GPU EconomyCloud GPU prices, read every hour
Latest read 97 providers tracked13,250 prices read in 24 h
H100 index$3.5334 providersWidestV100 32GB20.3×Mover, 7dRTX A4000+8.5%

NVIDIA H100 SXM vs NVIDIA H100 PCIe

What each costs to rent by the hour across the providers we track — NVIDIA H100 SXM on 39 providers, NVIDIA H100 PCIe on 12.

As of , the NVIDIA H100 SXM has a median rate card of $3.52 per GPU-hour versus $2.89 for the NVIDIA H100 PCIe — $0.63 (18%) more.

Side by side

NVIDIA H100 SXMNVIDIA H100 PCIe
ArchitectureHopperHopper
VRAM80 GB80 GB
Form factorSXMPCIe
Median rate card, per GPU-hour$3.52$2.89
Cheapest listed rate$1.06
Google Cloud, spot, GPU only
captured
$1.75
Hyperstack, reserved, from
captured
Median per GB of VRAM, per hour$0.044$0.036
Providers listing it3912

Providers that list both

On-demand rate cards, per GPU-hour - community capacity and marketplace floors out, as in every median; each price links to the provider's pricing page with its read time.
ProviderNVIDIA H100 SXMNVIDIA H100 PCIeH100 PCIe vs H100 SXM
HexGrid$2.99
Secure Cloud
captured
$2.75
Secure Cloud
captured
−$0.24 (−8%)
Hyperstack$3.20
captured
$2.50
captured
−$0.70 (−22%)
Scaleway$3.54
fr-par-2
captured
$3.21
fr-par-2
captured
−$0.34 (−9%)
Barrack AI$3.59
captured
$2.89
captured
−$0.70 (−19%)
RunPod$3.99
Secure Cloud
captured
$2.89
Secure Cloud
captured
−$1.10 (−28%)
Lambda$3.99
captured
$3.29
captured
−$0.70 (−18%)
GPU.ai$3.99
from
captured
$1.98
from
captured
−$2.01 (−50%)

Medians are median rate cards — each provider once, at its cheapest on-demand rate, community capacity and marketplace floors out — the same figure as on each card's page; the cheapest listed rate may be spot or a community tier and is labelled as such. Compare providers directly on the overview, or see each GPU's full table: NVIDIA H100 SXM, NVIDIA H100 PCIe. On your own page: embed this comparison (/embed/gpus/h100-sxm-vs-h100-pcie, an iframe, no script).

Data updated: