ERCOT’s Record July Peak Is Set to Shape 2027 Transmission Charges for Large Texas Customers, and the Peak Interval May Be Drifting Later

Hourly peak load in ERCOT reached 91.1 GW at 6:00 p.m. Central on July 22, according to the Energy Information Administration, 6% above the previous record of 85.5 GW set on August 10, 2023. The Texas Tribune reported ERCOT’s own figure for the day at 91,308 MW, with July 21 close behind at 87,403 MW. Natural gas supplied 48% of generation at the peak and solar supplied 32%.

Coverage treated the number as a reliability story. For a subset of Texas commercial and industrial customers, it is also a billing story, and the bill in question is for calendar 2027.

The mechanism. ERCOT allocates transmission cost recovery through four fifteen-minute intervals a year. Under the Four Coincident Peak method, ERCOT identifies the single highest fifteen-minute interval of system demand in each of June, July, August and September, then publishes calculations for each Distribution Service Provider. A demand-metered customer’s instantaneous load during those four intervals is averaged, and that average becomes the transmission billing determinant applied over the following calendar year. KilowattLogic puts the practical threshold at roughly 700 kW of demand, which means 4CP reaches transmission-voltage and larger distribution customers rather than the full commercial rate class.

The determinant is not a monthly maximum and not a rolling twelve-month ratchet. It is four moments, fixed by the weather and by everyone else’s load.

The 2025 intervals. Energy Edge Consulting published ERCOT’s official 2025 intervals, which set the transmission charges now appearing on 2026 bills: June 19 at 17:00 and 77.49 GW, July 30 at 17:00 and 81.91 GW, August 18 at 17:00 and 83.97 GW, and September 4 at 17:30 and 80.08 GW. Three of the four landed on the hour at 5:00 p.m.

The July 22, 2026 record fell in the 6:00 p.m. hour. That is one hour later than three of the four intervals that set this year’s charges, and thirty minutes later than the fourth.

Solar output. Utility-scale solar is the most plausible mechanism for a later peak. Solar met 32% of generation at the July 22 peak, which is what a fleet large enough to flatten the mid-afternoon looks like on a settlement report. As solar output declines into the evening while air conditioning load persists, the system peak slides toward sunset. That reading is interpretation rather than something the EIA report asserts, and one record hour is one data point. It is consistent with the direction of the 2025 intervals, in which September printed at 17:30 rather than 17:00 as the days shortened.

The practical consequence for anyone managing 4CP exposure is that curtailment windows built around a 3:00 p.m. to 5:00 p.m. assumption are aimed at where the peak used to be.

Precedent. This has happened before: the largest operational peak of a month does not automatically become that month’s official interval. Writing for Grid Status in June 2025, Connor Waldoch found that the official June 2024 4CP interval ranked 33rd on real-time operational load. The gap came from batteries: “During the anticipated 6/27 interval, storage charging was greater than 500 MW, but on the 30th, storage charging was only at 63 MW. Once storage load is removed, the result was a 230 MW difference, with 6/30 @ 17:45 having the higher peak load.”

Waldoch’s conclusion was blunt: “With the rise of storage, real-time data is more important than ever for avoiding 4CP intervals.”

A 230 MW swing decided which interval counted on a system carrying more than 77 GW. Grid-scale storage charging behavior, dispatched on price rather than on anything a commercial customer can see coming, now sits between the operational peak a building manager watches and the settlement interval that actually bills.

The manual playbook. The standard 4CP strategy is a notification service plus a human decision: a forecaster flags a probable peak day, the facility drops load for an hour or two, and the customer buys four cheap determinants for the price of four uncomfortable afternoons. That approach requires being right about the day, right about the hour, and now implicitly right about how much the storage fleet happens to be charging when the candidate intervals print.

An asset that shaves a wider window automatically does not need to win any of those three guesses. It pays for that immunity in throughput and cycles rather than in forecasting accuracy, which is a different cost structure, not a free one. The distinction is between hedging a timing risk and predicting it.

The rate. The charge applied to the determinant is rising alongside the determinant itself. KilowattLogic projects a summer 2026 transmission rate above $5.50 per kW per month. The load driving that rate base is not slowing: ERCOT projects roughly 175,000 MW of demand by 2032, attributed by the Texas Tribune to data centers, population growth and industrial projects, and AEP alone submitted 45 GW into ERCOT’s latest interconnection batch process, per RTO Insider. The transmission built to serve that growth is recovered through the same 4CP allocation. Load growth concentrated in large new customers therefore raises both the numerator of the transmission bill and, eventually, the height of the peaks that set who pays it.

Texas is not alone in setting records this summer. The Southwest Power Pool hit 57.9 GW on July 27 at 5:00 p.m. Central, and the EIA cautioned that ERCOT’s figure “could be surpassed again this summer, especially if Texas experiences another heat wave.”

Two limits. First, TDU tariffs differ in how and to whom 4CP determinants are applied, so the exposure is specific to a customer’s utility and service voltage rather than uniform across Texas commercial accounts. Second, September remains open. Three of the four candidate 2027 determinants have already been struck, and the fourth has not, which makes the remaining window the only part of the calculation still available to influence.

The 2027 transmission bill for a large Texas commercial customer will rest largely on four fifteen-minute windows chosen by the weather, at least one of which has been misidentified in real time in a prior year, and none of which ERCOT confirms until its official calculations are published after the season.

Sources