VISION

One API key. Every GPU cloud behind it.

You shouldn't have to pick a GPU provider. Describe the compute you need and Opengrid routes the job to whichever provider wins on price, availability, region and reliability.

POST /v1/route · 8× H100 · US · < $2.50/hr · done within 12 hours

Illustrative request. The market view ships first; the routing API is in development.

The price of a GPU moves in weeks, not years.

−44%AWS cut its H100 on-demand price in a single announcement, Jun 2025
+40%One-year H100 contract price, Oct 2025 to Mar 2026
$2.89–$11.06Per hour for the same H100, depending on who you rent it from
209GPU clouds. Aggregators reach about 20 of them

The right provider this quarter is the wrong one next quarter. Yet today, moving to it means starting over:

A new account and a new contract
A new API and new tooling
A new bill to reconcile
Re-testing who is actually reliable

So teams pick once and overpay for as long as they stay.

Today's tools give half the answer.

Market data, no control.

Indices and price sites show you a price. They cannot run your job.

Compute, no market data.

Aggregators and clouds run your job. They cannot tell you if the price is fair or the provider reliable.

Opengrid is one key across both: the market view and the routing.

How routing works.

01
Describe
GPU type, count, region, price ceiling, deadline. One request.
02
Discover
Live prices and availability pulled from every connected provider.
03
Score
Rank options on price, location, reliability and performance.
04
Place
Provision on the winning provider through its own API.
05
Bill
One invoice and one dashboard, whichever provider ran it.

Your code never changes when the market does. Switch providers by changing nothing at all.

Where this goes.

NOW
See the market.
Pricing, availability, capacity and regions across providers in a single live view.
NEXT
Route the workload.
One API key, every GPU cloud behind it. Opengrid picks where the job runs.
LATER
Trade compute.
Real routed jobs produce real prices: a live price for every GPU and region, to buy, sell or hedge.

The underlying GPU shouldn't matter. The underlying cloud shouldn't matter. What matters is getting the compute you need, where you need it, when you need it, at the right price.

Every GPU. Every provider. One market.

Get early access