A geothermal plant sells one simple promise to the grid: power whenever the grid needs it, day or night, rain or shine.
That pitch sits right next to nuclear and hydro on every slide deck.
It is why utilities sign decades long contracts for it.
It is also why the newest federal numbers landed with a thud.
Those numbers do not show a steady machine.
They show a fleet that just posted one of its sharpest single year slides in the measured record.
Why a geothermal plant’s output drifts away from its rated power
A geothermal plant does not burn fuel it can simply reorder more of. It pulls hot water or steam from a specific patch of rock, and that rock has a finite heat budget that thins out the longer a well produces. As pressure and temperature in the reservoir fall, the same wellhead delivers less punch to the turbine.
So operators compensate with new infield wells, reinjection of spent fluid, and chemical treatments that fight mineral buildup inside the pipes. Those fixes cost money and buy time, but they rarely restore a field to its opening day performance.
That is the quiet mechanical drift baked into almost every geothermal contract from the first day it is signed.
Flash plants face an added problem: scale. Minerals dissolved in the hot brine precipitate out as the fluid cools and depressurizes, coating the inside of pipes and heat exchangers like a kettle that is never descaled. Binary plants fight a gentler version of the same fouling problem, which is why both plant types tend to lose a bit of punch every single year they run.
What the newest numbers actually show at the national level
The federal agency that tracks every power plant in the country reports a single figure, called the capacity factor, that measures how much electricity a plant actually generated against how much it could have generated running flat out all year. For geothermal, that number is the gap between the sales pitch and the invoice.
According to the most recent annual data compiled from plant level reporting, the national geothermal capacity factor fell to 64.6 percent last year, down from 69.4 percent the year before and 69.0 percent the year before that.
That is a drop of nearly five percentage points in a single year, on a fleet that utilities still buy power from under contracts written assuming numbers in the 80s or 90s.
Put in plain terms, a plant rated to deliver 100 megawatts around the clock actually averaged closer to 65 megawatts over the year. Meanwhile gas peaker plants and battery projects covered in four-hour battery storage pricing comparisons are increasingly the backstop grid planners lean on when that geothermal megawatt count comes up short.
Where this comes from and why the industry still calls it reliable
This is not a one year blip buried in a press release. The figures trace back to monthly and annual filings that every geothermal operator in the country is legally required to submit.
Those are the same reporting forms researchers and federal analysts have used for more than a decade to track the fleet’s performance over time. Because the agency began publishing this detailed technology level breakdown in December 2013, analysts can watch the fleet’s average climb, plateau, and fall again across different years, rather than relying on any single company’s marketing claims.
That long run of government data is exactly why the latest dip stands out: it breaks a run of years where the number had held roughly steady in the high 60s. Researchers have flagged similar stretches of decline before, including one analysis of EIA figures that found the national capacity factor had slid across 2008 through 2012, so a fleet wide dip is not unprecedented, even if the size of this one year drop is unusual.
Even so, geothermal still beats most other renewable sources on this same measure, which is the industry’s standard defense. A capacity factor in the 60s still comfortably outpaces solar or wind, which is why utility planners still value geothermal contracts even as the specific number they are getting slips. The claim to beat nuclear or gas outright, though, no longer holds once nuclear plants are still running in the 90s.
The limits of the comparison and where the gap matters most
None of this means every geothermal field is failing. Some fields, especially newer binary projects built with reinjection designed in from day one, hold their output far better than the older flash plants that dominate the historic fleet.
The national average blends decades old fields with brand new ones, which flattens out just how steep the decline can be at any single site.
Still, the gap matters most to whoever signed the power purchase agreement. A utility that priced a decades long contract on an assumed 80 to 85 percent capacity factor, similar to the modeling choices developers lean on in project finance, ends up paying for megawatts that never show up on the meter. That shortfall gets made up somewhere, often through the same kind of rate cases that have reshaped utility bills from Oklahoma to Virginia in recent filings.
And the timing bites hardest exactly when it should not. Geothermal output tends to sag most in summer, when cooling towers work harder in the heat and power prices on the grid are already at their highest. That is the opposite of what a round the clock resource is supposed to deliver, and it leaves grid operators scrambling for backup precisely when demand peaks.
What it means for the next generation of geothermal projects
None of this kills the case for geothermal power. It does mean the newest wave of projects, including enhanced geothermal systems that fracture hot rock rather than relying on a natural reservoir, are being watched closely for whether they hold their output better than the older fleet did.
Federal research money has already followed that question, with recent funding rounds aimed squarely at next generation drilling and reservoir management techniques meant to slow the same decline described here.
Whether those newer designs age better than the plants built decades ago will not be clear for years, since the whole point of a capacity factor is that it is measured over a full year of real operation, not a single test run.
For now, the lesson for anyone reading a grid report or a utility rate filing is simple: round the clock power on paper and round the clock power on the meter are not always the same number, and the gap between them is exactly what federal data now tracks in public view every year.