What is Enhanced Geothermal (EGS)?

Conventional hydrothermal power needs heat, fluid, and permeable rock in the same place. A hot formation with little natural flow cannot deliver enough fluid to a turbine merely because the rock is hot. DOE defines EGS as creating a human-made reservoir by injecting fluid under controlled conditions to open new fractures or reopen existing ones. The enhancement is the fluid pathway through the rock, not the creation of heat.

Picture two wells reaching a hot, tight rock formation. One injects water. The other recovers heated fluid after it crosses the engineered fracture network. If the wells do not communicate, or if flow takes a short path with too little heat exchange, the surface plant will not see the expected output. The power plant is therefore the last part of a connected system: rock, wells, circulation, heat exchanger, turbine, and grid connection.

EGS is distinct from a sealed underground heat exchanger. DOE’s technology comparison says closed-loop systems circulate fluid inside pipes and do not rely on a fracture network in the rock. Both use subsurface heat, but their reservoir and fluid-management risks differ. “Next-generation geothermal” can cover both, so a project announcement alone is not enough to tell which design is proposed.

Site characterization comes before drilling. Developers need the temperature at depth, rock type, stress direction, existing fractures, available water, and a route for surface facilities. DOE’s EGS development outline starts with a geologic model, then an injection well, reservoir creation, a production well, and operation. A hot spot on a map is a reason to investigate, not proof that a field can produce contracted power.

During stimulation, water enters the formation under controlled pressure to increase permeability. Sensors track how the fracture system responds. The operator then tests whether the production well receives enough hot fluid from the injection well. DOE’s FORGE site in Utah studies drilling, stimulation, well connectivity, and flow testing precisely because these steps determine whether an engineered reservoir behaves as intended. A power purchase agreement cannot substitute for a successful flow test.

At the surface, the plant converts heat into electricity. DOE explains that a binary-cycle plant transfers heat from geothermal fluid to a secondary working fluid; that secondary vapor drives the turbine. This lets lower-temperature geothermal resources generate power without sending reservoir fluid through the turbine. A project may use another conversion configuration when temperature and fluid conditions warrant it. The key output for a buyer is net electricity after pumps and plant auxiliaries, measured at the agreed delivery point.

A vertical well samples one path through the resource. A horizontal or deviated well can expose a longer interval of hot rock and intersect more engineered flow paths, provided the geometry and stress field cooperate. Fervo says its Cape Station design adapts horizontal drilling methods from oil and gas. That is a project-specific engineering choice, not a promise that the same well design works in every basin.

The central field test is sustained circulation. At Project Red, Fervo reported a 30-day test in 2023 with a 63-liter-per-second flow rate and heat sufficient for 3.5 MW of electric production. Those are measured pilot results. They do not scale automatically by multiplying the number of wells: adjacent wells can interact, rock temperature and pressure vary, and parasitic pumping load grows with flow.

DOE reported an extended circulation test at FORGE in 2026. Longer tests matter because a short burst of hot fluid says less about sustained output and cooling of the local rock. A data-center buyer evaluating an EGS proposal should ask for the test duration, measured temperature and flow, expected decline or recovery assumptions, pump electricity, and how additional wells change the reservoir model.

The distinction between first power, commercial operation, and later phases matters. Fervo announced first power at Cape Station in September 2026: the first GeoBlock synchronized to the grid and exported electricity. Its September announcement described Phase I as three 33 MW blocks, with the first expected to reach contractual commercial operation by October 1 and the other two by January 1, 2027. Fervo’s commercial-operation announcement reports that GeoBlock 1 reached commercial operation on September 30, 2026. As of November 2026, the remaining Phase I blocks and the 400 MW Phase II target for 2028 are planned milestones, not completed output.

This is useful evidence for a buyer because it separates a proven first block from a full-field schedule. A campus contracting for a future 100 MW delivery should check whether its obligation depends on one block, all three blocks, or grid delivery from a larger portfolio. The plant could be generating while the customer still waits for transmission or a direct electrical connection.

