Search/H100

H100

80GB 700W 8 providers 570 offers

NVIDIA's Hopper flagship with 80 GB HBM3 and ~3.35 TB/s of bandwidth — the workhorse for large-scale LLM training and high-throughput inference.

Cheapest
$1.138/hr
Verda
Median
$18.753/hr
across 570 offers
P90
$80.000/hr
90th percentile
Prices updated 1 hour ago8 providers trackedHow we collect prices
Memory
80 GB
TDP
700 W
Bandwidth
3350 GB/s
FP16
495 TFLOPS
Brand
NVIDIA
Series
NVIDIA
Year
2023
Arch
Hopper

Price by provider

cheapest rate · $/GPU/hr · lower is better
Verda
0 regions
$1.138/hr
GCP
0 regions
$1.424/hr
AWS
0 regions
$1.887/hr
Nebius
0 regions
$2.150/hr
RunPod
0 regions
$2.390/hr
Cudo
0 regions
$2.653/hr
Lambda
0 regions
$3.290/hr
OCI
0 regions
$80.000/hr

8 providers

ProviderPrice RangeMedianTypeCPU / RAMAction
VE
Verda
Best price
$1.138$26.000
$3.900 Spot
30-176
120-1480GB
Browse
GC
GCP
$1.424$126.643
$36.008 Spot
26-208
234-1872GB
Browse
AW
AWS
$1.887$92.467
$17.594 Spot
16-192
256-2048GB
Browse
NE
Nebius
$2.150$30.800
$10.525 Spot
16-128
200-1600GB
Browse
RU
RunPod
$2.390$31.584
$10.915 Spot
8-224
125-2008GB
Browse
CU
Cudo
$2.653$2.653
$2.653Mixed
9-9
160-160GB
Browse
LA
Lambda
$3.290$31.920
$8.380Mixed
26-208
215-1933GB
Browse
OC
OCI
$80.000$80.000
$80.000Mixed
224-224
2048-2048GB
Browse
Best Deal Right Now
VE
Verda
$1.138/hr
per GPU / hour
Configurationx1 · H100
System RAM120-1480GB
Region
Launch on Verda

Estimated Running Cost

At cheapest rate · per GPU

1 hour$1.14
24 hours$27.31
1 week$191.18
1 month$19937.76

Compare

See how H100 stacks up against other GPUs.

H100 vs H100