Skip to content
Study 2: Beyond 19 GW → Cross-Market Excursion

ERCOT Says the Opposite

We ran the same spatial analysis on ERCOT's 6-zone model. The result is directionally opposite to PJM: in ERCOT, concentrating data centers in DFW is the best strategy, not the worst.

Cross-Market Excursion
Confidence level: Directional.

Zone capacity fractions derived from EIA-860 generator data (2023) and ERCOT load zone geography: North Texas (NTX) 81% firm, Houston (HOU) 93% firm, West Texas (WTX) 14% firm, South Texas (STX) 40% firm, Panhandle (PAN) ~0% firm. Based on public regulatory filings (1,897 ERCOT generators, 134 GW installed), not validated against ERCOT's internal nodal model. Hours-unserved curves are simulation outputs; zone ranking is robust to parameter variation.

We ran the same spatial analysis on ERCOT's 6-zone model. The directionally opposite result:

PJM

Worst zone
Dominion / NoVA
Best strategy
Distribute away
Why

Limited local generation, relies on transfer-constrained imports

ERCOT

Worst zone
Panhandle (wind-heavy)
Best strategy
Concentrate in DFW
Why

DFW has the most gas capacity relative to demand. Wind-rich zones lack firm capacity for nighttime DC load.

Finding 1

ERCOT's optimal siting is the opposite of PJM's. In PJM, distribute away from Northern Virginia. In ERCOT, concentrate in gas-heavy zones like North Texas/DFW. The difference: PJM's bottleneck is transmission, ERCOT's bottleneck is generation.

ERCOT Per-Zone Tipping Points

ERCOT per-zone tipping points — all DC in one zone

The directional finding: There is no universal answer to "where should data centers go." In PJM, distribute. In ERCOT, concentrate where gas capacity exists. The right strategy depends on the grid structure.

Finding 2

There is no universal answer to "where should data centers go?" The optimal location depends on the grid's specific constraint—transmission capacity (PJM), generation mix (ERCOT), or both. This is why ADM starts with the grid, not the data center.