H100 SXM vs H100 PCIe
H100 SXM has 1.7× the bandwidth.
The NVIDIA H100 SXM (Hopper, 2022) and the NVIDIA H100 PCIe (Hopper, 2022) share the same Hopper silicon generation, so this comes down to configuration and price rather than architecture. Memory is effectively a wash at 80 vs 80 GB, so capacity does not decide this one. On memory bandwidth, the number that governs LLM serving speed, the H100 SXM leads at 3.35 vs 2 TB/s. Raw compute favors the H100 SXM by 1.3x, which matters for training, prefill-heavy traffic, and diffusion work more than for chat-style decoding.
Price is where it settles: the H100 PCIe rents from $1.60/hr against $1.90/hr for the H100 SXM, a 1.2x gap that the performance numbers above do justify for the right workload. Per gigabyte of VRAM, the H100 PCIe is the cheaper rental, worth knowing if your model is capacity-bound rather than speed-bound. For how these numbers translate into tokens per dollar, the inference cost estimator runs both cards against any model. Everything below is the underlying data.
NVIDIA · Hopper · 2022
NVIDIA H100 SXM
The workhorse of the AI boom. Still the most widely available serious training and inference GPU.
From $1.90/hr at Hyperstack
NVIDIA · Hopper · 2022
NVIDIA H100 PCIe
Hopper in a standard PCIe slot: ~75% of SXM compute at half the power, often meaningfully cheaper.
From $1.60/hr at Hyperstack
Head to head
■ H100 SXM ■ H100 PCIe, bars share one scale across the whole catalog.
| H100 SXM | H100 PCIe | |
|---|---|---|
| Memory | 80 GB HBM3 | 80 GB HBM2e |
| Bandwidth | 3.35 TB/s | 2 TB/s |
| FP16 dense | 990 TF | 756 TF |
| FP8 dense | 1,979 TF | 1,513 TF |
| TDP | 700 W | 350 W |
| Interconnect | NVLink 4 · 900 GB/s | PCIe Gen5 · 128 GB/s (NVLink bridge optional) |
| Cheapest rental | $1.90/hr | $1.60/hr |
| $/hr per GB VRAM | $2.4¢ | $2.0¢ |
Rule of thumb: for LLM serving, prefer the GPU with more memory bandwidth per dollar; for training and prefill-heavy work, prefer FLOPS per dollar; and if the model doesn't fit in VRAM, none of the other numbers matter. Sanity-check with the VRAM calculator.