AI companies are building huge natural gas plants to power data centers. What could go wrong?

The tech industry is experiencing a significant wave of investment in natural gas power plants, driven by the surging demands of AI data centers. Companies like Microsoft, Google, and Meta are racing to secure energy sources to support their growing operations, with Microsoft partnering with Chevron to potentially generate 5 gigawatts at a new facility in West Texas. Meanwhile, Google is collaborating with Crusoe to construct a 933 megawatt plant in North Texas, and Meta is expanding its Hyperion data center in Louisiana with seven additional natural gas plants, pushing its capacity to a staggering 7.46 gigawatts. This frantic push towards natural gas is unfolding primarily in the southern U.S., which boasts some of the world's largest natural gas reserves. Recent estimates from the U.S. Geological Survey suggest that one area alone could power the entire country for ten months. However, this rush has sparked a shortage of essential turbines for the plants, leading to projected price increases of up to 195% compared to 2019 levels, according to Wood Mackenzie. The consultancy warns that new turbine orders won't be fulfilled until 2028, with delivery times extending up to six years. Tech companies are banking on the assumption that the demand for AI will remain robust, necessitating vast amounts of power from natural gas sources. Yet, this strategy may carry significant risks. Although the U.S. has ample natural gas supplies, production growth in key shale regions has recently slowed, raising concerns about long-term availability. The relationship between natural gas and electricity prices is also complex; with natural gas accounting for about 40% of U.S. electricity generation, fluctuations in gas prices could ultimately affect tech companies' operational costs. Moreover, while some tech giants might try to mitigate scrutiny by connecting their gas plants directly to their data centers—bypassing the grid—the reality is that their consumption could still strain overall energy resources. This situation raises critical questions about the sustainability of relying on a finite resource like natural gas, especially as other industries that rely heavily on this fuel may find themselves at a disadvantage. The potential for supply disruptions, particularly during extreme weather events, poses further challenges. A severe winter could spike household demand for heating, forcing suppliers to choose between powering AI centers or ensuring residential homes stay warm. As the AI sector continues to expand, the implications of this natural gas scramble could ripple through the economy, affecting not just tech companies but also the broader landscape that relies on these energy resources. With the stakes so high, the tech industry must carefully consider whether their aggressive investments in natural gas are prudent in the long run—or if they are merely falling prey to the fear of missing out on the AI revolution.