Superhot Geothermal Wells at Krafla and Reykjanes, Iceland

Published June 14, 2026

The Deep Harvest, Chapter 3: Into the Deep: Superhot Rock

At Krafla in Iceland, a well called IDDP-1, drilled for the Iceland Deep Drilling Project, ran into magma at about 2.1 kilometers down. When the well was flow-tested, steam came up at about 452 degrees Celsius and 142 bar, roughly 140 times the air pressure at sea level. That single borehole carried about 36 megawatts of electric output, 7 to 12 times what a conventional Icelandic geothermal well delivers.

The well did not last. Its casing, the steel pipe set in the hole to hold it open and carry the fluid to the surface, failed under repeated heating and cooling and the corrosive chemistry of the fluid. The project tried again at Reykjanes. IDDP-2 reached about 4.66 kilometers and found fluid at about 426 degrees Celsius and 340 bar, which confirmed supercritical conditions. Above 374 degrees Celsius, and under enough pressure, water stops behaving as separate liquid and steam and becomes a single dense fluid. Wells that tap that kind of fluid in superhot rock are expected to give 5 to 10 times the power of an ordinary geothermal well. At Reykjanes the casing failed too, and production was cut short.

In both Icelandic wells, the heat was there. What failed was the materials: the casing and its fight with hot, corrosive fluid. A model from the Clean Air Task Force and the University of Twente puts the superhot rock resource at about 63 terawatts if just 1 percent of it were used, and the report carries its own caveats on how confident that figure is. Whatever the exact number, casing, mineral scale in the pipes, the cooling of the rock around the well, and where and how the well is placed decide how much of that heat anyone can sell.

In June 2025, Fervo drilled its Sugarloaf appraisal well to a true vertical depth of 15,765 feet, or 4,807 meters, in 16 days, 79 percent less time than Department of Energy baselines. More than 90 percent of Fervo's workforce came from the fossil fuel industry. Quaise is testing a different tool, a beam of millimeter-wave energy in place of a drill bit. Its field demonstrations have gone about 100 meters into granite, at rates up to about 5 meters an hour, against about 0.1 meter an hour for conventional drilling in granite.

Fervo's wells are enhanced geothermal systems, or EGS. Instead of waiting for a site with natural hot water, an EGS project opens fractures in hot rock and pumps water through them to collect the heat. At Project Red, Fervo ran a 30-day test that held a flow of 63 liters a second at 191 degrees Celsius and produced 3.5 megawatts of electricity. EGS power cost about $140 per megawatt-hour in 2024, averaged over a plant's life, and the National Renewable Energy Laboratory has laid out pathways for that cost to fall.

Pumping water into fractured rock has a known failure mode. In Basel in 2006, a large injection of water was followed by a magnitude 3.4 earthquake. At Pohang, the seismicity has been analyzed in detail. A developer proposing EGS has to answer for both. One answer is to stop opening the rock. Eavor's closed-loop design circulates fluid through a sealed loop of pipe underground, where the hot fluid rises and the cooler fluid sinks without needing to fracture anything. At Geretsried, built on an abandoned hydrothermal site, Eavor began selling electricity in late 2025 from a surface footprint of about one hectare. It is too new for multi-year data, so nobody can yet say how fast the rock around that loop cools.

The fluid that comes out of an open well brings its own trouble. Hot geothermal water carries dissolved silica, and as the water cools on the way through the plant, the silica comes out of solution and coats the pipes. Chemical inhibitors can slow it, but each one works only within a limited range of calcium in the water. Underneath, the rock itself cools when heat is drawn out faster than it flows back in. That cooling is called drawdown, and on the timeline of a project a single reservoir's heat is finite.

A plant that can keep its wells open has more to sell than electricity. The Deep Harvest treats a geothermal site as a campus: data centers that need power around the clock, desalination plants, lithium recovery, industrial users that need heat, and thermal storage. One form of that storage is a Carnot battery, which stores electricity as heat and turns it back into electricity later. A proposal for a network of superhot rock plants includes pulling lithium, silica, and other minerals out of the produced fluid. That is a proposal. Whether commercial superhot rock sites can recover those minerals at a profit has not been shown. Desalination brine is another possible mineral feed on the same campus, which places the water plant and the mineral plant next to the same heat.

Buyers who need steady power and heat around the clock will get it from superhot rock only where the casing holds against heat and corrosion, the pipes stay clear of scale, and the rock keeps its temperature, the test the Krafla well failed on its casing. Every open well that passes it also brings up a stream of hot fluid carrying dissolved minerals and gases, and that stream is the next resource on the campus. The Deep Harvest sets up that campus in Chapter 3, Into the Deep: Superhot Rock.