Hyperscalers' natural gas bets could backfire as prices triple

After years of investing in renewable energy sources, Amazon, Google, Meta, and Microsoft have pivoted sharply toward natural gas to power their massive AI data centers. Meta announced in March that it would build a 7.5-gigawatt natural gas power plant in Louisiana to supply its Hyperion data center. Within days, Microsoft and Google each disclosed plans to build gigawatt-scale natural gas plants in Texas, and Amazon followed with plans for a 7.6-gigawatt facility in the same state. The shift reflects the companies' calculation that cheap natural gas prices made on-site power generation economical.

But Noreva, an energy research firm, has published a forecast suggesting this bet may prove costly. The firm predicts that natural gas prices could triple in some U.S. regions over the coming years, rising above $10 per million BTUs—more than double the current range of $2 to $4.50 per million BTUs (the widely traded Henry Hub in Louisiana sits just under $3). Peter Gardett, Noreva's CEO, told TechCrunch that the energy markets have been "lulled into a sense that gas prices can't go up," but simple arithmetic reveals a much tighter market ahead.

The root causes are structural. For years, natural gas prices stayed stable because demand was relatively flat and new supply additions offset declining production at aging wells. That equilibrium is breaking. Gardett expects that energy companies will be able to add more supplies, but not at the pace of the past—and new wells are becoming more expensive to drill. Simultaneously, the domestic U.S. gas market is now connecting to the global market as liquefied natural gas exports rise. West Texas, where hyperscalers are concentrating their operations, is particularly exposed: historically, natural gas there was a byproduct of oil drilling, sold at local discounts because few pipelines existed to move it. New pipelines are now routing West Texas gas to export terminals, linking regional prices to global markets. As Gardett explained, that connection means even modest price swings near hyperscaler data centers will ripple elsewhere, creating regional differentials that could push some hubs above $10 per million BTUs for extended periods.

The financial implications are substantial. Fuel represents about half the cost of electricity from a large power plant, so doubling or tripling natural gas prices could dramatically increase the operating cost of these "bring your own power" data centers. That could raise token costs (the cost of AI computation), or it could force hyperscalers to abandon on-site generation and draw from the grid, driving broader electricity prices higher. At least one investor Gardett spoke with was "surprised" by how much natural gas price risk hyperscalers are willing to absorb. For now, futures markets do not anticipate big price moves, suggesting the bet remains "not an unreasonable" one—but Gardett is unconvinced it will pay off. Beyond the financial risk, hyperscalers face reputational exposure: 80% of consumers are already worried about data centers' impact on utility bills, mostly related to electricity. Higher natural gas bills could amplify that backlash. As Gardett wryly noted, the outcome means that "on future Alphabet earning calls, you will hear them talk about the correlation between natural gas pricing and Google results, which is strange, but that's where we are."