Skip to content
InfratoApp
InfratoApp / Compare / H100-80GB vs H100-SXM-80GB

H100-80GB vs H100-SXM-80GB

H100-80GB is the cheaper of the two: $1.7945/hr at Lium, against $2.25/hr for H100-SXM-80GB at TensorDock — 1.3× the price.

Cheapest on each

H100-80GB
21 SELLERSView H100-80GB →
$1.79/hr
cheapest, at Lium
H100-SXM-80GB
13 SELLERSView H100-SXM-80GB →
$2.25/hr
cheapest, at TensorDock

Aligned facts

Same sheet, same order, both cards — including the gaps.

AttributeH100-80GBH100-SXM-80GB
VRAM80 GB80 GB
Bandwidth○ no answer3.35 TB/s
FP16○ no answer989 TFLOPS
FP8○ no answer1979 TFLOPS
TDP○ no answer700 W
NVLink○ no answerNVLink 4
ArchitectureHopperHopper
Released20222022

A blank is not a zero. Where one card states a figure and the other states none, neither wins the row — the second number is unknown, and ranking an unknown as nothing is how a comparison starts to lie.

Every seller, both cards

One row per seller, cheapest tier, on-demand list prices only. Sellers of both cards lead, cheapest combined rate first.

SellerH100-80GBH100-SXM-80GB
TensorDock
Hyperstack
Massed Compute
Runcrate
Oblivus
Lium○ no answer
HyperAI○ no answer
Sesterce○ no answer
Koyeb○ no answer
Spheron○ no answer
DeepInfra○ no answer
Civo○ no answer
Canopy Wave○ no answer
Cerebrium○ no answer
Beam○ no answer
Enverge○ no answer
DigitalOcean○ no answer
Fal.ai○ no answer
Lyceum○ no answer
Replicate○ no answer
Crusoe○ no answer
Vast.ai○ no answer
Thunder Compute○ no answer
Hyperbolic○ no answer
Modal○ no answer
RunPod○ no answer
Lambda○ no answer
Verda (DataCrunch)○ no answer
CoreWeave○ no answer

What this page will not tell you

On-demand list rates only · spot and committed terms priced separately and not mixed in here · no performance-per-dollar figure appears on this page, because it needs a throughput benchmark we do not hold and dividing a price by a vendor’s own TFLOPS claim produces a number that looks authoritative and means nothing