The pitch is almost irresistible: no fuel, no weather dependence, power around the clock from heat sitting inside the Earth for billions of years.
Unlike a solar farm, a geothermal plant keeps running at midnight. Unlike a gas plant, it never needs a delivery truck. But crack open the fine print on how these plants actually work, and how the newest enhanced generation pushes the physics further, and a physical consequence appears that operators rarely put in their brochures.
The question is whether the rock itself will cooperate quietly, or push back.
Why pushing fluid into hot rock under pressure can wake a fault
Enhanced geothermal systems manufacture their own underground reservoir. Because most of the hottest rock on Earth is also the densest and least porous, there is nowhere natural for water to flow through and collect heat. The fix is hydraulic stimulation: pumping fluid down an injection well under high pressure until the rock fractures along new pathways, creating an artificial heat exchanger kilometers below the surface.
That fracturing is not just a byproduct; it is the product. The engineering goal and the seismic side-effect are, in a physical sense, the same event. Fluids injected under high pressure cause the rock mass to fail, stimulating fractures that improve fluid connectivity.
But that same rise in pore fluid pressure can also reactivate pre-existing faults, potentially inducing earthquakes of significant size. The same mechanism applies, with greater intensity, in conventional hydrothermal fields where fluid is re-injected at scale. Operators can choose where to drill but cannot always choose which faults lie nearby, or how close to failure those faults already are.
What the rock beneath a California geothermal field actually shows
The Salton Sea Geothermal Field in Southern California has been running commercially since 1982, making it one of the longest-operating fields in the United States. Unlike EGS projects, it draws on a natural brine reservoir, but its 11 plants reinject large volumes of fluid, exposing the same pressure-driven hazard. The field sits inside one of the most seismically active corridors on the continent, laced with active fault structures, which makes it the best-documented working laboratory for the question every geothermal developer eventually faces.
Researchers found that there was a 2% chance of a magnitude 6.07 earthquake nucleating inside the field within a fifty-year period prior to energy production, and there is an equal probability that a magnitude 6.70 earthquake nucleates there during the production era. That is a shift of more than half a point.
The shift matters because earthquake energy does not scale linearly. Each full magnitude point represents roughly thirty-two times more released energy than the one below it. Half a point is not a rounding error.
The pattern is also directional: seismic activity is greatest to the east of the Brawley fault, where the amount of injection exceeds the amount of production. More injection correlates with more shaking, and EGS plants inject more fluid than conventional geothermal operations by design.
Where the evidence comes from and what the broader record shows
The Salton Sea findings analyzed 50 years of seismic history in the field between 1972 and 2022 and were published in Scientific Reports in January 2025. The work was supported by the Department of Energy’s Office of Energy Efficiency and Renewable Energy, and it is among the most granular assessments of a working US geothermal field in the public record.
The pattern is not unique to California. In the Basel Deep Heat Mining project in Switzerland, deep geothermal field development triggered over 10,000 earthquakes, some reaching magnitude 3.4; the resulting events led to $9 million in insurance claims and ultimately caused abandonment of the project.
Then came a more sobering case. A magnitude 5.5 earthquake near Pohang, South Korea in 2017 injured 90 people and caused $52 million in damage. The event left approximately 2,000 houses damaged and has been linked to a nearby deep geothermal project under development. The lead researcher on one of the two peer-reviewed studies said “The Pohang earthquake was larger than any predicted by existing theories”, he had expected quakes to be small, in line with the largest previous EGS-linked event at Basel, magnitude 3.4, and it took only five days to be proved wrong. Both Basel and Pohang are now the canonical cautionary examples cited in every major geothermal risk guideline published since.
Where the catch bites hardest and what operators cannot fully control
The deeper complication is that seismic risk from fluid injection does not always behave predictably. Some large earthquakes, called runaway seismic events, do not conform to theoretical models of maximum expected magnitude. Engineers can throttle the pump, but they cannot always throttle the fault.
Traffic-light monitoring systems, named for their red, amber and green thresholds, are the current industry standard, allowing near-real-time warnings to adjust pressure and pumping rates. But those systems depend on knowing in advance where the faults are and how stressed they already are, knowledge that is expensive to acquire and imperfect even when acquired.
The regulatory framework is still catching up. The legal landscape for EGS resembles the early, under-regulated years of shale gas extraction, with sparse permitting records and novel drilling technologies outpacing existing rules. That gap is especially acute as the industry scales: the largest enhanced geothermal deal ever signed in the US reached 396 megawatts, and dozens of smaller projects are advancing through permitting across the West.
What the industry is building toward and what remains genuinely unresolved
None of this has stopped EGS from advancing. The Department of Energy’s Utah FORGE test site near Milford has been running stimulation experiments designed to map the relationship between injection and seismicity in real time. That research feeds directly into the federal geothermal center in Colorado that opened earlier this year to coordinate applied safety work across the industry.
Induced seismicity risk from EGS is estimated to exceed that from conventional geothermal fluid injection, which itself, as the Salton Sea record shows, produced a documented shift in worst-case hazard. Managing that risk across dozens of new plants in geologically varied terrain is the test the industry has not yet run at scale.
The honest version of the geothermal pitch includes a clause most marketing materials omit: these operations carry real earthquake potential. The science knows this, the engineering is working on it, and the regulatory frameworks are catching up. The fuel cost stays at zero, but the subsurface accounting is more complicated than the brochure suggests, and getting it right is the foundational work the next decade of geothermal growth depends on.
