Skip to content

NVIDIA H100

The default datacenter GPU for training and large-model inference.

The lowest price you can rent right now is $2.50/GPU/hr, at Hyperstack in CANADA-1.

Cheapest H100 in stock right now

As of Nov 16, 22:30 UTC

$2.50/GPU/hr

$0.67 (21%) belowthe 7-day median of $3.17

  • Low $2.50
  • Median $3.17
  • High $4.39
Line chart with 168 data points. Use the left and right arrow keys to move between data points, Page Up and Page Down to jump by a tenth of the data, Home and End to jump to the first and last, Enter or Space to activate the current point, and Escape to dismiss the readout.

$2.50 at Hyperstack was in stock 20% of the last 7 days

Rent it at Hyperstack n3-H100x1 in CANADA-1
  • Hopper
  • 80 GB HBM3
  • 8 providers

The lowest H100 price is one you can get: $2.50/GPU/hr.

Hyperstack's $2.50/GPU/hr instance was in stock 56% of the last 30 days. Spot starts at $2.00/GPU/hr.

How we pick these prices · What availability means

Provider by provider

Each provider's lowest on-demand H100 price, and how often that instance was in stock over the last 30 days.

ProviderLowest priceInstanceRegionIn stock, last 30 daysLowest in stock at last scan
Hyperstack$2.50/GPU/hrn3-H100x1CANADA-156%$2.50/GPU/hr
RunPod$2.89/GPU/hrH100 PCIeUS-KS-2Only reports in stock$4.39/GPU/hr
Scaleway$3.16/GPU/hrH100-1-80Gfr-par-267%$3.16/GPU/hr
Lambda$3.79/GPU/hr1x H100 (80 GB PCIe)us-west-337%$3.79/GPU/hr
DigitalOcean$4.41/GPU/hrH100 GPU - 1Xams347%Sold out
AWS$6.88/GPU/hrp5.48xlargeus-east-1Not publishedNot published
Azure$6.98/GPU/hrStandard_NC40ads_H100_v5eastusNot publishedNot published
Oracle Cloud$10.00/GPU/hrBM.GPU.H100.8globalNot publishedNot published
Google Cloud$11.06/GPU/hra3-highgpu-8gus-central1Not publishedNot published

AWS, Azure, Oracle Cloud, and Google Cloud list it from $6.88/GPU/hr, 2.8× the lowest price here. That's the published on-demand rate, not a scan error: reserved capacity costs less, and spot starts at $0.90/GPU/hr at AWS. Neither reports stock.

How often each provider had one in stock

Line chart with 719 data points. Use the left and right arrow keys to move between data points, Page Up and Page Down to jump by a tenth of the data, Home and End to jump to the first and last, Enter or Space to activate the current point, and Escape to dismiss the readout.

What the NVIDIA H100 is

NVIDIA's Hopper datacenter GPU. Hopper added FP8 and the Transformer Engine, and it has about three times the peak BF16 throughput of the A100 it replaced. Providers sell it as SXM, in 8-GPU boards joined by NVLink, and as a PCIe card. The PCIe card has less memory bandwidth and a lower power limit, and the H100 NVL variant has 94 GB.

Architecture
Hopper
Memory
80 GB HBM3
Memory bandwidth
3.35 TB/s
BF16 compute
989 TFLOPS, dense
Lowest precision
FP8
For rent here as
PCIe, PCIe NVLink, and SXM

