GPU cloud comparison

Rent the right GPU.
Understand the price.

Compare a full GPU setup, see what an hour or a working day could cost, and go straight to the provider when you are ready.

Find your configuration

Start with the GPU. Check the whole price.

Not sure which GPU? Get help →
Showing 6 of 8 configurations · 28 in catalogue

We currently earn no commission from these links. How we fund this site

DigitalOceanGPU machine

NVIDIA H100 SXM 80GB

1× H100-SXM GPU Droplet

1 GPU80 GB per GPUon-demand

Price checked 14 Oct 2026 · valid until 21 Oct 2026, 00:39 UTC
GPU variant inferred from the provider’s official specifications

$4.41 / hour

USD · 1 GPU · full configuration

Compute only · extras excluded
Try it for 1 hour$4.41
An 8-hour day$35.28
Compute estimate: a billing rule is not published. Storage, transfer and tax extra.
Price details & source

Included in this rate

1× H100-SXM GPU Droplet

GPU compute and listed CPU/RAM allocation

Check these costs

Storage, transfer, tax and optional services may add to compute.

Applicable taxes; Extra storage and optional services

Source notes

Price checked 14 Oct 2026 · valid until 21 Oct 2026, 00:39 UTC
GPU variant inferred from the provider’s official specifications

Region not confirmed. Availability not checked.

Billed per second; minimum charge not published. Compute examples are estimates and do not assume the unpublished rule.

GPU Droplets stay billed while powered off; billing ends only when the Droplet is destroyed.

Minimum period is unresolved: the pricing FAQ says 5 minutes; the billing documentation says 60 seconds or $0.01, whichever is higher.

DigitalOcean lists this GPU as NVIDIA HGX H100; HGX boards use SXM modules.

Official pricing source ↗
DigitalOceanGPU machine

NVIDIA H100 SXM 80GB

8× H100-SXM GPU Droplet

8 GPUs80 GB per GPUon-demand

Price checked 14 Oct 2026 · valid until 21 Oct 2026, 00:39 UTC
GPU variant inferred from the provider’s official specifications

$35.28 / hour

USD · 8 GPUs · full configuration

Compute only · extras excluded
Try it for 1 hour$35.28
An 8-hour day$282.24
Compute estimate: a billing rule is not published. Storage, transfer and tax extra.
Price details & source

Included in this rate

8× H100-SXM GPU Droplet

GPU compute and listed CPU/RAM allocation

$4.41 per GPU-hour × 8

Check these costs

Storage, transfer, tax and optional services may add to compute.

Applicable taxes; Extra storage and optional services

Source notes

Price checked 14 Oct 2026 · valid until 21 Oct 2026, 00:39 UTC
GPU variant inferred from the provider’s official specifications

Region not confirmed. Availability not checked.

Billed per second; minimum charge not published. Compute examples are estimates and do not assume the unpublished rule.

GPU Droplets stay billed while powered off; billing ends only when the Droplet is destroyed.

Minimum period is unresolved: the pricing FAQ says 5 minutes; the billing documentation says 60 seconds or $0.01, whichever is higher.

DigitalOcean lists this GPU as NVIDIA HGX H100; HGX boards use SXM modules.

Official pricing source ↗
HyperstackGPU machine

NVIDIA H100 SXM 80GB

8× H100 SXM VM

8 GPUs80 GB per GPUon-demand

Price checked 14 Oct 2026 · valid until 21 Oct 2026, 00:39 UTC

$25.60 / hour

USD · 8 GPUs · full configuration

Compute only · extras excluded
Try it for 1 hour$25.60
An 8-hour day$204.80
Compute estimate: a billing rule is not published. Storage, transfer and tax extra.
Price details & source

Included in this rate

8× H100 SXM VM

GPU compute and listed CPU/RAM allocation

$3.20 per GPU-hour × 8

Check these costs

Storage, transfer, tax and optional services may add to compute.

Applicable taxes; Extra storage and optional services

Source notes

Price checked 14 Oct 2026 · valid until 21 Oct 2026, 00:39 UTC

CANADA-1. Availability not checked.

Billed per minute; minimum charge not published. Compute examples are estimates and do not assume the unpublished rule.

