North Milford Valley, in western Utah, is home to dormant volcanoes, subterranean lava deposits, and smatterings of obsidian—black volcanic glass—that Paiute peoples once collected for arrowheads and jewelry. Scalding groundwater still bubbles to the surface in places. In such a landscape, you remember that the planet’s hard exterior, where we spend our entire lives, is so thin that we call it a crust. Its superheated interior, meanwhile, burns with an estimated forty-four trillion watts of power. Milford was once a lead-, silver-, and gold-mining town, but when I visited the area on a sunny spring morning a scientist named Joseph Moore was prospecting for something else: heat.
Heat mined from underground is called geothermal—“earth heat,” in ancient Greek—and can be used to produce steam, spin a turbine, and generate electricity. Until recently, humans have tended to harvest small quantities in the rare places where it surfaces, such as hot springs. Moore’s mission, as a geologist at the University of Utah and the project leader of the Frontier Observatory for Research in Geothermal Energy (FORGE), is to “develop the roadmap that is needed to build geothermal reservoirs anywhere in the world.” This road is long, and much of the map remains blank. The biggest problem is drilling miles through hot rock, safely. If scientists can do that, however, next-generation geothermal power could supply clean energy for eons.
During my trip, Moore’s corps of consultants and roughnecks were drilling the fifth borehole of their experimental project. Their rig, armed with a diamond drill bit, towered like a rocket over the rural landscape; miles of solar panels and wind turbines receded into the distance. The hole, which would eventually be L-shaped, was five thousand feet deep, and the team had another five thousand to go, horizontally. But, before they could drill any farther, they needed to install a hundred-and-fifty-ton steel tube in the hole, using special heat-resistant cement to glue it into place. The tube was like a massive straw that was meant to transport hot water and steam from an artificial underground reservoir—without contaminating local groundwater or triggering earthquakes.
At 6:15 P.M. on May 3rd, cement had started flowing into the hole. Four hours later, part of the cement folded in on itself. The next morning, the cement supply ran out; the men had miscalculated how much they needed. This brought the three-hundred-million-dollar operation to a maddening halt. Moore, in bluejeans and a FORGE-branded hard hat, called his supplier. The nearest batch of suitable cement was five hundred miles away, in Bakersfield, California. The truck would not arrive until after dark.
Right now, geothermal energy meets less than one per cent of humanity’s electricity and heating needs—a puny, almost irrelevant portion. Fossil fuels power about eighty per cent of human activity, pumping out carbon dioxide and short-circuiting our climate to catastrophic effect. Converts argue that geothermal checks three key boxes: it is carbon-free, available everywhere, and effectively unlimited. Crucially, it is also baseload, which means that, unlike solar panels or wind, it provides a constant flow of energy. Companies and governments have taken notice. “Over the last two years, I have watched this exponential spin-up of activity in geothermal,” Tony Pink, a drilling expert in Houston, told me, in 2023.
But there is a glaring risk of moon shots: often, they miss. “There’s basically zero chance that you’re going to develop a moon-shot technology and have it be commercial in five years, on a large-scale, worldwide,” Mark Jacobson, a Stanford engineering professor and the author of “No Miracles Needed: How Today’s Technology Can Save Our Climate and Clean Our Air,” told me. That’s how long humanity has to lower emissions before climatic devastation, according to his calculations. “There’s a very decent chance you can do that with wind and solar,” he said. Perhaps, when resources and time are finite, trying and failing—or simply taking too long—could be worse than not trying at all.
In 1890, two brothers began drilling for hot water in Boise, Idaho; their company soon built the world’s first geothermal district heating system. In 1904, an Italian entrepreneur named Piero Conti constructed a geothermal turbine in Tuscany. It powered only five light bulbs, but Conti went on to build Larderello 1, the first-ever geothermal plant and the power source for a railway and two nearby villages. These early experiments illustrated one of geothermal’s key advantages: it can be used for both electricity and heating, which collectively account for around thirty-eight per cent of global climate emissions.
In the nineteen-sixties and seventies, Pacific Gas and Electric, a California utility company, built an industrial-scale geothermal electricity plant at the Geysers, a Yellowstone-esque collection of natural steam vents in Northern California. It became the largest of its kind, producing a third of America’s geothermal output. Geothermal increasingly seemed like a path to U.S. energy independence, and industry heavyweights such as Chevron and Phillips opened plants. But problems soon stacked up. High temperatures caused drilling equipment to fail. Profits proved paltry and investment scarce. “Everything was a challenge, and there was never enough money to solve all those engineering problems,” Jeff Tester, a longtime geothermal researcher, told me. Tester worked on a geothermal project at Los Alamos, which aimed to perfect directional drilling. “We were never able to do it because the techniques in the field were too immature,” he said.
