Super-Hot Rock Geothermal Could Reach Data Centers, but the Engineering Is Still Ahead

The United States has up to 57 terawatts of super-hot rock geothermal generating capacity, according to TechCrunch, which reported the figure on November 20, 2026. The same report puts the output of a single super-hot rock well at 15 megawatts of electricity or more.
Super-hot rock geothermal means drilling into rock far hotter than the reservoirs conventional geothermal plants use, then using that heat to drive turbines. A terawatt is one trillion watts and a megawatt is one million, so the resource figure is an estimate of what the rock could hold, not a measure of power being produced today.
The resource estimate is the backdrop for a cluster of commercial and technical announcements, most of them still at the plan stage. Quaise Energy said it had landed Google as a customer. Google is to pay more than $500 million for 50 megawatts from a Quaise project in central Oregon that taps heat three miles deep. The deal is an offtake agreement, a contract to buy a project's future output.
Drilling is where the technical risk sits. Mazama Energy, which is backed by Khosla, said it had worked out how to keep its drill heads cool while tapping super-hot rock. Its Carbon Chill technology uses carbon dioxide in place of water-based drilling mud, the fluid normally pumped down a well to cool the bit and carry rock fragments back up. Quaise, for its part, plans to replace its conventional drill bit with a gyrotron-powered bit, a high-power microwave source, starting at about 2.5 miles deep. That bit would bombard the rock with millimeter-wave radiation, very short radio waves, to turn it into ash rather than grinding through it mechanically. Think of a microwave oven aimed at stone, though the scale is entirely different.
Reliability is the third thread. XGS Energy received a Department of Energy grant to refine materials, well designs and models for improved reliability at the depths where super-hot rock exists. Fervo Energy, which went public earlier in 2026, gave broader markets their first look at enhanced geothermal, according to the same report.
In my view, the headline's framing, that super-hot rock could end data center energy woes, claims more than these items establish. A resource estimate, a single-well output figure, one announced offtake agreement and a set of drilling methods under development describe a pipeline, not a supply of power operating at scale. The gap between a well that can produce 15 megawatts on paper and a field that delivers 50 megawatts under contract is where most geothermal projects have historically been tested.
The broader context here is that a hyperscaler, a large cloud operator, is willing to commit capital to a geothermal project that does not yet exist at the depths being targeted. That is a signal about where buyers think firm, carbon-free supply, meaning power available around the clock without emissions, will come from. If the bit-cooling approaches and the millimeter-wave drilling concept hold up in the field, the Oregon project and the resource base behind it could give data center operators electricity that is not tied to a particular weather pattern or fuel contract. Whether those engineering steps convert to megawatts on schedule is the question the next few years of field results will answer.