Other GPUs to consider

Head to head:H100 vs A100H100 vs H200H100 vs B200H100 vs RTX PRO 6000

197 NVIDIA H100 instances

Availability
n3-H100x1Hyperstack · CANADA-1
$2.50/GPU/hr
—
available
n3-H100x1-bigrootHyperstack · CANADA-1
$2.50/GPU/hr
—
available
H100-1-80GScaleway · fr-par-2
$3.16/GPU/hr
—
available
H100-2-80GScaleway · fr-par-2
$3.16/GPU/hr
$6.31/hr total
available
H100-SXM-2-80GScaleway · fr-par-2
$3.65/GPU/hr
$7.29/hr total
available
1x H100 (80 GB PCIe)Lambda · us-west-3
$3.79/GPU/hr
—
available
H100 SXMRunPod · AP-IN-1
$4.39/GPU/hr
—
available
H100 SXMRunPod · AP-JP-1
$4.39/GPU/hr
—
available
H100 SXMRunPod · CA-MTL-1
$4.39/GPU/hr
—
available
H100 SXMRunPod · EU-FR-1
$4.39/GPU/hr
—
available
H100 SXMRunPod · US-GA-2
$4.39/GPU/hr
—
available
H100 SXMRunPod · US-NE-1
$4.39/GPU/hr
—
available
p5.48xlargeAWS · us-east-1
$6.88/GPU/hr
$55.04/hr total
not reported
p5.4xlargeAWS · us-east-1
$6.88/GPU/hr
—
not reported
p5.48xlargeAWS · us-east-2
$6.88/GPU/hr
$55.04/hr total
not reported
p5.4xlargeAWS · us-east-2
$6.88/GPU/hr
—
not reported
p5.48xlargeAWS · us-west-2
$6.88/GPU/hr
$55.04/hr total
not reported
p5.4xlargeAWS · us-west-2
$6.88/GPU/hr
—
not reported
Standard_NC40ads_H100_v5Azure · eastus
$6.98/GPU/hr
—
not reported
Standard_NC80adis_H100_v5Azure · eastus
$6.98/GPU/hr
$13.96/hr total
not reported
Standard_NC40ads_H100_v5Azure · eastus2
$6.98/GPU/hr
—
not reported
Standard_NC80adis_H100_v5Azure · eastus2
$6.98/GPU/hr
$13.96/hr total
not reported
Standard_NCC40ads_H100_v5Azure · eastus2
$6.98/GPU/hr
—
not reported
Standard_NC40ads_H100_v5Azure · westus2
$6.98/GPU/hr
—
not reported
Standard_NC80adis_H100_v5Azure · westus2
$6.98/GPU/hr
$13.96/hr total
not reported
Standard_NC40ads_H100_v5Azure · westus3
$6.98/GPU/hr
—
not reported
Standard_NC80adis_H100_v5Azure · westus3
$6.98/GPU/hr
$13.96/hr total
not reported
p5.48xlargeAWS · eu-north-1
$7.36/GPU/hr
$58.89/hr total
not reported
p5.4xlargeAWS · eu-north-1
$7.36/GPU/hr
—
not reported
Standard_NCC40ads_H100_v5Azure · centralus
$7.89/GPU/hr
—
not reported
Standard_NCC40ads_H100_v5Azure · southcentralus
$7.89/GPU/hr
—
not reported
p5.48xlargeAWS · ca-central-1
$7.91/GPU/hr
$63.30/hr total
not reported
Standard_NC80adis_H100_v5Azure · indonesiacentral
$8.17/GPU/hr
$16.33/hr total
not reported
Standard_NC80adis_H100_v5Azure · malaysiawest
$8.17/GPU/hr
$16.33/hr total
not reported
Standard_NC40ads_H100_v5Azure · indonesiacentral
$8.17/GPU/hr
—
not reported
Standard_NC40ads_H100_v5Azure · malaysiawest
$8.17/GPU/hr
—
not reported
p5.48xlargeAWS · ap-south-1
$8.26/GPU/hr
$66.05/hr total
not reported
p5.4xlargeAWS · ap-south-1
$8.26/GPU/hr
—
not reported
Standard_NC40ads_H100_v5Azure · northcentralus
$8.38/GPU/hr
—
not reported
Standard_NC80adis_H100_v5Azure · northcentralus
$8.38/GPU/hr
$16.75/hr total
not reported
Standard_NC40ads_H100_v5Azure · northeurope
$8.38/GPU/hr
—
not reported
Standard_NC80adis_H100_v5Azure · northeurope
$8.38/GPU/hr
$16.75/hr total
not reported
Standard_NC40ads_H100_v5Azure · southcentralus
$8.38/GPU/hr
—
not reported
Standard_NC80adis_H100_v5Azure · southcentralus
$8.38/GPU/hr
$16.76/hr total
not reported
Standard_NC40ads_H100_v5Azure · centralus
$8.59/GPU/hr
—
not reported
Standard_NC80adis_H100_v5Azure · centralus
$8.59/GPU/hr
$17.17/hr total
not reported
Standard_NC40ads_H100_v5Azure · canadacentral
$8.59/GPU/hr
—
not reported
Standard_NC80adis_H100_v5Azure · canadacentral
$8.59/GPU/hr
$17.18/hr total
not reported
p5.48xlargeAWS · ap-northeast-1
$8.60/GPU/hr
$68.80/hr total
not reported
p5.4xlargeAWS · ap-northeast-1
$8.60/GPU/hr
—
not reported
Page 1 of 4

Agents can read the Markdown version, which lists every current instance.