NextFin News - The operator of Texas' power grid is forecasting demand to approach its all-time record this week as a late-summer heat wave drives air-conditioning load toward the breaking point — and wind generation, the state's largest source of electricity by installed capacity, is expected to sag during the very hours the grid needs it most. ERCOT warned that peaks would run in the upper 80,000- to lower 90,000-megawatt range as widespread highs top 100 degrees, a setup that brushes against the 91,089 MW record set just weeks ago on July 22.
The warning carries a specific operational edge. In its August Monthly Outlook for Resource Adequacy, ERCOT flagged an "extremely low wind scenario" for the month, modeling August peak demand at 78,465 MW and noting that lower wind generation was forecast during afternoon peak hours. That combination — demand rising with temperature while supply falls with the calm — is the classic Texas stress test. It is also the scenario that separates a grid that has merely added capacity from one that has added resilience. For power traders, the setup is the textbook "wind drought": the one weather pattern that attacks both sides of the balance sheet at once, lifting load and stripping supply in the same hours.
The Record That Just Got Set — and the Prices That Didn't Follow
The backdrop to this week's forecast is a record that fell in July. On July 22, 2026, ERCOT load climbed past 91 GW for the first time, peaking at 91,089 MW according to the grid operator's official records — shattering the long-standing high of 85,508 MW set in August 2023. The record had already been broken once the day before, when load reached 87,533 MW, and ERCOT's own day-ahead model had forecast only 89 GW for the 22nd. The miss mattered: at 101.6 degrees Fahrenheit during the 5 p.m. hour, historical data showed that 100-101 degree days had previously topped out around 85 GW. The same heat that used to produce an 85 GW day now produces a 90-plus GW day. Weather sets the shape of the curve; something else has been raising the floor underneath it.
What did not happen was a price explosion. In August 2024, when the grid last approached these levels, real-time prices hit the system-wide offer cap of $5,000 per megawatt-hour and averaged roughly $3,000/MWh at the 6 p.m. peak. On July 22, 2026, real-time prices stayed comparatively muted, reaching only $378/MWh, with upside pressure emerging only after battery discharge fell off around 10 p.m. Day-ahead power for the 9 p.m. hour at the North hub settled at $237.83/MWh — a scarcity premium, but not a panic. The day-ahead difference in highs between the two events was about $300/MWh, a quiet revolution in a market famous for its spikes.
Load on the 22nd broke the prior day's peak of 87,533 MW early in the afternoon, just after 12 pm, and didn't slow down. It kept climbing, tracking well above the forecast into the early evening, eventually peaking above 91 GW.
The gap between the two outcomes is the story. Solar output hit a record 34,665 MW the day before the demand record, and battery discharge hit a record 11,980 MW on the record day itself. Solar at times supplied more than 45% of all electricity on the grid during the heat wave, and the increase in output from solar and batteries "kept substantial downward pressure on prices during a period with such a high load," as grid analysts put it. Natural gas remained the largest single source of generation through the evening ramp, followed by storage, coal, and wind — but the marginal megawatts that set the price increasingly came from stored energy rather than from gas peakers screaming into a shortage.
Why a Wind Drought Is the One Setup ERCOT Still Fears
Wind is not a marginal resource in Texas; it is structural. Combined wind and solar generation is on track to surpass natural gas as the system's single largest source of electricity in 2026, with the U.S. Energy Information Administration projecting 189 million megawatt-hours from wind and solar versus 177 million MWh from gas. Natural gas still supplied 43% of ERCOT demand in the first nine months of 2025, down from 47% in 2023 and 2024, but it remains the dispatchable backbone when the sun sets and the air stays still. Wind accounts for roughly a quarter of the grid's generating capacity, and when a summer wind drought hits, the system can lose 15,000 to 25,000 MW of expected wind generation at precisely the moment demand is highest — enough to power several million homes.
That is why ERCOT's low-wind warning matters more than a generic heat forecast. In an energy-only market with a $5,000/MWh offer cap, the loss of that much wind forces expensive gas peakers online and can push real-time prices toward the cap within minutes. Wind droughts are the most dangerous price event for index-rate customers precisely because they combine high demand with low supply simultaneously; a hot, still day is the one scenario where every resource is stressed at once. The July 22 event showed the mitigation working — but it also showed where the stress migrated. Net load, which is total demand minus wind and solar generation and represents what dispatchable plants must actually supply, reached a record 75,733 MW around 8 p.m., well above the previous winter record of 68,398 MW set on a cold February morning in 2025. The tightest hours have shifted from mid-afternoon to the evening.
