Register your GPU.

Compute fleet counts the GPUs Compute rents from Vast that are up now. The community pool counts the Runpod GPUs that run every coin’s images and are up now; coin GPU budgets pay for them, and any community can fund more. Neither is self-reported. Community machine reports count matching inventory received within 90 seconds. Paused, expired, revoked, and disconnected nodes do not contribute to the totals. Reports confirm a connection; hardware qualification and unique-device verification are separate.

Compute capacity counts recently connected, operator-approved GPUs with matching runtime and model qualification, available job capacity, and enabled dispatch. The pilot qualifies one GPU per node. Throughput needs a workload benchmark; memory and GPU counts cannot establish TFLOPS.

Completed work counts managed and contributor jobs with an accepted result, stored output, and settled charge. Contributor acceptance also records unpaid USD compensation. Preview jobs are excluded. Jobs in progress include dispatch and result validation; this is not a GPU utilization measurement.

Updates refresh every 15 seconds while this section is visible. An unsuccessful refresh retains the previous observation and marks it out of date. Individual machines appear in the history below only under an anonymous alias; no owner, wallet, machine name, address, or private job details are published.