Compute pricing — a published index, not a quote

Pricing on the Open Compute Network is an administered tariff published openly rather than a negotiated rate. Each resource a unit of work asks for is priced separately — processor time, memory, scratch disk, accelerator time, bytes moved — with premiums for low latency, higher reliability and hardware attestation. A floor computed from the contributing machine's own measured power draw prevents any machine being offered work that would cost it more in electricity than it earns.

Every dimension is priced separately and billed by what was used, not what was requested.
The price index is published continuously with its history.
The network takes a commission on settled work; the buyer is quoted gross either way.
No machine is offered work below its measured electricity cost plus at least 30%.

Elsewhere on this site

The Cube Floor — The Open Compute Network. The live figures on this page are loaded from https://api.thecubefloor.com, which is public and needs no credentials to read.