Grid archive

What the utility published, and when

PG&E and the Texas Comptroller overwrite these publications in place. They keep no history, so nobody can tell you what changed. This archive snapshots them and diffs consecutive captures.

Exported 2026-09-12 · 4 datasets · 2,118,185 rows held

Datasets
4
one snapshot each so far
Rows held
2,118,185
across every dataset
Change events
0
a diff needs two snapshots
Utilities covered
1
PG&E, plus one state roster

Data centre supply pipeline

Every data-centre building a county, city or regional authority publishes, normalised onto one lifecycle vocabulary and shaped after datacenterHawk’s own API. No megawatts: that figure is a lease fact and nobody publishes it.

Utility publications

Empty is the correct result today: each dataset holds exactly one snapshot, and a diff needs two. The archive starts the day it starts.
What this dataset does not contain lists every deliberate omission with its reason — megawatts, lease rates, vacancy, absorption and carrier-on-net fiber — and what was checked before each was ruled out.
PG&E writes -99.99 into numeric columns it will not publish. Every row of the large-load pipeline carries that sentinel in its expected-load column, so no page here states a megawatt figure for a load project. What these datasets support is where, what type, how far along, and what moved.
    Grid Archive — ZoningLabs