Major technology corporations, including Amazon, Google, Meta, and Microsoft, are increasingly turning to natural gas as a primary energy source for their expanding artificial intelligence data centers. This strategic pivot marks a departure from their historical focus on renewable energy such as wind and solar. However, a recent analysis casts doubt on the long-term viability of this approach, predicting a potential tripling of natural gas prices in specific areas of the United States. Such a surge in energy costs could severely impact the financial models of these tech giants, which are investing heavily in AI infrastructure, and potentially lead to widespread repercussions across the energy market.
The shift towards natural gas for AI data centers is a significant development, especially given these companies' prior commitments to renewable energy. This strategic move is largely influenced by the perceived affordability and availability of natural gas, making it an attractive option for powering energy-intensive AI operations. For example, Meta announced plans to construct a 7.5-gigawatt natural gas power plant in Louisiana to support its Hyperion data center. Similarly, Microsoft and Google are pursuing gigawatt-scale natural gas power plants in Texas, while Amazon intends to build a 7.6-gigawatt facility in the same state. These substantial investments highlight the tech sector's growing reliance on fossil fuels to meet the escalating energy demands of AI.
Despite the current stability of natural gas prices, experts like Peter Gardett, CEO of Noreva, an energy research firm, caution against complacency. Gardett points out that the energy market's prevailing sentiment that gas prices cannot rise indefinitely is a misconception. He argues that a straightforward analysis of supply and demand dynamics indicates a much tighter gas market in the near future compared to previous years. This tightening is attributed to several factors: a decrease in the rate of new supply growth, increasing exports of liquefied natural gas, and the surging demand from hyperscalers. Consequently, Noreva projects that natural gas prices could exceed $10 per million BTUs in certain key hubs, a significant increase from current prices ranging between $2 and $4.50, with the Henry Hub in Louisiana currently below $3.
The implications of such price increases are profound. Fuel typically constitutes about half the operational cost of large power plants. Therefore, a doubling or tripling of natural gas prices could render the "bring your own power" model for AI data centers considerably more expensive. This could, in turn, drive up the cost of AI services or compel hyperscalers to connect to existing power grids, thereby escalating overall electricity prices. Historically, these technology companies have been hesitant to undertake massive capital expenditures. However, the current boom in data center construction is pushing them into unfamiliar territory within the energy markets. Gardett notes that the willingness of hyperscalers to assume such significant natural gas price risks has surprised even seasoned investors, as their actions deviate from typical consumer behavior in the energy sector.
The appeal of regions like Texas and Louisiana, with their historically low natural gas prices, has drawn hyperscalers to establish their data centers there. In West Texas, for instance, natural gas has often been an unmarketed byproduct of oil extraction, leading to discounted prices due to limited pipeline infrastructure. However, this situation is rapidly evolving as new pipelines are being constructed, primarily to facilitate exports. This enhanced connectivity will integrate West Texas more deeply into national and international natural gas markets, meaning local price fluctuations will increasingly affect broader markets. Gardett predicts that significant price differentials will emerge, pushing prices above $10 per million BTUs for extended periods in certain areas.
Such a scenario could exacerbate public concerns about data centers' energy consumption. Consumers are already worried about the impact of data centers on their electricity bills. A substantial increase in natural gas prices could extend this anxiety to natural gas expenses, creating a new wave of public backlash. As hyperscalers rapidly embed themselves in the fossil fuel industry to meet AI's power demands, they venture into an area where they possess limited expertise. This lack of experience could soon have a material impact on their business operations and financial performance. Gardett starkly illustrates this potential shift by suggesting that future earnings calls for companies like Alphabet might include discussions on the correlation between natural gas prices and Google's search results, reflecting the deep entanglement of technology and energy markets.