Then, in the two-thousands, came a rapid expansion in fracking, a drilling method that injects fluid into hard rock, creating fractures through which fossil fuels can flow. A “shale revolution” unlocked huge stores of oil and gas, causing small earthquakes, polluting groundwater, and spewing greenhouse gases along the way. A useful side effect, however, was a Cambrian explosion of new drilling practices—such as horizontal drilling and magnetic sensing—that inspired a geothermal resurgence. In 2005, the Department of Energy commissioned Tester, then a professor at M.I.T., to revisit the subject; his team’s report, “The Future of Geothermal Energy,” calculated that just two per cent of the heat in the four miles below U.S. soil could meet the entire country’s energy needs—two thousand times over. “That report was inspirational to a lot of people, including myself,” Mark McClure, the C.E.O. of a company that provides software and consulting to the oil, gas, and geothermal industries, told me.
The next generation of geothermal-energy projects—including FORGE, in Utah—was not about harvesting natural steam but, rather, drilling to create steam. These projects, known as enhanced geothermal systems, or E.G.S., generally circulate water between underground wells. Hot rock then turns it into superheated steam, which can spin turbines and generate electricity.
Engineering such a system in hard, deep rock is a profound challenge. Moore, the FORGE geologist, said that wells need to be linked with a network of fractures in the rock, which he likened to cracks in a pane of glass. Consequently, many of his colleagues are former frackers. One of them, Paul Stroud, took me on a tour of the drilling rig; we walked thirty feet up a steep staircase to a walled-in control room known as “the doghouse.” “This is an older rig,” he told me, almost apologetically. We could see sharp, snowy peaks in the distance. An array of computer screens displayed the drill bit’s depth, speed, torque, and rate of penetration. The long list of variables spoke to the complexity of carving horizontal wells in deep, hot expanses of granite.
The cement truck from Bakersfield arrived around 8:30 P.M. By ten-thirty, the men were pouring cement again, gluing the enormous metal straw in place. Next, the team scanned the borehole with gamma rays. What drillers do not want to see are spaces around the cement column. The scan showed no spaces.
Well 16B(78)-32 was completed in July, 2023, not far from another well that had been drilled earlier. It was more than two miles long. The day after its completion, engineers successfully circulated fluid through artificial fractures between the two wells. The men were ecstatic. “Connectivity is the whole goal,” Moore said. “We achieved it.” But he was also thinking about all the steps that had to follow. No heat had been mined yet, no energy produced. “We need to drill more wells,” Moore told me soon afterward. “We need to get into hotter environments.”
Since then, many of these techniques have been proven. In the summer of 2023, Fervo, an E.G.S. startup, set a record for continuous flow rate between multiple wells. They broke it again last year. “It’s a really big deal,” McClure told me. The firm soon started building a four-hundred-megawatt commercial power plant near the FORGE site, gleaning much of its technical know-how from Moore’s team. In 2026, Fervo intends to start providing around-the-clock power to utilities in Southern California. Tech companies such as Google, Meta, and Microsoft have incorporated geothermal power into notoriously energy-hungry A.I. data centers. And Eavor Technologies, a Canadian company operating in southern Germany, is developing a commercial E.G.S. project that would provide heat and electricity to tens of thousands of homes. Moore is more bullish on geothermal than ever; in October, the D.O.E. awarded FORGE another eighty million dollars to continue its research into 2028. “I encourage this in every way possible,” he said. “We need to continue.”
A small number of engineers hope to progress beyond enhanced geothermal to “deep geothermal,” which would reach depths of six to twelve miles and temperatures in excess of eight hundred degrees Celsius. (This is about as hot as the tip of a candle flame.) Believers view deep geothermal as a climate savior; skeptics view it as a technically infeasible pipe dream. “I’m still on the negative side of ‘Will this work in reality?’ ” Philip Ball, a geothermal consultant, told me. Moore called the idea “bullshit.”
On a pleasant spring morning in 2023, I visited a deep-geothermal lab in Houston, a city that oil built. I plugged in my rental, a Chevy Bolt, near a warehouse operated by Quaise, a startup with fifty-five employees that began in Cambridge, Massachusetts. From the parking lot, I could see a neighbor’s two white horses.
Carlos Araque, a bearded and smooth-talking engineer from Colombia who founded Quaise, told me that he left the oil industry in 2017 to search for a renewable energy source “with the real potential to displace fossil fuels.” He predicted that by 2050 humanity’s energy consumption will have doubled. “Wind and solar won’t scale to the level that’s required by the civilization we’ve built on fossil fuels,” he said. “It’s physics.” A pound of oil contains far more energy than a one-pound rechargeable battery. He also claimed that solar and wind projects have run out of affordable land—something that Jacobson, the Stanford professor, deemed “incorrect.” (Jacobson estimates that renewables require about half a per cent of the Earth’s landmass to cover humanity’s energy consumption.)