Per-GPU list rate; the full configuration costs this rate multiplied by its GPU count.

Billed per minute; no minimum charge is published.

A stopped (SHUTOFF) VM stays fully billed; hibernation stops GPU, CPU and RAM charges, but the root disk and a kept public IP are still billed.

Official pricing source ↗
LambdaGPU machine

NVIDIA H100 SXM 80GB

1× H100 SXM instance

1 GPU80 GB per GPUon-demand

Price checked 14 Oct 2026 · valid until 21 Oct 2026, 00:39 UTC

$4.29 / hour

USD · 1 GPU · full configuration

Compute only · extras excluded
Try it for 1 hour$4.29
An 8-hour day$34.32
Compute estimate: a billing rule is not published. Storage, transfer and tax extra.
Price details & source

Included in this rate

1× H100 SXM instance

GPU compute and listed CPU/RAM allocation

Check these costs

Storage, transfer, tax and optional services may add to compute.

Applicable taxes; Extra storage and optional services

Source notes

Price checked 14 Oct 2026 · valid until 21 Oct 2026, 00:39 UTC

Region not confirmed. Availability not checked.

Billed per minute; minimum charge not published. Compute examples are estimates and do not assume the unpublished rule.

The source groups instance tables in 8×, 4×, 2× and 1× configurations.

First-come self-service capacity; availability was not checked.

Billed in one-minute increments; no minimum charge is published.

Instances cannot be stopped; a running instance is billed until it is terminated.

Official pricing source ↗
LambdaGPU machine

NVIDIA H100 SXM 80GB

8× H100 SXM instance

8 GPUs80 GB per GPUon-demand

Price checked 14 Oct 2026 · valid until 21 Oct 2026, 00:39 UTC

$31.92 / hour

USD · 8 GPUs · full configuration

Compute only · extras excluded
Try it for 1 hour$31.92
An 8-hour day$255.36
Compute estimate: a billing rule is not published. Storage, transfer and tax extra.
Price details & source

Included in this rate

8× H100 SXM instance

GPU compute and listed CPU/RAM allocation

$3.99 per GPU-hour × 8

Check these costs

Storage, transfer, tax and optional services may add to compute.

Applicable taxes; Extra storage and optional services

Source notes

Price checked 14 Oct 2026 · valid until 21 Oct 2026, 00:39 UTC

Region not confirmed. Availability not checked.

Billed per minute; minimum charge not published. Compute examples are estimates and do not assume the unpublished rule.

The source groups instance tables in 8×, 4×, 2× and 1× configurations.

First-come self-service capacity; availability was not checked.

Billed in one-minute increments; no minimum charge is published.

Instances cannot be stopped; a running instance is billed until it is terminated.

Official pricing source ↗
ModalCode on demand

NVIDIA H100 SXM 80GB

H100 SXM5 GPU task

1 GPU80 GB per GPUserverless

Price checked 14 Oct 2026 · valid until 21 Oct 2026, 00:39 UTC

$3.949 / hour

USD · 1 GPU · GPU resource rate

GPU only · CPU + memory extra
Try it for 1 hour$3.95
An 8-hour day$31.59
Compute estimate: a billing rule is not published. Storage, transfer and tax extra.
Price details & source

Included in this rate

H100 SXM5 GPU task

GPU resource only

Check these costs

CPU, memory and other resources add to this GPU-only charge.

CPU; Memory; Storage; Network egress above plan allowance; Paid plan subscriptions; Applicable taxes

Source notes

Price checked 14 Oct 2026 · valid until 21 Oct 2026, 00:39 UTC

Region not confirmed. Availability not checked.

Billed per second; minimum charge not published. Compute examples are estimates and do not assume the unpublished rule.

Hourly equivalents use exactly 3,600 seconds. A GPU-resource rate is not a complete application bill.

Modal lists region selection at 1.15–1.75× and non-preemptible execution at 3× base prices; those multipliers are not applied here.

CPU (minimum 0.125 core) and memory (minimum 128 MiB) are billed separately per container, as are load time and the 60-second idle keep-alive.

Official pricing source ↗

Showing 6 of 8 matching configurations

Continue a previous comparison