Natural gas, long the overlooked sibling of crude oil in the American energy story, is stepping into a starring role as the world’s demand for artificial intelligence (AI) and digital services ignites a new era of power consumption. The surge is most visible in the heart of the Appalachian Basin—specifically the Marcellus and Utica shale plays spanning Pennsylvania, West Virginia, and Ohio—where gas production is booming, and local economies are being reshaped by the rapid buildout of data centers and liquefied natural gas (LNG) infrastructure.
According to Fortune, this boom is driven by three converging forces: the explosive rise of data centers powering generative AI, record-high LNG exports, and the ongoing retirement of aging coal-fired power plants. As a result, top gas producers like Expand Energy, EQT, Range Resources, and Antero Resources—all with deep Appalachian footprints—have seen their market values soar between 25% and 75% over the past year. Meanwhile, oil-focused companies are lagging, grappling with stagnant prices and weaker demand growth.
“With the resource-rich potential in this [Marcellus] basin and the growing demand component for AI and data centers and power, it really is setting us up well to help shape this AI revolution that’s going to take place here in the United States,” said Dennis Degner, CEO and President of Range Resources, in remarks to Fortune. He described how the industry’s focus has shifted from seasonal weather patterns to the massive, year-round power demands of AI and data infrastructure, along with the doubling of LNG exports in the coming years. “Those are all big, diverse demand components that really get us excited about the durability of our business model.”
The numbers are staggering. The Appalachian region now produces just over a third of the nation’s natural gas, but barely any oil. Its proximity to Virginia’s so-called Data Center Alley—and an increasing number of AI infrastructure projects within Appalachia itself—has made it a strategic hub for the new digital economy. The International Energy Agency projects U.S. electricity consumption will jump by 25% from 2023 to 2035, and by 60% by 2050, driven largely by AI and data centers. At the same time, U.S. LNG exports are on track to double by 2030, reaching at least 30 billion cubic feet per day.
Paul Rady, Antero’s chairman and CEO, summed up the industry’s expectations in his recent earnings statement: “Natural gas demand [is] set to soar 25% by 2030, led by LNG growth and then by data center power thirst.” That’s a remarkable leap for a sector that already produces 107 billion cubic feet of gas per day—double the output from just two decades ago.
The top gas producers are not only exceeding their production estimates this year but are doing so with remarkable efficiency. Range Resources, for example, aims to grow production by 20% by the end of 2027 while running just two drilling rigs. For context, Exxon Mobil, the oil giant, operates at least 35 rigs in the Permian Basin. As Gabriele Sorbara, energy analyst at Siebert Williams Shank & Co., put it: “These [Marcellus] wells are just massive. The fundamentals for gas are very strong. You’re going to have massive tailwinds.”
Yet, even as optimism runs high, there’s caution in the air. Producers remain conservative about ramping up output, wary of overbuilding pipelines or locking in fixed-price deals with data center developers. Milder weather and rising gas storage levels have recently caused a dip in prices and stock values, but the bullish outlook persists. Range, for instance, sends about half its Pennsylvania gas to the Gulf Coast for LNG exports, but sees most new growth coming from regional data center demand due to pipeline limitations.
July 2025 saw a major political boost for the region when former President Trump touted $92 billion in energy and AI investments in Pennsylvania, involving hyperscalers, power generators, and industrial developers. Range Resources recently partnered with Imperial Land to supply new gas-fired power generation for data centers, while Pennsylvania’s Homer City complex is being converted from a 1.9-gigawatt coal plant into what will soon be the largest gas-fired power plant in the country—capable of delivering up to 4.5 gigawatts to a sprawling data center campus. EQT, the largest Marcellus gas producer, has also inked deals to supply both Homer City and the planned Shippingport Power Station, both of which are transitioning from coal to gas.
This transformation is not just about numbers and deals—it’s about the physical and environmental footprint of the digital revolution. As Paris Marx, writing for Bioneers, points out, “AI doesn’t run on magic—it runs on energy, water, and massive physical infrastructure.” The rapid rise of generative AI, kicked off by the November 2022 launch of ChatGPT, has been fueled by centralized computing power, massive data collection, and colossal capital investment. Amazon began building its first massive cloud warehouses in 2006, and by the end of 2024, the number of hyper-scale data centers worldwide had nearly doubled to 1,136, with another 500-plus in the pipeline—about half of them in the U.S.
These hyper-scale data centers are voracious consumers of resources. Google’s facilities, for example, use an average of 550,000 gallons of water per day, primarily for cooling, and the company’s total water use at its data centers jumped 20% between 2022 and 2023. Microsoft’s use rose by 34% in the same period. Globally, data centers now account for between 2% and 3% of all energy consumption, a figure that rises to around 5% in the U.S. alone. According to the International Energy Agency, data centers, crypto, and AI infrastructure consumed 460 terawatt-hours of electricity in 2022—roughly the same as France—and are projected to use more than double that by 2026, approaching Japan’s total annual energy use.
This insatiable appetite for power is putting pressure on local grids and ecosystems. Communities near clusters of new data centers are increasingly pushing back, citing strains on electricity and water supplies. In Ireland, for instance, 21% of all metered electricity now goes to data centers, leading to wintertime grid alerts and calls for a moratorium on new facilities.
Despite these concerns, the buildout continues. The gas industry, emboldened by strong fundamentals and vast reserves, is confident it can sustain this growth for decades—so long as it avoids the classic boom-and-bust cycle. “We can do this for decades to come, and now you’re talking about a [data center] demand component that’s coming that’s heavily dependent on reliability, repeatability, and the [gas] inventory,” Range’s Degner told Fortune. “Of course, Range thrives on all three.”
Expand Energy, now the nation’s largest natural gas producer after the merger of Chesapeake Energy and Southwestern Energy just ten months ago, has a major presence in both Appalachia and Louisiana’s Haynesville Shale. “It’s a pretty exciting time for natural gas,” said Expand CEO Nick Dell’Osso in a recent earnings call. “You have people recognizing the value that gas plays in the economy, the efficiency that gas creates for the growth in power demand, which is all tied to our growing economy fueled by the innovation associated with AI.”
Yet, as the world races to power the AI revolution, critics urge a broader conversation about priorities and sustainability. Paris Marx writes, “How much computation do we actually need? Do we really need to build out endless data centers to support a flood of AI tools with questionable uses—tools that often serve tech companies’ bottom lines more than the public good?” The environmental costs and societal trade-offs of this new energy-data nexus are only beginning to surface, and the stakes could not be higher.
As Appalachia’s gas fields fuel the digital future, the region stands at the crossroads of prosperity and responsibility—a testament to how the world’s hunger for computation is fundamentally reshaping both the American landscape and the global energy equation.
