Get the Effective Utilization Index history
The daily Effective Utilization Index series. Each row carries occupancy, the effective utilization rate, the index level, the market price, the price per useful GPU-hour, revenue per available GPU-hour, and the demand/supply decomposition against the previous published day.
demandEffect and supplyEffect sum exactly to occupancyChange, which is the point: occupancy rising because buyers took more GPUs is a different market from occupancy rising because a provider withdrew capacity, and a headline rate cannot tell them apart.
Days with fewer than 24 readings are omitted from the series and are not used as a decomposition baseline, so a scrape gap cannot manufacture a demand spike on the following day. The window defaults to the trailing 30 days and is capped at 400 days.
Overview
The daily Effective Utilization Index series, with each day’s move decomposed into a demand effect and a supply effect that sum exactly to the observed change. That decomposition is the reason to read this series rather than a plain utilization rate: occupancy rising because buyers took more GPUs is a different market from occupancy rising because a provider withdrew capacity. See the Effective Utilization Index concept page for the methodology.Authorizations
API key passed as a Bearer token. Create one in the dashboard.
Query Parameters
Restrict to a single GPU model. Defaults to every public model, aggregated.
H100 SXM, H200, A100 SXM4, B200, RTX 5090, RTX PRO 6000 WS Occupancy region. Defaults to global.
"global"
Inclusive start of the window in YYYY-MM-DD. Defaults to 29 days before endDate.
"2026-07-01"
Inclusive end of the window in YYYY-MM-DD. Defaults to today.
"2026-08-12"
Rebase anchor for the index level. Defaults to 2026-01-01.
"2026-01-01"
Response
Daily Effective Utilization Index rows for the requested window.

