Texas is staring down an electricity crisis that has no modern precedent. Under the most aggressive growth scenarios, the state’s grid operator projects that peak power demand on the ERCOT system could roughly quadruple by 2032, driven overwhelmingly by a wave of data center construction that is already straining the interconnection queue. A federal reliability assessment published by the Federal Energy Regulatory Commission warns that the pace of new load requests threatens to outrun the generation, storage, and transmission buildout needed to keep the lights on.
The numbers are staggering. ERCOT’s 2024 summer peak hit roughly 85 gigawatts. Under high-growth forecasts that account for large-load interconnection requests already in the pipeline, peak demand could climb past 150 GW within the decade. Much of that surge traces to hyperscale data centers planned by companies including Google, Microsoft, Amazon, and Meta. These facilities consume hundreds of megawatts around the clock. They have limited ability to curtail usage during emergencies.
A federal warning with teeth
FERC’s 2024 long-term system assessment treats ERCOT’s isolation as a defining vulnerability. Unlike the Eastern and Western Interconnections that allow neighboring states to share power during shortfalls, the Texas grid stands largely alone. Every gap between projected demand and available supply carries outsized consequences. There is no safety valve of imported electricity when reserves run thin.
The assessment found that large industrial and digital loads are compressing what used to be decade-long demand growth into just a few years. Interconnection requests now routinely propose adding hundreds of megawatts at a single point on the grid. That pattern can create localized transmission bottlenecks even when statewide generation looks adequate on paper. FERC’s modeling flagged these dynamics as a direct threat to system reliability, not a theoretical concern for future planning cycles.
When FERC issues findings like these, grid operators and utilities typically treat them as actionable. The agency has statutory authority over bulk-power-system reliability, and its assessments carry regulatory weight that distinguishes them from industry white papers or think-tank projections.
Reserve margins already under pressure
ERCOT’s own planning documents reinforce the federal warning. The grid operator’s December 2024 Capacity, Demand, and Reserves (CDR) report showed reserve margins tightening significantly under high-load scenarios. In those projections, the margin of surplus generation above forecast peak demand narrows to levels that grid planners consider uncomfortably thin, raising the probability of emergency conditions during extreme summer heat or unexpected plant outages. The CDR is the primary tool ERCOT uses to signal whether enough generation is on track to meet future demand, and the December 2024 edition painted a notably less comfortable picture than earlier versions.
Reserve margins matter because they represent the grid’s cushion. A healthy margin means there is enough spare capacity to absorb surprises. A slim one means a single large plant trip or a hotter-than-expected afternoon can push the system into conservation alerts or, in the worst case, controlled outages. The December 2024 CDR made clear that under scenarios incorporating the flood of large-load interconnection requests, that cushion is eroding faster than new generation can rebuild it.
State lawmakers push back on the forecasts
Not everyone in Austin accepts the projections at face value. State Sen. Charles Schwertner, who chairs the Senate Business and Commerce Committee, has publicly challenged ERCOT’s large-load forecasting methods. He warned that inaccurate numbers carry real risks in both directions. Overestimating demand could trigger billions of dollars in unnecessary generation and transmission buildout, costs that would ultimately land on ratepayers. Underestimating it could leave the grid short during peak periods, raising the specter of rolling blackouts like those that killed more than 200 Texans during Winter Storm Uri in February 2021.
That political pressure is already producing changes. ERCOT and the Public Utility Commission of Texas are revising how they count and verify large-load interconnection requests, an acknowledgment that the sheer volume of applications has outpaced the old methodology. Lawmakers have also floated placing new requirements on large businesses seeking to connect to the grid. That would mark a significant shift in a state that has long prided itself on minimal energy-market regulation.
During the 2025 legislative session, bills including SB 1 targeted the large-load interconnection process directly. The proposals would require developers to demonstrate firm power supply commitments before receiving grid access. The outcome of those efforts will shape whether Texas can screen speculative projects out of the queue before they distort planning assumptions.
Where the data gets thin
Several critical pieces of this story remain unresolved. The exact share of projected load growth attributable to data centers, as opposed to petrochemical plants, hydrogen production, or cryptocurrency mining, has not been broken out in publicly available ERCOT or FERC documents with granular detail by customer type. Industry estimates vary widely. Most trace back to secondary analyses rather than official datasets.
Geography matters just as much as aggregate demand. Where new facilities plan to connect determines whether existing transmission infrastructure can handle the load. West Texas and the Gulf Coast already face well-documented bottlenecks. If data centers cluster in constrained corridors, reliability risks grow faster than statewide totals alone would suggest. If they locate near robust substations or regions with surplus generation, the system-level impact could be less dramatic.
Corporate energy strategies add another layer of uncertainty. Google, Microsoft, Amazon, and Meta have all announced Texas data center investments in press releases, but detailed commitments on on-site generation, battery storage, or demand-response participation have not been filed in ERCOT or PUCT proceedings in a way that allows independent verification. Until those details surface, it is hard to judge whether private-sector mitigation will meaningfully offset the grid strain regulators are warning about.
On the supply side, developers have proposed a substantial volume of new gas-fired plants, renewable projects, and battery installations. But interconnection queues are notorious for containing far more capacity on paper than ever gets built. Without clarity on which projects have secured financing, permits, and realistic construction timelines, matching specific supply resources to the projected demand surge remains guesswork.
What Texas ratepayers should watch
For the roughly 27 million Texans who depend on the ERCOT grid, the stakes are immediate and personal. Tighter reserve margins translate into higher wholesale electricity prices, which flow through to retail bills. During the handful of hours each summer when demand peaks and reserves shrink, scarcity pricing can spike wholesale costs above $5,000 per megawatt-hour. That mechanism is designed to incentivize new generation, but it also punishes consumers in real time.
Data centers bring jobs, tax revenue, and technology investment. They also consume enormous amounts of power around the clock, with little flexibility to cut back during emergencies without disrupting services that businesses and governments treat as essential. That combination makes them both attractive economic anchors and challenging grid neighbors.
The tension is real, but it is not yet a guaranteed crisis. FERC’s assessment signals that ERCOT’s existing planning tools may be out of step with the speed and scale of new demand. State-level debates highlight the financial and political costs of getting the forecasts wrong. Until more granular data on load types, locations, and corporate energy commitments becomes public, claims about exactly how much of the problem stems from data centers versus other industries should be treated as provisional.
A shrinking window for grid planning corrections
What is not provisional is the shrinking margin for error. Texas built its electricity market on the premise that price signals and competition would attract enough generation to meet demand. That model has never been tested against load growth this fast. The window for course correction is closing. Every month of delayed planning is a month closer to a summer when the grid may not have enough power to go around.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
