$1.60 /hr
From · Hyperstack
Lower clocks and HBM2e instead of HBM3 mean about 60% of the SXM's memory bandwidth. Fine for single-GPU work; the weaker interconnect hurts multi-GPU training.
Decision guide: A100 vs H100 for inference →
Where to rent a H100 PCIe
On-demand prices per GPU-hour; marketplace rates refresh daily, list prices reviewed 2026-09. Reserved and spot run 30-70% lower. Full breakdown with the used market and rent-vs-buy math: H100 PCIe pricing.
| Provider | $/GPU-hr | Notes |
|---|---|---|
| Hyperstack | $1.60 | neocloud |
| FluidStack | $1.85 | neocloud |
| Vast.ai | $2.56 | Verified-listing rate, refreshed daily |
| RunPod | $2.89 | neocloud |
| OVHcloud | $2.99 | hyperscaler |
Specifications in context
Bars scaled against the best value in the whole catalog (B200 / MI325X era).
| Architecture | Hopper (2022) |
|---|---|
| Memory | 80 GB HBM2e, 2 TB/s |
| Interconnect | PCIe Gen5 · 128 GB/s (NVLink bridge optional) |
| Form factor | PCIe |
| Partitioning | MIG, up to 7 isolated instances |
What fits on one H100 PCIe
Weights + ~2 GB runtime overhead against 72 GB usable VRAM. Longer contexts and bigger batches need more, check the VRAM calculator.
| Model | Params | Highest precision that fits |
|---|---|---|
| Llama 3.2 1B | 1.24B | FP16 |
| Llama 3.2 3B | 3.21B | FP16 |
| Llama 3.1 8B | 8.03B | FP16 |
| Llama 3.3 70B | 70.6B | 4-bit |
| Qwen2.5 7B | 7.62B | FP16 |
| Qwen2.5 14B | 14.8B | FP16 |
| Qwen2.5 32B | 32.8B | 8-bit |
| Qwen2.5 Coder 32B | 32.8B | 8-bit |
| Qwen2.5 72B | 72.7B | 4-bit |
| QwQ 32B (reasoning) | 32.8B | 8-bit |
| Mistral 7B | 7.25B | FP16 |
| Mixtral 8x7B | 46.7B | 8-bit |
| Gemma 2 9B | 9.24B | FP16 |
| Gemma 2 27B | 27.2B | FP16 |
| Phi-4 14B | 14.7B | FP16 |
| gpt-oss-20b | 20.9B | FP16 |
Best for
- Single-GPU fine-tuning and inference
- Cost-sensitive H100-class workloads
- Standard server chassis without SXM boards