BurstDock
GPU COMPUTE. ON YOUR TERMS.WORKLOADS / USAGE / CONTROL

GPU compute.
Without the hunt.

From RTX to NVIDIA Blackwell. Find the GPU, memory and price that fit your next breakthrough. Train, fine-tune, render and run inference with BurstDock.

Deploy a GPU
91GPU models & variants
44Countries to deploy in
$0.0613/hrStarting price · USD
BIG COMPUTE. CLEAR PRICES.

Pick your power.

Explore GPU pricing
80 GB VRAM

H100 SXM

From$2.4306/hr
1 GPU server · 4 countries
140 GB VRAM

H200 NVL

From$4.5139/hr
1 GPU server · 4 countries
179 GB VRAM

B200

From$7.0443/hr
1 GPU server · 1 country
269 GB VRAM

B300

From$11.7318/hr
1 GPU server · 1 country
80 GB VRAM

A100 SXM4

From$0.4945/hr
1 GPU server · 4 countries
45 GB VRAM

L40S

From$1.0093/hr
1 GPU server · 4 countries
24 GB VRAM

RTX 4090

From$0.4473/hr
1 GPU server · 12 countries
32 GB VRAM

RTX 5090

From$0.5436/hr
1 GPU server · 22 countries

USD per server-hour, including configured storage. Starting prices reflect available configurations; your selected GPU, location and disk determine the total.

For the people building what’s next.Explore GPU Cloud
01 / LESS FRICTION. MORE FOCUS.Your work. In one place.

Your work is complex.
Your workspace
shouldn’t be.

Big work starts with an idea. Keeping it moving takes visibility: into what’s running, what it’s using and who has access.

BurstDock brings workload management, usage tracking and prepaid credit together. One workspace. A clearer view of your compute.

Open your workspace
01

Your workloads.

Create, start and stop from one place. Follow each workload from request to completion.

CONTROL
02

Your usage.

See runtime and usage by session. Keep reservations, available credit and activity in view.

INSIGHT
03

Your boundaries.

Organize your team with defined roles and workspace access. Put the right controls in the right hands.

OWNERSHIP
02 / SEE HOW SCALING WORKSSCALING EXPLAINED

Traffic goes up.
Capacity follows.

Trigger a spike. Stop the traffic.
Compare fixed, autoscale and serverless.

01 — Choose how to scale

Add GPUs when demand grows. Scale down after a quiet period; keep one ready.

20requests / sec
No traffic100 req/s
0s
EXAMPLE WORKLOAD, IN MOTIONServing traffic
INCOMING20 req/s
QUEUE0
PROCESSED0 / sec
GPU 01Ready · idle
GPU 02Ready · idle
GPU 03Off
GPU 04Off
GPU 05Off
GPU 06Off
GPU 07Off
GPU 08Off
READY GPUs2 / 8
UTILIZATION0%
GPU TIME0 sec
DemandReady capacityreq/s
Recent historyNow
Send a traffic spike to see capacity follow demand.
ILLUSTRATION MODEL

Example workload: 8 workers, 10 requests/s per worker, 2s startup and 4s scale-down delay. These values explain the animation; GPU choices and prices are listed above and on the pricing page. GPU time includes startup and idle time. Switching modes keeps the current workload so you can see the difference.

03 / BUILT FOR GPU WORKFrom idea to compute.

Your workloads.
Built to move.

BurstDock gives developers and teams a clear way to deploy GPU compute and keep workloads, access, usage and spend organized.

From the first experiment to production workloads, everything stays focused on the compute you need and the work you are building.

Developer resources
GPU COMPUTEAI WORKLOADSDEVELOPER FOCUSED
What does BurstDock bring together?

Workload lifecycle management, organization workspaces, role-based team access, prepaid credit and session-level usage tracking. The integration layer also exposes workload operations to compatible tools.

What is the current GPU availability?

Compare popular GPU models, memory and starting prices. Choose your exact configuration in your workspace. Explore GPU pricing

How can we explore an integration?

Our developer documentation covers integrations, workflows and supported tooling. Read the integration documentation

YOUR NEXT BIG IDEA STARTS HERE.

Let’s make
it burst.