Texas Power Crisis 2021: What Happened and Who Paid

The 2021 Texas power crisis was a week in February when an Arctic storm, Winter Storm Uri, knocked out roughly half of the state's generating capacity, forced rolling blackouts that peaked at about 20 GW of shed load and held the wholesale price of electricity at its $9,000 per megawatt-hour cap for days. Texas's health department counted 246 deaths related to the storm. The bills that followed bankrupted the state's largest power cooperative and a retailer whose customers paid wholesale prices directly, while the two largest generators lost billions and survived.

On the evening of Sunday 14 February 2021, ERCOT, the operator of most of the Texas grid, set a winter demand record of about 69 GW (reported). Within hours its plants began to fail. In the early hours of 15 February grid frequency fell below 59.4 hertz for several minutes, and ERCOT later said the system had come within minutes of an uncontrolled blackout that could have lasted weeks or months (reported). The physical crisis lasted four days. The financial one lasted years.

What happened in the Texas power crisis of 2021?

The cold arrived from 10 February. As temperatures fell, the gas system froze from the bottom up. Wellheads, gathering lines and processing plants iced over, and some lost electric power because they sat on circuits that were being cut in the rolling outages. Gas production fell, and the gas-fired plants that depended on that gas tripped. Every other fuel type had problems too: coal piles and instruments froze, wind turbines without cold-weather packages stopped, and at the South Texas Project a frozen feedwater sensor shut down a nuclear reactor.

Texas had been warned. After a February 2011 cold snap, a FERC–NERC inquiry recommended that Texas plants be winterized. Winterization had been voluntary.

Why did the price stay at $9,000 for so long?

ERCOT runs an energy-only market. It has no capacity market; plants earn their fixed costs when prices spike in scarce hours, up to an offer cap that was $9,000/MWh at the time. In a normal year ERCOT's average real-time price was in the tens of dollars per MWh. The scarcity price is meant to reward plants for being available.

On 15 February the Public Utility Commission of Texas ordered ERCOT to set the price at the cap while it was shedding load, on the logic that electricity nobody could get should be priced at scarcity. Load shedding ended on 18 February, but the price stayed at the cap until the morning of the 19th. ERCOT's independent market monitor, Potomac Economics, later concluded that prices had remained at the cap about 32 hours longer than they should have, adding about $16 billion of costs, and recommended a correction. The commission declined to reprice.

The design flaw was not the price. It was the assumption that a price could make plants run. As the book puts it, "Uri showed that a scarcity price cannot make a frozen plant run." A signal that rewards availability does not guarantee it; only physical standards and fuel security do that. Chapter 8 on power markets explains why energy-only markets lean so heavily on those few scarce hours.

Who lost money in Winter Storm Uri?

The losses did not track efficiency. They tracked two things: whether a company's plants physically ran, and what it had promised to deliver.

At the same time, Henry Hub gas stayed near $3 per MMBtu, while Texas hubs traded at hundreds. The gap shows how local the crisis was.

Brazos: a cooperative with a merchant bill

Brazos is a generation-and-transmission cooperative, a member-owned non-profit descended from the rural electrification program of the 1930s. It had little equity because it was not built to carry risk. When its own supply fell short and it had to buy power at $9,000/MWh to serve its member co-ops, a single week's bill exceeded several years of revenue. Its creditors, not its members, absorbed much of the loss.

Griddy: the household as the hedge

Griddy sold households access to the wholesale price for a monthly fee. In ordinary times that was cheap. In Uri it turned households into unhedged traders. The book works through a stylized household on a pass-through tariff using 60 kWh a day for electric heat over four days at $9 a kilowatt-hour: about $2,160 for energy alone, against about $29 on a normal fixed rate. Pass-through retail did not remove the risk. It moved the tail onto the people least able to carry it.

