Traditional geothermal waited on lucky geology — a natural pocket of hot water already open enough to flow. Enhanced geothermal systems do something less romantic and more industrial: they drill into hot rock that would otherwise sit unused, open pathways the way an oilfield crew would, and circulate fluid through heat that never clocks out for the night. The applied point is not a new physics lecture. It is firm power sold on a schedule.
Fervo Energy is the clearest commercial test of that idea. At its Project Red pilot in northern Nevada, the company drilled a horizontal injection–production well pair, stimulated the rock with multistage hydraulic fracturing and proppant, and in 2023 ran a commercial-scale well test that produced on the order of 3.5 megawatts electric-equivalent — enough to prove the method was not a science-fair curiosity. The toolkit looks familiar to anyone who has watched Permian Basin pads: lateral wells, plug-and-perforate stimulation, real-time drilling data, Helmerich & Payne rigs. The product is different. Hot brine moves through a closed loop, dumps heat into a power plant, and goes back underground.
Cape Station scales the method
Cape Station, in Beaver County, Utah near Milford, is where that method scales. Phase I is an approximately 100-megawatt build made of three modular 33-megawatt units Fervo calls GeoBlocks. By mid-2026 the company had mechanically completed GeoBlocks 1 and 2; commissioning on the first unit had already connected wells to the plant, moved geothermal brine through heat exchangers, and spun turbines. Company reporting targets test power from GeoBlock 1 in the fourth quarter of 2026, with full Phase I production aimed by year-end and the remaining blocks ramping in early 2027. Phase II — about 400 megawatts across eight 50-megawatt GeoBlocks — entered construction in early 2026 and is aimed at commercial operation in 2028. Together the first two phases are a half-gigawatt greenfield project, widely described as the largest enhanced geothermal system under construction.
Offtake pays before the turbines finish
The commercial half of the story matters as much as the drilling. In late August 2026 Fervo signed a 15-year, 396-megawatt power purchase agreement with Google for Cape Station capacity, to be delivered in four 99-megawatt tranches with target commercial dates beginning in the third quarter of 2028 — framed as the largest enhanced geothermal PPA to date, with an option for Google to take roughly another 600 megawatts. That sits on top of earlier contracted offtake with buyers including Southern California Edison, Shell, and a prior Google / NV Energy arrangement in Nevada. After the September Google deal, public tallies put Fervo’s binding contracted capacity on the order of a gigawatt. Hyperscalers want carbon-free power that does not wait on the weather; utilities want new firm supply without pretending the grid can stretch forever. Enhanced geothermal sells into both needs.
Costs are still high compared with gas or mature renewables — Phase I all-in costs have been reported around $7,000 per kilowatt, with Phase II targeted near $5,500 and a longer-term goal of $3,000 — but the learning curve is the industrial one: faster wells, longer laterals, hotter rock, standardized plants. At Cape, Phase I averages already compressed drilling times sharply versus the Nevada pilot; Phase II wells are pushing toward nearly 20,000 feet of measured depth in hotter resource. That is oilfield repetition applied to heat, not a one-off research hole.
For appliedscience.com, the hook is simple. Enhanced geothermal stopped being “maybe someday heat” when someone treated hot granite like a completion program and signed megawatt-scale offtake with names that pay invoices. Cape Station is still commissioning as of September 2026, so first electrons are a near-term proof, not a finished fleet. The science that already shipped is the method: horizontal wells, stimulated rock, modular power blocks, and buyers who contracted before the turbines finished spinning.