Enhanced geothermal has been ten years away for about thirty years. On Sept. 24 it stopped being ten years away.
Fervo Energy synchronized the first GeoBlock at Cape Station in Beaver County, Utah, roughly 33 MW, to the grid. The company calls it the first utility-scale enhanced geothermal system anywhere to export power, and no one has disputed the claim. Phase I runs to about 100 MW across three GeoBlocks; commercial operations were expected by Oct. 1, with the other two blocks targeted for the start of 2027. A 400 MW Phase II is under construction for 2028.
Why enhanced geothermal is a drilling problem, not a heat problem
Conventional geothermal needs a rare combination: hot rock, natural permeability and water already in place. That’s why the resource sits in a handful of western basins and why it supplies about 0.4% of U.S. generation. Enhanced geothermal drops the permeability requirement. You drill into hot dry rock, create the flow paths, and circulate your own water.
The engineering that made it viable came from shale. Horizontal drilling and multi-stage stimulation are oilfield techniques, and Fervo’s cost curve has followed the same shape shale did: each well pair cheaper and faster than the last. That’s the part worth watching, because it’s the only reason the economics might close.
The cost numbers, and who gave them
Fervo told Canary Media that Phase I delivers at about $7,000 per kilowatt and that Phase II is projected at $5,500. For comparison, gas plants averaged roughly $2,157 per kilowatt in 2025. Those figures came from the company, not from an independent assessment, and the Phase II number is a projection.
At three times the capital cost of gas, geothermal wins only on what gas can’t offer: firm, carbon-free output with no fuel price and a flat production curve. That combination is worth a large premium to exactly one class of buyer right now, which is why Fervo has roughly 1,000 MW of contracted and in-progress offtake and why Google is behind Phase II. Shell Energy takes the Phase I power.
What still isn’t settled
Capacity figures don’t reconcile across sources. The project has been described as 500 MW and as up to 900 MW depending on whether contracted, installed or planned capacity is being counted. Anyone citing a number should cite where it came from.
And this is generation, not buildings. It isn’t the same technology as the ground-source loops going into Lennar subdivisions in Colorado at about $25,000 a home, which move heat rather than make electricity. The two get filed under the same word and share almost nothing but a drill rig.
Fervo went public in May 2026. The next number that matters is the delivered cost of the second and third GeoBlocks, because a learning curve you can’t repeat is just a first-of-a-kind discount.