It is 2 a.m. somewhere in Nevada, and a data center full of AI chips is wide awake, drawing power as if the sun were up. The sun is not up. Solar panels are asleep, the wind is undecided - and yet the lights stay on. Some of that electricity is coming from three miles underground, from rock that has been hot since before there were lights to keep on. That rock now has Fervo Energy's name on it.
A geothermal company that finally got the economics to work
For most of its history, geothermal energy had a real estate problem. It worked beautifully - clean, constant, available day and night - but only in the rare places where the earth was generous enough to push heat near the surface. Iceland, sure. A few corners of California and Nevada. Everywhere else, the heat was there; it was just too deep and too dry to bother with.
Fervo Energy's pitch is that "too deep and too dry" stopped being a dealbreaker. By borrowing horizontal drilling and reservoir engineering from the shale boom, the company drills down and then sideways through hot dry rock, creates an engineered reservoir, and circulates water to pull heat back up as steam. The result is firm, 24/7, carbon-free power that can be built in far more places than old-school geothermal ever could.
Fervo's mission is to leverage innovation in geoscience to accelerate the world's transition to sustainable energy.
— Fervo Energy, on what it is actually forThe grid wants power that doesn't take breaks
Renewables had a dirty secret that everyone politely ignored: the sun sets and the wind stops. As more of the grid went solar and wind, the value of power that is always on - "firm" power, in the trade - kept climbing. Utilities needed something that could run at 3 a.m. in February without burning gas. Data center operators, suddenly the hungriest customers on the planet thanks to AI, needed it even more.
The usual answers were nuclear (slow, expensive, politically exhausting) and batteries (great for hours, not for seasons). Geothermal could do the job - it just had to escape its geography. That gap is the whole reason Fervo exists, and every part of the company points back at it.
Cape Station is expected to be the largest next-generation geothermal development in the world.
— The bet, stated plainlyTwo Stanford grads who spoke fluent drilling
Tim Latimer started his career as a drilling engineer, the kind of person who knows what a drill bit costs and how badly a stuck pipe ruins a week. He went to Stanford, met Jack Norbeck - a reservoir scientist who would become Fervo's CTO - and the two of them made a contrarian wager in 2017: the techniques the oil and gas industry spent billions perfecting could be turned toward the climate instead of away from it.
It was a strange thing to believe at the time. Geothermal was a sleepy corner of clean energy, the kind of sector that got polite nods and little funding. Latimer and Norbeck bet that the science was ready and only the engineering economics were missing. The trick was not inventing something new so much as pointing something old at a new target.
What makes Fervo unmistakably Fervo
- It listens to rock. Distributed fiber-optic cables run thousands of feet down each well, sending back real-time data on what the reservoir is doing.
- It fracks for heat, not oil. The horizontal drilling DNA is straight from shale country - same playbook, very different payload.
- It gets cheaper every time it drills. Each well teaches the next one, and the cost curve has bent hard downward.
Heat, sold by the megawatt
Fervo's first real proof was Project Red in 2023 - the first commercial enhanced geothermal system built with horizontal wells, generating about 3.5 megawatts. Modest in size, enormous in meaning: it showed the approach worked outside a lab.
Then came Cape Station, in the Utah desert near Milford. It is designed to deliver an initial 500 megawatts of firm clean power, with as much as 2 gigawatts permitted at the same site. The company doesn't just sell hardware; it develops, builds, owns, and operates the plants, then sells the electricity under long-term contracts. The product, in the end, is a promise: power that shows up, every hour, with no carbon attached.
A short history of borrowing the earth's heat
// MILESTONES, ABRIDGED
The numbers that turned skeptics into customers
The argument for next-gen geothermal lives or dies on cost. Early geothermal wells were too expensive to compete; the whole question was whether drilling enough of them would make each one cheaper. Fervo's answer, between Project Red and Cape Station, is the most persuasive thing it has: drilling speed up roughly 70%, well costs down close to 75%.
Customers noticed. The company holds binding power purchase agreements covering 658 megawatts, with counterparties that read like a who's-who of the grid: Southern California Edison, Google, NV Energy, and Shell. Those contracts add up to a multi-billion-dollar backlog - revenue that is signed but not yet flowing.
The cost curve, bending the right way
// FERVO DRILLING PERFORMANCE, PROJECT RED → CAPE STATION
Figures are approximate and drawn from company statements and reporting around the 2026 IPO. Bars are scaled for comparison, not to a single common axis.
The earth has been a fully charged battery for four and a half billion years. Fervo's job is just to find the outlet.
— The simplest way to explain the whole companyClean power that behaves like the old kind
Fervo's stated mission - accelerating the transition to sustainable energy through geoscience - sounds like every other climate company's mission until you notice the specific thing it is solving. It is not trying to add more intermittent renewables to a grid that already has plenty. It is trying to deliver the one thing the clean grid still lacks: power that is on all the time, everywhere, without a smokestack.
That is also why the company leans on public-private partnerships and community engagement around its sites. A geothermal plant is not a software product you can ship and forget; it is steel in the ground in someone's county, and the politics of that are part of the work.
The AI era needs a power source that never sleeps
Here is the irony Fervo gets to enjoy: the same artificial intelligence boom that is straining the grid is the best thing that ever happened to its business. Data centers want enormous amounts of clean, constant power, and they want it on contracts long enough to justify building new plants. Geothermal, suddenly, is exactly the right shape for the demand.
Whether next-gen geothermal becomes a pillar of the grid or a promising footnote depends on execution that is still underway - more wells, more megawatts actually delivered, costs that keep falling. Fervo has made the case on paper and in the market. Now it has to make it in the dirt.
Back to that data center in Nevada, still awake at 2 a.m. The chips keep humming. The power keeps arriving. And three miles down, in rock that has been hot since long before anyone thought to drill it, Fervo Energy has quietly turned a geography problem into a business - and the planet's oldest heat into the grid's newest habit.