Back
Five years ago, Winter Storm Uri brought the Texas power grid to its knees. Temperatures plunged across the state for nearly a week, power plants froze, natural gas supply lines failed, and the grid operator came within minutes of a total system collapse. More than 4 million Texans lost electricity, many for days. Over 200 people died. It was the worst infrastructure failure in modern Texas history.
In the years since, Texas has quietly built one of the largest renewable energy and battery storage fleets in the world. According to capacity data from the Electric Reliability Council of Texas, the state has added roughly 31 gigawatts of solar capacity and 17 GW of battery energy storage — enough to power millions of homes. Over the same period, the legislature mandated weatherization of power plants and natural gas infrastructure, ERCOT improved its operational procedures, and new market mechanisms were introduced to better coordinate solar and storage.
The results speak for themselves. Since Uri, the Texas grid has faced three major winter storms that each set new all-time winter peak demand records. In every case, the grid held. No rolling blackouts. No load shedding. No emergency curtailments. Demand kept climbing, and the grid kept delivering.
This track record matters because a prominent Texas think tank, the Texas Public Policy Foundation, has published a widely circulated analysis arguing that ERCOT’s reliance on solar and battery storage is making the grid less reliable in winter. The analysis is authored by Brent Bennett and uses real ERCOT data. But as this article will show, Bennett’s own numbers contradict his conclusions — and the actual performance of the grid over the past five years contradicts them even more decisively.
The following chart I worked up offers a quick summary: Texas’ reliability has increased dramatically in recent years in direct proportion to the renewables and battery storage it has added.
The above data tells the story. At the time of Uri, ERCOT had roughly 5 GW of solar and less than 1 GW of battery storage. When Winter Storm Elliott arrived in December 2022, it had 14 GW of solar and 2 GW of storage. By Winter Storm Heather, in January 2024: 22 GW and 4 GW. By Winter Storm Kingston, in February 2025: 30 GW and 9 GW. And now, as we pass the fifth anniversary of Uri: approximately 35 GW of solar and 15 GW of battery storage.
During each of these storms, peak winter demand set a new record — climbing from 74,525 MW during Elliott to 78,349 MW during Heather to 80,525 MW during Kingston. Just three weeks ago, the grid sailed through another major winter storm with over 11,000 MW of operating reserves and ERCOT said it did ​“not anticipate any reliability issues on the statewide electric grid.”
In none of these events did ERCOT order load shedding. This is the track record that Bennett’s analysis asks you to ignore.
Now let’s turn to Bennett’s projected numbers for 2030. His Figure 1 posits that ERCOT could have 103,802 MW of firm output against a speculative peak demand of 110,000 MW — his estimate, not ERCOT’s. That’s a gap of roughly 6 GW. His projected battery fleet by 2030? Forty-three gigawatts.
Read that again: a 6-GW shortfall covered by 43 GW of batteries.
Bennett’s response to this rather obvious mismatch is to reframe the question entirely. Instead of asking whether batteries can cover peak demand windows — which is what they’re designed to do — he converts the entire battery fleet into a single energy metric: 77 GWh, which he says is ​“equivalent to running a single 1 GW thermal power plant for the duration of this three-day storm.” It’s a striking comparison. It’s also irrelevant to how batteries actually operate in ERCOT.
Nobody designs, operates, or dispatches battery storage as a 72-hour baseload resource. Batteries are designed to shave peaks, provide rapid frequency response, and bridge the morning and evening demand ramps when solar output is low. A 43-GW battery fleet can inject enormous amounts of power during exactly the narrow peak windows that Bennett’s own Figure 2 identifies as the problem periods. During Winter Storm Heather, ERCOT’s post-storm analysis confirmed that batteries were ​“partially supplementing the lack of solar generation available” during the coldest pre-sunrise hours — the exact scenario Bennett says they can’t handle.
Perhaps the most revealing aspect of Bennett’s analysis is what he doesn’t discuss: the massive existing fleet of gas, coal, and nuclear generation that forms ERCOT’s backbone. He projects 103,802 MW of firm winter output in 2030. That fleet — overwhelmingly fossil and nuclear — carries the grid through the vast majority of every storm hour in his model. The assumed thermal outage rate is only 12% — a figure drawn from ERCOT’s reliability assessments — meaning 88% of the thermal fleet performs through the modeled storm.
Bennett constructs a scenario in which batteries fail by defining success as continuous 72-hour discharge, while simultaneously taking for granted the thermal fleet of 80-plus GW that keeps the lights on during the bulk of his modeled event. The batteries aren’t replacing that fleet. They’re supplementing it during the peak demand windows that the thermal fleet alone can’t quite cover — which is precisely the role that ERCOT’s system planning envisions for them.
The contrast between Bennett’s theoretical model and actual ERCOT performance is stark. During Winter Storm Elliott, solar contributed roughly 8 GW at peak, and real-time prices dropped from over $3,000/MWh to under $100 within 90 minutes of sunrise. During Heather, large flexible loads curtailed voluntarily, demonstrating the demand-side response that Bennett barely acknowledges. ERCOT CEO Pablo Vegas has specifically identified the growth in battery capacity as ​“perhaps the most significant factor affecting grid stability,” while University of Texas energy professor Michael Webber credited ​“significant investments in more solar and more batteries and demand response” as key factors in the grid’s most recent winter storm performance.
None of these experts are claiming the grid faces zero risk. ERCOT’s probabilistic risk assessment, as reported in NERC’s winter reliability assessment, puts the chance of controlled load shed this winter at about 1.8% — low, but not zero. The question is whether Bennett’s framework for evaluating that risk is sound, and on that point, the data he himself relies on says no.
Bennett’s piece concludes that ERCOT needs ​“market design changes that redirect revenue away from wind and solar and toward resources that can work in all types of weather conditions.” That’s a policy preference dressed up as an engineering conclusion. His own data doesn’t support it.
What his data actually shows is that ERCOT has a manageable peak-demand gap that battery storage is well positioned to address, supplemented by a massive thermal fleet that provides the overwhelming majority of firm capacity during winter events. The December 2025 launch of ERCOT’s Real-Time Co-optimization Plus Batteries (RTC+B) market is specifically designed to optimize exactly this kind of coordination — dispatching storage where and when it creates the most grid value.
The real question isn’t whether batteries can run for 72 hours straight. No one is asking them to. The question is whether the combination of 100-plus GW of firm thermal capacity, a rapidly growing battery fleet, improving demand-response capabilities, and better weatherization standards can keep the lights on during winter storms. The last five years of actual performance — including three consecutive record-breaking winter peaks — provide a clear answer.
Bennett’s analysis works only if you accept his premise that battery storage should be evaluated as a baseload replacement rather than what it actually is: a fast-dispatching, peak-shaving complement to the thermal fleet, which helps dramatically in firming up renewables like wind and solar. Reject that premise, and his crisis narrative dissolves into the numbers he himself provides.
‍
.svg)
