The Natural Gas Gamble: Hyperscalers’ Folly in a Carbon-Conscious World

Hyperscalers like Amazon, Google, and Microsoft have recently announced plans to build massive natural gas power plants to fuel their AI ambitions. However, a new forecast from Noreva suggests that the trend of stable natural gas prices may be short-lived.

According to Peter Gardett, CEO of Noreva, simple arithmetic reveals a tightening gas market. Declining supply growth and rising exports of liquefied natural gas (LNG) could lead to a tripling of prices in certain regions. As of writing, natural gas prices range from $2 to $4.50 per million BTUs. If Noreva’s forecast proves correct, that number could soar above $10.

The consequences for hyperscalers are far-reaching. Natural gas accounts for approximately half the cost of electricity from a large power plant. With prices potentially doubling or tripling, the economics of “bring your own power” AI data centers would become increasingly unattractive. This could lead to higher token costs or push hyperscalers to connect to the grid, driving up electricity prices for all consumers.

The narrative surrounding natural gas is complex and multifaceted. On one hand, it has been touted as a cleaner alternative to coal, offering a bridge to a low-carbon future. However, the environmental implications of increased production and consumption cannot be ignored. Gardett notes that the connection between domestic gas markets and global markets will have far-reaching effects on prices.

Hyperscalers’ reliance on natural gas is striking given their historically cautious approach to capital expenditures. The sudden shift towards large-scale data center building has led them deeper into energy markets, an unfamiliar territory for most of these companies. This gamble raises questions about their ability to navigate the nuances of the fossil fuel world.

The consequences of this miscalculation are not limited to hyperscalers alone. As natural gas prices surge in certain regions, consumers will feel the pinch on their utility bills. Eighty percent of consumers already express concern about data centers’ impact on electricity costs; they may soon find themselves worrying about natural gas bills as well.

Gardett wryly observes that “On future Alphabet earning calls, you will hear them talk about the correlation between natural gas pricing and Google results.” This quip highlights the absurdity of the situation: hyperscalers are now unwitting participants in the fossil fuel world, with their very businesses tied to the whims of a volatile market.

Hyperscalers must reassess their reliance on natural gas. They would do well to remember that their AI ambitions come with significant environmental implications. The time has come for these companies to think critically about their energy strategies and consider more sustainable alternatives.

The future of data centers will be shaped by the choices made today. As hyperscalers navigate this treacherous landscape, they will not emerge unscathed from this gamble on natural gas.