The AI industry is turning to an old energy favorite to bring its data centers online fast

As artificial intelligence companies scramble to secure energy for their expanding data centers, natural gas is experiencing a resurgence. The demand for larger and more power-intensive data centers has surged, leading to tighter construction timelines that the U.S. power infrastructure struggles to accommodate. Vivian Lee, a managing director at Boston Consulting Group, emphasized the importance of speed in this race for energy. "Bringing a site online even a year earlier can have a meaningful economic impact," she stated. Typically, establishing a data center takes two to three years, contingent on community support. However, if local approval is lacking, the timeline can extend significantly. In contrast, grid upgrades may require four to eight years, prompting AI firms to seek quicker alternatives for their energy needs. One of the fastest solutions currently available is leveraging existing natural gas infrastructure. Natural gas plants can be constructed or expanded more rapidly than nuclear facilities, connect seamlessly to an established pipeline network, and offer a more reliable energy source compared to renewables. For instance, Meta is expanding its Hyperion data center in Louisiana by adding seven natural gas plants. Additionally, Chevron and Engine No. 1, which formed a partnership last year, have welcomed Microsoft to collaborate on powering a data center campus in West Texas through natural gas. Google is also partnering with Crusoe Energy to establish natural gas plants for its "Goodnight" data center in the Texas Panhandle. Jamie Webster, a senior director at BCG, noted that the primary metric driving this shift is the swift availability of power. "That's why gas is back in focus," he remarked. Although natural gas emits less carbon dioxide than coal or oil, it remains a fossil fuel contributing to climate change, making its adoption by Silicon Valley particularly noteworthy. Historically, American tech giants have positioned themselves as pioneers in the renewable energy sector. Companies like Google, Amazon, Microsoft, and Meta entered into substantial wind and solar agreements to meet the rising electricity demands of their data centers. However, these decisions may have been influenced more by financial considerations than by environmental principles. Webster explained that the focus on renewables was not solely driven by sustainability but also by their potential to lower energy costs over time. Over the past decade, advancements in solar, wind, and battery technologies have led to dramatic cost reductions, fundamentally altering the energy landscape. Nevertheless, the current economic climate is shifting back towards natural gas as AI companies secure significant funding for infrastructure development while still facing challenges in revenue generation. Lee noted that while renewables are deemed "essential," they are often viewed as inadequate on their own. To mitigate the environmental consequences of increased natural gas usage, carbon capture technology could play a crucial role. This process captures carbon dioxide emissions from power plants and either stores them underground or repurposes them, potentially allowing for continued natural gas use while reducing its impact on climate change. However, Webster cautioned that carbon capture technology is still in the nascent stages of widespread implementation. As the world navigates a "structural supercycle" driven by the demands of data centers and electrification, Webster highlighted the growing pressure on energy supply, with natural gas often emerging as a rapid response to meet this pressing need.