The Data Center Power Index

Eight North-American ISOs ranked across queue depth, average power cost, build verdicts, and curtailment risk. Click any ISO for the full diagnostic.

Constraint Score combines queue wait time, reserve margin proximity to NERC floor, demand-growth YoY, and 30-day grid-emergency frequency. Excess Power Score is the contrarian metric: reserve-margin headroom, generation additions queued <12mo, renewable curtailment volume, queue approval rate, stranded interconnection at retiring plants, and behind-the-meter industrial generation. Updated daily from ISO public filings + DC Hub's grid extractors.

DCPI launched as a US-only, ISO-by-ISO read. It isn't one any more — metros outside North America are scored on the same two axes, by the same daily recompute, and appear in the same ranking as the US ones. Every figure in this section is measured when the page renders; if one can't be measured, the claim ships without the number instead of with a stale one.

Read the halves separately: markets scored and grid regions measure the index itself. Countries and facilities measure the physical footprint the market list is built from — DC Hub tracks facilities in many more countries than the index currently scores, and the market universe is drawn from that footprint, not the reverse.

Where each figure comes from, so you can check it. Markets scored is the canonical distinct-market count — the same figure /api/v1/stats/canonical publishes as dcpi_markets_scored. It sits below the 323 entries listed further down this page because retired alias slugs and the rural aggregate regions collapse out of it. Grid regions is the row count of /api/v1/dcpi/iso-comparison, computed from this same published score set. Footprint counts are countries_covered and facilities_distinct on /api/v1/stats/canonical. Free for press citation, like the rest of the index.

Every market in the DC Hub Power Index — open any market's free detail page: