Listen again · In Conversation
Powering the Data Center Boom October 10, 2026 · 10:29 · With Maya, Alex, and Emma
Alex joins us from the reporting desk, and Emma brings the energy perspective. Thanks for listening. We’re looking at what rising data-center demand could mean for the choice between batteries and natural-gas power.
Transcript Tap a sentence to hear it from there.
Maya Welcome to Slow English FM. This is our round table, where we look at one issue from more than one angle. I’m Maya. Alex joins us from the reporting desk, and Emma brings the energy perspective. Thanks for listening. We’re looking at what rising data-center demand could mean for the choice between batteries and natural-gas power. Maya The immediate news is a striking cost comparison. TechCrunch reports that a Wood Mackenzie study found four-hour battery systems cost less than open-cycle natural-gas turbines in all 43 markets it surveyed. That may change how some power projects are considered as demand grows. But the comparison is not proof that batteries can replace every power plant, or meet every need. Alex, what exactly is being compared? Alex The comparison is between two specific options. A four-hour battery is a storage system rated to deliver its stored electricity over four hours. An open-cycle gas turbine is a type of natural-gas power plant. TechCrunch says the battery systems came out cheaper across the 43 surveyed markets. That is useful evidence about costs. It does not tell us that every battery, every gas plant, or every project has the same price. Maya So the result is broad, but it still has a boundary: those two technologies, in those markets. Emma, why is the gas-turbine comparison relevant to data centers? Emma Because the demand for electricity is growing, and gas turbines have often been used to help meet peaks. A peak is a time when demand is especially high. TechCrunch reports that open-cycle turbines are often used as peaking power plants. It also says orders for turbines now face long waits, and that backlogs have pushed up prices for new gas power plants. My interpretation is that a project planner facing both higher prices and a long wait may take another look at storage. But that does not establish what any particular data center will choose. Maya Alex, can we see the demand pressure in the U.S. figures? And let’s be careful to separate forecasts from what has already happened. Alex Yes. The U.S. Energy Information Administration, or EIA, forecast that commercial electricity demand would grow 2.8 percent year over year through 2027. Commercial demand includes data centers. That is a forecast, not a measured result. In a separate September outlook, the EIA forecast U.S. electricity sales of 4,135 billion kilowatt-hours in 2026 and 4,211 billion in 2027. It linked expected growth in commercial and industrial electricity use to data-center development and increased manufacturing activity. Maya Those are national figures. Does the pressure look the same everywhere in the country? Alex Not according to that September forecast. The EIA said the West South Central region, which includes Texas, would account for nearly 20 percent of U.S. electricity-sales growth in 2026 and nearly 40 percent in 2027. Those are shares of forecast growth, not shares of all U.S. electricity use. The EIA also noted that Texas had paused new data-center grid connections for a regulatory review. That gives us two pieces of context: expected demand growth is concentrated in part of the country, and Texas had paused new connections while a review was under way. Emma And that is where I would be cautious about treating a national cost comparison as a complete answer. A lower reported cost for one type of battery does not, by itself, show how a particular region should meet its demand. The regional forecast and the pause in Texas connections point to a local question about grid connections and rising sales. The evidence here does not settle that question. It does show why the choice of new power equipment matters to a discussion about growth. Maya Emma, Alex has described where demand is expected to grow. How does solar fit into the cost picture? Emma TechCrunch reports that solar was the cheapest form of new power in every market in the Wood Mackenzie survey. It also notes that U.S. tariffs and import restrictions are putting pressure on solar prices. That makes the comparison more interesting: the reported battery-versus-turbine result is not the same as saying batteries are the cheapest source of new electricity. Solar was cheapest in that survey. My reading is that planners are weighing different tools, and that changes in equipment prices or trade rules may affect the comparison. But the supplied figures do not tell us how a specific project would combine them. Alex I agree that the categories matter. The reported solar finding concerns the cheapest form of new power. The battery finding compares four-hour storage with open-cycle turbines. Those are not identical comparisons. A battery stores electricity for later delivery; a turbine generates electricity by burning natural gas. The result does not tell us that one option can do everything the other can do, or that the lowest reported cost alone decides a project. Maya That distinction is important. One question is what new power costs. Another is what equipment can provide, and when. Emma, what can we responsibly say about the wider time scale of data-center demand? Emma The EIA’s longer-term outlook gives a range for data-center server electricity use in 2050: from 446 to 818 billion kilowatt-hours across its cases. Those are projections, not current consumption figures, and the range shows that the outlook depends on which case is considered. We should not turn the higher end into a prediction that is certain to happen. Still, it illustrates why long-term planning is part of the discussion, alongside near-term questions about turbine waits and battery costs. Maya Alex, is this demand surge entirely new? Alex No. The EIA estimates that U.S. electricity demand grew about 1.7 percent a year from 2020 to 2025. It estimates growth of 0.1 percent a year from 2005 to 2019. The agency identifies data-center electricity use as one driver of the more recent growth. Those figures describe different periods. They help show a change in the pace of demand, but they do not mean data centers are the only cause. Emma And that last point matters when we talk about consequences. The EIA also connects expected commercial and industrial growth with manufacturing activity. So it would be too simple to describe every increase in electricity demand as coming from data centers. The data-center boom is part of the story, but the outlook names another source of growth too. Maya Let me bring the two angles together. Alex, we have national and regional demand forecasts, plus historical growth. Emma, we have the cost comparison, turbine backlogs, and solar’s position. What is the most careful conclusion for someone following this debate? Alex The careful conclusion is that the demand question is real in the forecasts, but forecasts are not results. The EIA expects U.S. electricity use to rise, and it points to data centers among the reasons for recent demand growth and expected commercial use. It also identifies a region that would account for a substantial share of forecast sales growth. Those details show why power planning is receiving attention. They do not tell us which technology every area will build. Emma And on the technology choice, the cost evidence challenges the assumption that new gas turbines are always the cheaper option. In the Wood Mackenzie survey reported by TechCrunch, four-hour batteries cost less than open-cycle gas turbines in all 43 markets. But solar was the cheapest form of new power in those markets, and the turbine comparison is limited to specific technologies. I would read this as a reason to compare options carefully, not as a verdict that batteries will replace gas plants. Maya So the story is not simply batteries versus gas. Rising demand creates pressure to plan, while prices, turbine waits, and the availability of different technologies shape the choices. The evidence points to a broad cost shift in the surveyed markets, but it does not answer every local power question. Alex brought us the demand forecasts and regional picture. Emma explained what the technology comparison does—and does not—show. Thanks to both of you, and thanks for listening to Slow English FM.
