Logo
Logo

H100
Pricing & Specs

H100
Pricing & Specs

H100
Pricing & Specs

H100
Pricing & Specs

TensorPool only charges a fraction of the money you save.

Basic Plan

Sign up in seconds

$1.99/hr for 1-8 H100s

$2.49/hr for multinode H100 clusters

As low as $1.40/hr for reservations

Never pay for idle GPUs

Normal support

No required cloud configuration

Enterprise

Contact Us

Use your own cloud or on-prem compute

Analytics & Monitoring Tools

Spend controls

Dedicated Support

99.9% Uptime SLA

Basic Plan

Sign up in seconds

$1.99/hr for 1-8 H100s

$2.49/hr for multinode H100 clusters

As low as $1.40/hr for reservations

Never pay for idle GPUs

Normal support

No required cloud configuration

Enterprise

Contact Us

Use your own cloud or on-prem compute

Analytics & Monitoring Tools

Spend controls

Dedicated Support

99.9% Uptime SLA

Basic Plan

Sign up in seconds

$1.99/hr for 1-8 H100s

$2.49/hr for multinode H100 clusters

As low as $1.40/hr for reservations

Never pay for idle GPUs

Normal support

No required cloud configuration

Enterprise

Contact Us

Use your own cloud or on-prem compute

Analytics & Monitoring Tools

Spend controls

Dedicated Support

99.9% Uptime SLA

Basic Plan

Sign up in seconds

$1.99/hr for 1-8 H100s

$2.49/hr for multinode H100 clusters

As low as $1.40/hr for reservations

Never pay for idle GPUs

Normal support

No required cloud configuration

Enterprise

Contact Us

Use your own cloud or on-prem compute

Analytics & Monitoring Tools

Spend controls

Dedicated Support

99.9% Uptime SLA

Compare All Plans & Features

Compare All Plans & Features

Compare All Plans & Features

Compare All Plans & Features

Basic Tools

Lambda Labs

TensorPool

Traditional Clouds (GCP, AWS, etc.)

Creation

Creation

On-demand GPU Costs / hr

Creation

Creation

Creation

H100 SXM 1-8 GPUs

$3.29

$1.99

$14

H100 SXM 1+ node

N/A

$2.49

$14

Basic Tools

Lambda Labs

TensorPool

Traditional Clouds (GCP, AWS, etc.)

Creation

Creation

On-demand GPU Costs / hr

Creation

Creation

Creation

H100 SXM 1-8 GPUs

$3.29

$1.52

$14

H100 SXM 1+ node

N/A

$1.52

$14

Google Colab

Core features

Setup Time

0 mins

24/7 Support

GPU Costs / hr

T4 8 v CPU

$0.79

T4 8 v CPU

$0.91

A100 40 GB 12 vCPU

$3.79

TensorPool

Core Features

Setup Time

0 mins

24/7 Support

GPU Costs / hr

T4 8 vCPU

$0.30

L4 8 vCPU

$0.40

A100 40 GB 12 vCPU

$1.52

Traditional Clouds (GCP, AWS, etc.)

Core Features

Setup Time

1 hr+

Ads Management

GPU Costs / hr

T4 8 vCPU

$0.77

L4 8 vCPU

$0.89

A100 40 GB 12 vCPU

$3.71

Basic Tools

Lambda Labs

TensorPool

Traditional Clouds (GCP, AWS, etc.)

Creation

Creation

On-demand GPU Costs / hr

Creation

Creation

Creation

H100 SXM 1-8 GPUs

$3.29

$1.52

$14

H100 SXM 1+ node

N/A

$1.52

$14

Frequently Asked Questions

Frequently Asked Questions

What is TensorPool?

What are the H100 machine specs?

What is the NFS performance?

What is the NVMe performance?

How can I use TensorPool at my company?

When should I use TensorPool?

Why should I use TensorPool?

How did TensorPool start?

Why don't I just use Google Colab?

How can TensorPool ensure I'll always get access to an H100 on-demand?

What is TensorPool?

What are the H100 machine specs?

What is the NFS performance?

What is the NVMe performance?

How can I use TensorPool at my company?

When should I use TensorPool?

Why should I use TensorPool?

How did TensorPool start?

Why don't I just use Google Colab?

How can TensorPool ensure I'll always get access to an H100 on-demand?

What is TensorPool?

What are the H100 machine specs?

What is the NFS performance?

What is the NVMe performance?

How can I use TensorPool at my company?

When should I use TensorPool?

Why should I use TensorPool?

How did TensorPool start?

Why don't I just use Google Colab?

How can TensorPool ensure I'll always get access to an H100 on-demand?

What is TensorPool?

What are the H100 machine specs?

What is the NFS performance?

What is the NVMe performance?

How can I use TensorPool at my company?

When should I use TensorPool?

Why should I use TensorPool?

How did TensorPool start?

Why don't I just use Google Colab?

How can TensorPool ensure I'll always get access to an H100 on-demand?