Beneath desert mesas and old volcanic fields there’s a steady, relentless heat — and a growing chorus of engineers, investors and policymakers saying we should use it. Call it geothermal’s second act: not the sleepy steam plants of the past but next‑generation systems that could deliver always‑on clean power to data centers, factories and entire towns.
From a trickle to a tide
Geothermal today is a niche: roughly 0.4% of U.S. electricity comes from heat tapped near the surface. But two trends are changing the math. One is demand — accelerated by sprawling data centers and AI services that need round‑the‑clock electricity — and the other is technology, including techniques borrowed from decades of oilfield drilling and new closed‑loop designs that sidestep some old constraints.
Investors can see it. Global capital flowing into geothermal has climbed from under $2 billion a year at the start of the decade and, according to industry research, could approach $9 billion by 2030 as next‑gen approaches move from labs to projects. Companies such as Fervo have drawn big rounds to finish utility‑scale builds, and startups like XGS Energy are signing deals to supply hyperscalers: XGS announced a sizable agreement to power a Meta data center in New Mexico and recently demonstrated a closed‑loop system that runs without touching native groundwater.
Two technical paths: fracture the rock, or surround the well
Traditional enhanced geothermal systems (EGS) use stimulated fractures to connect hot rock with injected water that becomes steam. The oil and gas world has handed geothermal useful tools — directional drilling, reservoir modeling, big rigs — but the approach carries tradeoffs: high drilling costs, water use in dry regions, and the risk that injections can trigger seismic events.
Newer designs aim to avoid some of that risk. Closed‑loop systems put a sealed circuit of steel tubing or engineered materials down a well so working fluid circulates without contacting the subterranean aquifer. XGS, for example, emphasizes a sealed, modular approach and a downhole thermally conductive material that boosts heat transfer; the company says that reduces water consumption and simplifies permitting because the system doesn’t “touch your water.” That appeals to operators in the drought‑stressed American West and to customers like data centers that prize predictable, continuous baseload power.
Meanwhile, advocates of superhot rock (SHR) geothermal — tapping rock hotter than 400°C at depths of many kilometers — argue that if engineers can master the tools and materials needed for those conditions, each well could yield five to ten times the energy of today’s conventional wells. The Clean Air Task Force laid out a detailed road map this year calling for coordinated international testbeds, high‑temperature drilling tools, and material science investments to make SHR commercializable.
The bottlenecks: money, materials and maps
Even with enthusiasm, hurdles are very real. Drilling to deeper, hotter zones is expensive; the high‑temperature environment requires new alloys and sensors; and developers still don’t always know precisely where the sweetest heat sits beneath a given landscape. Geologists caution some of the early hype — mapping high‑quality reservoirs at scale remains difficult, and power plants still need transmission lines to reach customers.
Permitting and politics are part of the picture, too. Geothermal has found an unusual degree of bipartisan support in Washington: tax incentives for geothermal survived recent budget negotiations even as aid for other renewables shifted. State governments are also moving — New Mexico lawmakers, for instance, now allow conversions of certain oil and gas wells for geothermal use, and the state has been flagged as having on the order of 160 gigawatts of dry, hot‑rock potential in recent assessments.
Where the electricity will actually go
Hyperscale data centers are already lining up as early buyers. They need carbon‑free, always‑available power to keep AI racks humming; that steady demand is a big reason geothermal is suddenly commercially interesting. The industry’s momentum sits beside other big bets on data‑center infrastructure — a reminder that electrification choices ripple across sectors (see the scale of cloud and AI infrastructure plans like Google’s Project Suncatcher). And as services such as Google’s Gemini push deeper into live search of user data and cloud workloads, the electricity appetite only grows (Gemini’s Deep Research is part of that story).
But electricity isn’t the only use case. District heating projects — small, practical demonstrations like the utility‑owned geothermal network pilot heating municipal buildings and homes in Framingham, Massachusetts — show how shallow geothermal and networked heat can cut bills and emissions for communities. That diversity of applications strengthens geothermal’s long‑term case: from gigawatts for AI to looped services that warm neighborhoods.
Investment, coordination and a very practical road map
What’s needed now looks less like a single miracle and more like coordinated effort: public R&D to de‑risk deep tools, international testbeds to standardize measurements, blended finance to seed facilities that validate commercial designs, and workforce training to staff the drills and labs. The Clean Air Task Force’s phased plan — from governance to deployment to continual improvement — reads like a blueprint for turning scattered pilots into an industry.
Private capital is responding: large project financings and big venture rounds show investors expect returns if technology and permitting line up. But early movers with subsurface and drilling expertise will have an outsized advantage: they write the standards, build supply chains and own early intellectual property.
For communities and utilities, practical questions will drive decisions. How much water does a project use? Can a plant be sited near load centers to avoid long transmission builds? What happens if a reservoir cools faster than models predict? The answers will come slowly, well by well.
Momentum doesn’t erase uncertainty. Still, between a clear market pull (AI, industrial electrification, municipal heating) and multiple technical routes forward (EGS, closed‑loop, SHR), geothermal has moved from an afterthought to a portfolio play for decarbonization. The next decade will likely decide whether this heat becomes a marginal supplement or a major backbone of clean, firm power — and whether the people who drill, finance and regulate can turn subterranean warmth into dependable electricity at scale.