Most of the adequacy debate in ERCOT runs through reserve margins and effective load-carrying capability. Both blend resources together by their expected contribution. I wanted a narrower, structural number. How much of the summer peak rests on capacity that can be dispatched on command, whatever the weather?
I call it the Thermal Coverage Ratio: net summer capacity of gas and coal units in the ERCOT balancing authority, divided by the annual summer peak. It is not a reserve margin, and nobody argues thermal should cover 100% of the peak. It measures dependence, not adequacy.
The result. In 2019, ERCOT’s summer peak was 74,820 MW and dispatchable thermal capacity was 73,487 MW. Coverage was 98%. At the July 2026 record of 91,134 MW, thermal capacity stood at 75,775 MW. Coverage was 83%. The gap between peak and thermal capacity grew from 1.3 GW to 15.4 GW.
What was built. Combined cycle capacity has been essentially flat since 2019, at about 39.6 GW. Net thermal growth came from combustion turbines (+3.6 GW) and reciprocating engines (+0.6 GW), offset by coal and gas steam retirements of about 2.1 GW. Texas built peaking capacity, not baseload.
What filled the gap. Batteries, mainly. ERCOT now has 19.2 GW of battery storage, more than the entire gap. But the fleet holds 31.7 GWh, an average duration of 1.65 hours, and only about 0.5 GW is rated four hours or longer. At a 6 pm peak with solar still producing, the gap is covered. On a long evening shoulder, duration becomes the binding constraint.
Method notes for planners. Capacity is from EIA Form 860M, with the Operating and Retired tabs combined to rebuild the fleet as of each year’s peak month. Nuclear (4,980 MW) is excluded because it doesn’t follow load. Including it adds about 5.5 points. Imports, demand response and storage are excluded by design. Removing units now listed as out of service lowers 2026 to 81.3%. The 2025 uptick to 89% reflects a milder summer, not new capacity.
This is the first market in a series. PJM follows. I’d value challenge from anyone who plans these systems. Is the 1.65-hour duration profile already reflected in how you think about evening adequacy, or is the megawatt figure still doing too much of the work?
Full analysis, method and a downloadable data pack listing every unit behind every number: https://alexmarshallenergy.com/perspectives/ercot-thermal-coverage-ratio