The grid operator's own reliability modeling reflects cautious confidence rather than complacency: a 0.09% probability of a grid emergency in June 2026 and 0.21% in July. Those are low numbers, but they are probabilities, not guarantees, and they describe single-month risk windows, not the cumulative risk of a multi-day event. ERCOT's summer outlook also noted that approximately 4,300 MW of 4CP load response was observed in 2025 — demand that vanishes when price signals tell large consumers to step back.
The Structural Shift: Batteries Bought Time, Not Duration
Here is the deeper transformation beneath the daily forecast. ERCOT has added wind, solar, and battery capacity faster than any other U.S. grid operator. The grid entered 2026 with approximately 13.9 GW of commercially operational battery storage, surpassing California to lead the nation. Between the August 2024 record peak and the July 2026 record, peak solar output jumped 12.7 GW and battery discharge jumped 8 GW. Batteries, which had essentially no presence at the 2024 peak, discharged at nearly 12 GW during the 2026 evening ramp. By the end of 2026, ERCOT projects more than 45 GW of solar and 27 GW of storage capacity — a sharp increase from roughly 20 GW of solar and 5 GW of storage in 2023.
But batteries have a duration problem, and it defines the limit of the current resilience. Most grid-scale batteries in Texas are two-to-four-hour systems, engineered to bridge the evening ramp from solar to gas, not to carry the grid through consecutive windless days. On July 22, real-time price volatility returned only after battery discharge fell off around 10 p.m. — a small signal with a large implication. Solar and batteries have lowered the price of surviving one hot day. They have not yet bought the system duration.
The demand response piece reinforces the point. During the July record week, roughly 2 to 3 GW of flexible load came off the system during peak hours as large consumers avoided the four highest-priced "4CP" intervals that set their transmission charges for the year. That is a meaningful cushion — but it is also a cushion that depends on price signals working, and on consumers still having load left to cut when the fifth hot day arrives. ERCOT is aware of the gap: the grid operator is developing a Large Load Curtailment Manager that would give operators advance warning when system conditions may require mandatory curtailment of large loads to avoid emergency alerts, a tool enabled by Senate Bill 6 passed in 2025.
The Second-Order Trade: What the Price Signal Is Really Saying
The first-order read of the July 22 event is simple: more solar and batteries meant lower prices at record demand. The second-order read is more consequential for anyone trading power, gas, or ERCOT-exposed equities. The market did not price an emergency; it priced a stressed hour. Day-ahead power at $237.83/MWh for the 9 p.m. hour is a scarcity premium, but it is nowhere near the cap, and it signals that traders believe the incremental generation and reserve margin ERCOT has brought online is, for now, keeping pace with the new baseline of demand.
That belief is a bet on duration, not on capacity. The capacity exists — ERCOT's modeling shows the system can clear a 91 GW day. What the market has not yet been forced to price is five consecutive days above 90 GW with wind output persistently below 10 GW. On that kind of run, the two-to-four-hour batteries that smoothed the July 22 evening ramp would be cycled to depletion, the 4,300 MW of 4CP demand response would already have been called, and the marginal price would be set by gas peakers bidding into a genuinely thin reserve margin. Natural gas prices themselves would become the swing factor: lower gas prices compared with last summer helped keep wholesale electricity prices muted in June, but a sustained heat dome across the South would lift gas demand for power burn at the same time it lifts cooling demand, tightening the very fuel the grid needs most.
This is the asymmetry traders should watch. On the upside, one more record-breaking solar day or one more evening of 12 GW battery discharge reinforces the "resilience is real" narrative and caps power prices. On the downside, a single multi-day wind drought tests whether the storage buildout has solved ERCOT's problem or merely deferred it into the evening hours. The market has priced the first scenario; it has not priced the second.
The Counter-Thesis: Has ERCOT Actually Solved Its Reliability Problem?
The bull case for the grid is data-backed and should not be waved away. ERCOT absorbed a 91 GW day — more than 5 GW above the old record — without scarcity pricing, without emergency conservation alerts, and without the $5,000/MWh spikes that defined the 2024 peak. The U.S. Energy Information Administration forecasts ERCOT demand rising 14% in the first nine months of 2026 compared with the same period in 2025, reaching 425 terawatthours, faster than any other U.S. grid operator. Generation has, so far, kept pace. Prices moving up modestly rather than spiking to the cap is the market saying the stress looks manageable — for now. Wind, solar, and nuclear contributed as much as 55% of total generation on some high-demand days in late June, a fuel-mix transformation that would have been unthinkable five years ago.