The USGS Great Basin assessment estimated a provisional 135 GW electric resource in hot, low-permeability rock if technology advances enough to commercialize it widely. That is resource potential, not installed generation or available capacity for a particular site. The operational record at a commercial block and a regional resource estimate answer different questions and should never be added together.

Geothermal heat is present day and night, so a producing field can supply a continuous power profile. DOE describes EGS as a potential source of firm, flexible power. Yet a customer still needs an electrical delivery arrangement and a plan for plant outages. A “firm” source in a resource brochure is not the same thing as an uninterrupted 100 MW at a campus bus.

For example, if a data center needs 40 MW every hour and an EGS plant contracts to deliver that amount, the contract must say what happens during a pump failure or turbine maintenance. The site may use the grid, another generator, storage for a short interruption, or workload curtailment. If the geothermal plant is remote, transmission rights and losses matter too. If the plant is co-located, the customer must still design switching, protection, and reserve supply.

Some geothermal projects sell into the grid rather than directly powering the named buyer. Fervo’s first Cape Station block exported to the grid; that fact does not make it a behind-the-meter data-center plant. Our behind-the-meter guide explains the meter and connection choices. When comparing bids, ask for the net capacity, hourly delivery obligation, outage provisions, and exact electrical point at which power is promised.

Four categories of uncertainty deserve separate evidence. First, the temperature at economically drillable depth must be high enough for the intended power cycle. Second, wells must connect through the reservoir with enough sustained flow. Third, pumps, heat exchangers, and the surface plant must convert that heat into enough net electricity. Fourth, the project needs water, permits, land access, and a route to the customer. DOE’s EGS research overview names geologic qualities, permeability, and well construction as central technical challenges.

The environmental review is also specific. Injecting fluid changes rock stress and can produce induced seismicity. DOE’s seismicity protocol notice discusses monitoring, public communication, and risk management; it does not say the risk is zero. The appropriate controls depend on the site and regulator. Water is circulated and commonly reinjected, yet losses and makeup requirements still belong in the water plan.

A project can be scientifically promising and still miss a customer’s date. Wells, plant equipment, transmission, and permitting each have their own schedule. The useful question is not whether EGS can theoretically operate in more regions than hydrothermal power. It is whether this site has a characterized resource, tested reservoir, buildable plant, and delivery path aligned with the load’s commissioning schedule.

Request a resource report that distinguishes measured well data from model outputs. Ask which wells are drilled, which have circulated fluid, and which results are projections. A proposed 200 MW field might be built as several blocks; the output and date of each block matter more than a single final-capacity headline. Cape Station’s documented block schedule shows how first power and later capacity can be years apart.

Then compare the project to the load. Is the quoted output gross at the turbine or net after pumping? Does it meet the data center’s peak or only its annual energy target? What covers planned maintenance and a forced outage? Where is power delivered and who holds any transmission service? A grid-delivered PPA may help a company’s energy procurement goals while the physical facility still needs a separate utility connection.

Finally, read the risk allocation. The operator can be responsible for drilling and plant performance while the customer remains responsible for grid service or backup supply. Ask for evidence of induced-seismicity monitoring, water rights and supply, and a commissioning test with measured net megawatts. None of those checks requires a promise that EGS is cheap or that every hot-rock site is buildable.

Do not treat heat-in-place as generation that is ready to buy. The USGS estimate is expressly conditional on successful commercial-scale technology. Do not confuse EGS with a sealed closed-loop design; only EGS depends on engineered permeability in the reservoir.

Do not describe a first-power announcement as full-project commercial operation. A synchronized 33 MW block and a planned 500 MW development are different milestones. And do not assume geothermal electricity arrives at a data center merely because it is contracted: trace the delivery point, transmission service, backup plan, and each block’s acceptance test.