Hyperscalers like Amazon, Google, Meta, and Microsoft have increasingly turned to natural gas to power their AI data centers. However, a new report from Noreva, an energy research firm, warns that rising demand and declining supply growth could lead to a tripling of natural gas prices in some U.S. regions. This could significantly increase the cost of running AI data centers, potentially affecting token costs and pushing hyperscalers to rely more on the grid, which could raise electricity prices. "I think everyone in the energy markets has been lulled into a sense that gas prices can’t go up," said Peter Gardett, CEO of Noreva. "You just need simple arithmetic to get to a much tighter gas market than you were in just a few years ago."
Hyperscalers have made substantial investments in natural gas infrastructure, including Meta’s 7.5-gigawatt plant in Louisiana, Microsoft and Google’s gigawatt-scale plants in Texas, and Amazon’s 7.6-gigawatt facility in Texas. These projects reflect a shift for companies historically reluctant to invest in large capital expenditures. Gardett noted that some investors were surprised by the level of natural gas price risk hyperscalers are willing to take. "They’re doing things that are not normal for an off-taker to do," he said. Noreva expects natural gas prices to surpass $10 per million BTUs in certain hubs, a significant jump from current prices of $2 to $4.50 per million BTUs.
Natural gas prices have remained stable due to balanced demand and supply, but Gardett warned that this equilibrium is changing. New wells are becoming more expensive, and the domestic gas market is now increasingly linked to global markets. This connection, combined with rising AI demand, could lead to significant price fluctuations. "You will get places where you get a lot of gas next to someplace where there’s none, and so you’ll get those big differentials," Gardett said. These differentials could drive prices above $10 per million BTUs for extended periods, potentially impacting hyperscalers’ operations and adding to the backlash against data centers.
Source: techcrunch