Vistra and NRG: hedges that became short positions

Vistra entered 2021 as the state's largest integrated generator and retailer, with about 18 GW in ERCOT (reported). It had sold forward much of its expected winter output. Its model assumed that in a spike its plants would run and earn the spike. When some gas units lost fuel or froze, the forward sales did not disappear. Vistra had to buy power at the cap to cover them, and to serve TXU Energy's fixed-price retail customers. It first estimated a hit of $900 million to $1.3 billion on 26 February 2021, then raised it to about $1.6 billion net on 26 April, blaming mainly the failure of the gas supply system (company).

The arithmetic is brutal. In the book's stylized example, 1,000 MW sold forward at $50/MWh that cannot run for 32 hours at $9,000/MWh costs about $286 million to replace; for 96 hours, about $859 million. Spread across several gigawatts and partial outages, that is how a profitable fleet loses about $1.6 billion in a week. NRG's position was similar but smaller. Both companies protected their fixed-price retail customers from wholesale prices and absorbed the cost, which is the opposite of what Griddy did.

Why did Vistra survive while Brazos failed?

Both lost money for the same physical reason. The difference was the balance sheet. Vistra and NRG had diversified fleets, access to credit and the liquidity to absorb a loss of that size; Vistra raised liquidity and later issued preferred stock (reported). Brazos, Griddy and Just Energy had none of that cushion.

The case teaches two lessons that run through the whole energy industry. First, risk is correlated. Uri was not one plant failing; the gas and power systems failed together, each making the other worse, at the moment demand peaked. Models that treat outages as independent understate the tail. Second, the worst outcomes fall on the thinnest balance sheets. A hedge or trading book that looks sound in normal weather can turn into a liability in the one week it is tested.

What changed in Texas after Uri?

Politics stepped in where the market had been designed not to need it.

• Weatherization. SB 3 required generators, and later critical gas facilities, to weatherize, with inspections and penalties.
• Price cap. The commission cut ERCOT's system-wide offer cap from $9,000 to $5,000/MWh, effective 2022.
• Governance. SB 2 replaced ERCOT's stakeholder board with an independent board.
• New products and money. ERCOT added firmer reserve products, and voters approved the Texas Energy Fund in November 2023 to lend to new dispatchable plants.
• Securitization. Texas authorized bonds to cover gas utilities' fuel costs and some ERCOT market defaults, repaid by customers over many years (reported).

Mandatory cold-weather standards were also adopted for generators across North America in 2022–24. Five years on, the market looks different. ERCOT's peak demand reached a record of about 91.1 GW on 22 July 2026, yet the real-time price at the peak was only about $378/MWh, because solar and about 14–15 GW of batteries covered the evening ramp (reported). Vistra, meanwhile, became one of the S&P 500's best performers in 2024 as data-center demand repriced firm power, a turn covered in our guide to how the energy industry makes money.

Which factor of the power equation failed?

The Power Equation holds that an energy company creates durable value only when four factors are true at once: supply at a cost the market will pay, a route to market, capital matched to the asset, and a licence to operate. Because they multiply, a zero in any one is fatal.

Winter Storm Uri is the rare case in which all four failed together. Supply could not run at any price, because plants and the gas system feeding them had not been built for the cold. The route to market turned into bills nobody could pay, because a market designed to send price signals sent them to households and co-ops that could not respond. The capital structures of a co-op and two retailers could not survive a week of tail risk. And the market's licence was lost and had to be rebuilt by statute. For a different kind of system failure, compare the PJM capacity-price shock in our explainer on the PJM capacity auction, and the wider design choice in regulated vs deregulated electricity markets. The full set of disasters is in the book's study of energy failures, disasters and bubbles, and company detail is in the Vistra and NRG Energy profiles.

The takeaway

The 2021 Texas power crisis was a physical failure that a market turned into a financial one. The same megawatt-hour was worth more than two hundred times its normal value, and whether that made a company rich or bankrupt depended on whether its plants ran and what it had promised, not on how efficient it was. Large integrated companies lost billions and survived; a cooperative, two retailers and many households did not. The lasting lesson is that a hedge is only as good as the physical asset behind it, and a market keeps its licence only while its outcomes are survivable. Chapter 1 of The Power Equation opens with Uri and builds the book's framework from it.