Gas Price Spikes Threaten AI Plans at Meta and Microsoft

A new forecast from energy research firm Noreva warns that tech giants investing heavily in natural gas to power their artificial intelligence infrastructure could face severe financial shocks. Hyperscalers have recently pivoted toward fossil fuels to meet the massive energy demands of AI. Meta announced plans for a 7.5-gigawatt natural gas plant in Louisiana for its Hyperion data center, while Amazon is planning a 7.6-gigawatt gas plant in Texas. Microsoft and Google have also committed to building their own gigawatt-scale gas plants in Texas.

However, Noreva projects that natural gas prices could spike above $10 per million BTUs at certain delivery hubs, a massive leap from current rates that hover between $2 and $4.50. The widely traded Henry Hub in Louisiana currently sits just under $3. This potential tripling of fuel costs is driven by a combination of surging AI energy demand, slowing domestic supply growth, and newly constructed pipelines that connect previously isolated cheap gas markets, like West Texas, to lucrative global export markets.

For AI practitioners and developers, this looming energy crunch could directly impact the bottom line. Because fuel accounts for roughly half of a large power plant's electricity generation costs, soaring gas prices will make these self-powered data centers far more expensive to operate. To recoup these expenses, hyperscalers may be forced to raise token prices for AI model APIs. Alternatively, if these companies pivot back to the public grid to avoid high fuel costs, they could drive up regional electricity rates, triggering broader public backlash.

This shift forces tech companies into unfamiliar territory as major players in volatile energy markets. Noreva chief executive Peter Gardett noted that tech firms are taking on unusual risks for energy buyers. If these projections hold true, future corporate earnings reports may soon tie AI profit margins directly to the fluctuating price of fossil fuels.