The adversarial case is that the true test has not arrived. The July 22 resilience depended on three things aligning on a single day: record solar during daylight, record battery discharge at dusk, and a heat wave that, while severe, did not stretch across multiple consecutive windless days. A multi-day wind drought late in the summer, when stored energy is depleted and gas capacity margins are thin, would expose the duration limit of the battery fleet. A repeat of the February 2021 freeze would expose the fact that solar and batteries contribute little firm capacity when the sun is down and the wind is still. The grid's risk has not disappeared; it has migrated to the evening hours and to longer-duration weather events. And beneath it all sits the planning reality that ERCOT's own December 2025 capacity report showed the planning reserve margin entering negative territory in the net peak load hour for summer 2026 — a warning that capacity additions, however impressive, are chasing load growth that may be accelerating faster than the build.
What Comes Next: The Demand That Does Not Revert
The heat wave and the wind drought are cyclical. They will revert when the weather breaks. The demand growth underneath them is structural, and it is not reverting. ERCOT President and CEO Pablo Vegas told the Texas Senate Business and Commerce Committee on July 29 that statewide peak demand could reach roughly 175,000 MW within six years — nearly double the current all-time record — driven by a wave of data centers alongside population and industrial growth. ERCOT's preliminary long-term forecast, filed in April, had projected 367,790 MW by 2032, a figure regulators described as almost certainly flawed; the grid operator has since paused official long-range forecasts while it rebuilds its methodology for large loads. The very fact that the forecast is being reworked is itself a signal: the old models, built on weather-driven load, no longer capture a grid where data centers draw power around the clock.
The interconnection queue tells the same story in megawatts. Vegas said there are 460,000 MW of generation seeking to connect to the ERCOT grid, about 70% of it solar and batteries and roughly 78,000 MW from natural gas plants. ERCOT's summer outlook projected about 1,725 MW of expected large-load growth between May and September 2026 alone, on top of the roughly 6,000 MW of large loads already observed taking service. The U.S. grid operator with the fastest load growth is building toward a future in which today's record is next year's baseline.
For the rest of this summer, the watch items are specific and quantifiable. First, whether wind output stays below 10 GW on any day demand exceeds 90 GW. Second, whether real-time prices exceed $1,000/MWh for more than four consecutive hours — a threshold that would signal the battery bridge is no longer sufficient. Third, whether the evening net-load record of 75,733 MW is challenged as the heat persists into late August and September, when ERCOT's MORA report continues to model low-wind scenarios against demand that the grid operator itself now expects to exceed 92,211 MW at the summer peak.
Scenarios for the Rest of 2026
The base case is that ERCOT clears the rest of the summer without an emergency, with occasional peaks in the upper 80,000 MW range and prices that spike briefly but settle as batteries and demand response kick in. In this world, the July 22 outcome becomes the template: record demand, muted prices, and a narrative that the storage buildout has tamed the Texas grid.
The downside case requires a specific trigger: a heat dome lasting five or more consecutive days with wind output persistently below 10 GW. In that scenario, battery state-of-charge depletes across successive evenings, 4CP demand response is exhausted, and the marginal price is set by gas peakers bidding into a thin reserve margin. Real-time prices would test $2,000 to $5,000/MWh, and ERCOT would face its first real test of the Large Load Curtailment Manager. This is the scenario the August MORA's "extremely low wind" modeling is designed to anticipate — and the one this week's forecast edges toward.
The upside case is that continued storage additions and the roughly 7,000 MW of new gas projects backed by the Texas Energy Fund come online faster than load growth, pushing the risk window further into the evening and eventually compressing summer price volatility to the point where the wind drought no longer moves markets at all. That is the bull case for ERCOT reliability — and it is a race between construction cranes and air conditioners that will not be settled this summer.
Bottom Line
ERCOT's forecast for record-level demand this week, paired with sagging wind, is the classic Texas stress test — but the grid that faces it in 2026 is not the grid that failed in 2021 or even the one that spiked in 2024. Solar and batteries have bought the system time and lowered the price of surviving a single hot day. They have not yet bought it duration. The real test of Texas grid resilience will come not on the hottest day, but on the fifth consecutive hot, windless day — and on that day, natural gas and demand response, not renewables, will still be the margin that matters.
Explore more exclusive insights at nextfin.ai.
