Geothermal, especially enhanced geothermal systems (EGS) of interest to drillers and fracturing experts, went through a period of irrational exuberance in its initial proposal phase, but this has slackened as it has entered the “prove it” phase.

Geothermal startups were coming out of the woodwork left and right with grand proposals for EGS pilot projects across Texas, California, and the southwest with funding from the DOE, traditional oil and gas operators, and private sources.

But detailed, long-term electrical generation results are not yet in hand. Fervo has been the most active and the most publicized of the EGS startups. Its Cape Station Phase I in Utah comprises three 33-Mw GeoBlocks. First power for GeoBlock 1 is expected in fourth-quarter 2026. GeoBlocks 2 and 3 are expected to reach initial power in early 2027.

Cape Station’s second phase comprises 400 Mw across eight 50-Mw GeoBlocks and is expected to come online in 2028. The company reports 1,054-Mw binding planning performance agreements (PPA) with utilities, hyperscalers, and community choice aggregators (entities which aggregate buying power on behalf of individual customers), with a recent PPA signed with Google to provide about 400 Mw for 15 years, starting in third-quarter 2028.

Sage Geosystems has not lagged far behind in funding or attention. It is evaluating geothermal for energy storage as well as power. The company sold the first electricity from its energy storage installation at the San Miguel Electric Cooperative Inc. (SMECI) plant in south Texas in second-quarter 2026. The storage site is currently being evaluated for subsurface pressure response, injection and production, water management, system efficiency, and reservoir behavior.

These new plants and contracts for power are encouraging, but they hinge on steam turbine availability and system operational costs. The turbines are a bottleneck in geothermal plant construction. 

Operational costs for EGS are another concern. The effects of geothermal steam or supercritical water on operational components, the hot dry rock source heat-depletion, long-term fracture conductivity between injector and producer, downhole water losses, and a host of other issues yet unknown threaten wholesale adoption of EGS systems. Fortunately, most of these issues lie within typical oil and gas engineering disciplines. Unfortunately, we are just at the beginning of understanding them for EGS, and investor patience may be waning. 

Scaling up geothermal to a functioning large-scale power plant requires tens of billions of dollars, an amount which appears frustratingly difficult to raise for this clean but commercially unproven technology. Fervo issued an IPO on May 13, 2026, and within days it rose to about $40/share to peak at $42.50 a week later. Despite a string of good news since then, the stock currently trades at about $15-19.

Contrast that trajectory with funding for data centers, AI companies, and hyperscalers, which measure in the hundreds of billions of dollars with only recent signs of retrenchment. Nvidia stepped in to steel spines with a GPU leasing strategy to keep the data-center ball rolling.

Said ball has uncertain power and cooling-water supplies and an unproven business model in the face of unprecedented leveraged borrowing. Geothermal, with its promising potential to directly power data centers, add power to the grid, and provide district and industrial heating, could use even a fraction of the funding.