Hyperscalers Might Reevaluate Natural Gas Consumption in Light of Revised Forecasts

In light of significant investments in renewable energy, leading tech companies such as Amazon, Google, Meta, and Microsoft are increasingly turning to natural gas to power their data centers that facilitate large-scale AI operations. Nonetheless, a recent research study has raised concerns about the potential risks linked to this rising dependency on fossil fuels by these tech giants.

Noreva, a research firm focused on energy, has issued warnings that natural gas prices in several regions across the U.S. could potentially triple in the near term. This anticipated price increase is primarily driven by heightened demand from hyperscalers, limited growth in supply, and increasing exports of liquefied natural gas (LNG), which could make hyperscalers vulnerable to sudden price fluctuations.

“Many in the energy sector have underestimated future gas prices,” remarked Peter Gardett, CEO of Noreva, during an interview with TechCrunch. “A simple calculation indicates that the gas market is much tighter than it was a few years back.”

Significant Investments by Hyperscalers

The affordability of natural gas is encouraging hyperscalers to fortify their presence in the market. In March, Meta disclosed plans to build a 7.5-gigawatt natural gas facility in Louisiana to support its Hyperion data center. Shortly after, Microsoft and Google announced their intentions to construct gigawatt-scale gas plants in Texas. Moreover, Amazon revealed plans for a 7.6-gigawatt gas plant also in Texas.

These typically cautious companies, known for their prudent capital expenditures, are now being pushed to make substantial investments in physical infrastructure while navigating complicated energy markets.

According to Gardett, at least one investor expressed surprise at the level of risk hyperscalers are prepared to take regarding natural gas prices. “Their behavior is quite atypical for an off-taker,” he noted.

Noreva forecasts that natural gas prices could surpass $10 per million BTUs at certain delivery points, a sharp contrast to current prices, which fluctuate between $2 and $4.50 per million BTUs, with the widely recognized Henry Hub in Louisiana priced just below $3.

Given that fuel expenses constitute nearly half of the electricity costs for large power plants, a doubling or tripling of natural gas prices could substantially raise operational costs for AI data centers that “produce their own power.” This scenario may lead to increased token prices or force hyperscalers to rely more on the grid, further amplifying their electricity expenses.

In the immediate future, natural gas prices seem stable, with futures contracts not indicating substantial volatility. “It’s a reasonable assumption,” Gardett stated, while expressing skepticism about its long-term stability.

Increasing Demand

The stabilization of natural gas prices can be attributed to a prolonged period of steady demand and limited supply growth, offsetting declines from aging wells, according to Gardett’s analysis. He asserts that energy companies can increase supply, though not to previous levels, and the drilling of new wells has become more costly.

“This alone wouldn’t significantly alter the economics. The pivotal factor is that we are now connecting the domestic gas market with the global market,” he emphasized. “Furthermore, there’s an upsurge in demand fueled by AI.”

Hyperscalers are attracted to Texas and Louisiana due to their lower natural gas prices. For instance, many gas wells in West Texas are primarily focused on oil extraction, with natural gas being a byproduct facing limited market opportunities. Historically, inadequate pipeline infrastructure has slowed the transportation of this gas, compelling producers to sell it at lower rates. However, substantial infrastructure enhancements are currently underway.

“Recently constructed pipelines are mainly intended for export markets,” Gardett explained.

As West Texas continues to strengthen its connections with regional and global natural gas markets, local demand will increasingly impact prices in different areas, and vice versa. Even minor price changes near hyperscaler data centers could significantly influence costs in various regions.

“Certain regions will have an abundance of gas, while others will face scarcity, resulting in significant price variations,” Gardett highlighted. Such disparities could drive prices in specific locations above $10 per million BTUs for extended durations.

In this evolving landscape, even if hyperscalers adjust to rising costs, their escalating consumption of natural gas might incite growing public discontent towards data centers. Currently, 80% of consumers are concerned about the impact of data centers on their utility bills, especially with regard to electricity costs. This concern will likely extend to natural gas pricing as well.

By rapidly electrifying their AI data centers, hyperscalers are becoming more entwined with the fossil fuel industry—an area where they possess relatively little expertise, yet one that could significantly affect their operations.

“In future earnings calls for Alphabet, discussions surrounding the connection between natural gas prices and Google’s performance may arise, which is unusual but reflects the current reality we are facing,” Gardett concluded.

Please be aware that clicking on links in our articles may lead to us earning a small commission. This does not affect our editorial independence.