UII UPDATE 438 | NOVEMBER 2025

Intelligence Update

Emerging tech: enhanced geothermal systems

EMERGING TECHNOLOGY SERIES

In this report, part of a series on emerging and potentially disruptive technologies that may be deployed in digital infrastructure, Uptime Intelligence assesses the use of enhanced geothermal systems (EGS) to power a facility.

image

Context

Data center operators need reliable low-carbon electricity — either generated on-site or delivered through the grid — to match demand. Common low-carbon energy sources such as wind and solar pose a challenge because their output is intermittent. Underground heat sources, by contrast, do not depend on the weather and can provide continuous (or firm) power where they are available.

Today, geothermal generation provides less than 1% of the world’s power. However, the International Energy Agency (IEA) estimates that improved techniques could increase this to about 11% (around 800 GW, or 6,000 TWh per year) by 2050. In the US, the Department of Energy projects that geothermal capacity could generate 300 GW by this date.

Relatively few locations have natural geological features that make geothermal energy accessible at reasonable drilling costs (see Figure 1). Enhanced geothermal systems (EGS) aim to expand access to affordable geothermal energy in more locations by using hydrofracturing (or fracking) techniques developed for the oil and gas industry.

Figure 1 Predicted suitability for geothermal energy exploitation

image

The technology

The heat of Earth’s interior can be harnessed as an energy source. In the mid-20th century, geothermal power plants were developed at locations with naturally occurring underground steam or hot water. Despite early promise, wider development stalled because conventional geothermal power plants require hot porous rocks that are within 0.6 miles (1 km) of the surface, accessible by drilling or natural vents. Today, most geothermal power plants are in places such as Iceland, New Zealand and California, where these resources exist.

Elsewhere, underground heat is much harder to access. Rock layers hot enough to generate geothermal power are typically found around 2.5 miles (4 km) below the surface and are usually crystalline and dry, meaning they do not naturally generate steam on their own.

Advanced geothermal techniques are being developed, predominantly in the US, to overcome some of these limitations. Current efforts focus on three main areas:

  • Improved exploration methods to find new geothermal sources.
  • Deeper drilling capabilities, extending beyond depths of 1.9 miles (3 km).
  • Hydrofracturing (or fracking) to increase permeability in hot, dry rock formations.

These techniques aim to create reservoirs in otherwise impermeable hot, dry rocks. Water can then be pumped down, heated by the rock and returned to the surface for power generation. Table 1 lists some leading technologies in development.

Table 1 Advanced geothermal technologies

image

In practice, these technologies overlap. Companies may combine elements of EGS, closed loop geothermal systems (CLGS) and geopressured geothermal systems (GGS) where appropriate. Sage Geosystems, for instance, develops GGS projects and also uses fracking techniques, which is typical of EGS.

EGS is currently the most advanced and widely deployed technique (see Figure 2). The US leads in its development, although some European governments are supporting progress through research funding, streamlined permitting and financial de-risking measures. Pilot projects have emerged in Saudi Arabia, the Philippines and Australia.

Figure 2 How EGS generates geothermal power

image

Implementations

Some large-scale data center operators are actively exploring the use of geothermal energy.

Google has signed a 115 MW power purchase agreement (PPA) with Fervo Energy. This deal will help expand Fervo’s 3.5 MW pilot in Nevada, which opened in 2024 with initial support from both Google and the Department of Energy. Fervo also has 320 MW of PPAs with Southern California Edison, supported by Shell’s energy division. This is part of Fervo’s long-term Cape Station project, which is a plan to construct a commercial 2 GW grid-scale plant in Utah by 2026.

Meta has agreed to take between 4 MW and 8 MW in 2027 from Sage Geosystems. This is understood to be a virtual power purchase agreement (VPPA), in which Meta buys energy attribute certificates (EACs) to offset emissions elsewhere, possibly from multiple projects. Sage is building a 3 MW geopressurized energy storage plant for San Miguel Electric Cooperative in Texas and plans a second project (with Department of Energy support). Sage expects to scale up to 150 MW in total by around 2029.

Meta is also supporting a closed-loop geothermal project by XGS Energy in New Mexico. This project is scheduled to supply 150 MW to the PNM grid in 2030.

Microsoft has partnered with Kenya Electricity Generating Company (KenGen) and AI technology group G42 to power a data center with conventional geothermal power in Kenya, as part of a $1 billion digital investment program.

Economics

Commercial geothermal projects in operation today have a levelized cost of energy (LCOE) ranging from $61 to $102 per MWh compared with $68 to $166 for coal and $115 to $221 for gas peaker plants, according to Lazard’s 2025 Levelized Cost of Energy report.

EGS projects are typically more expensive than conventional geothermal due to the costs of fracking and deep drilling. However, the total project costs vary depending on the amount of drilling and fracking required at a given site.

The general costs of EGS projects are falling as drilling speed and efficiency improve, along with advances in borehole topology. For example, Fervo reported a 60% reduction in the cost of its horizontal drilling in 2024. As new techniques emerge and make previously inaccessible resources commercially viable, EGS development is expected to expand into new regions.

The US Department of Energy’s Enhanced Geothermal Shot research program, announced in 2022, aims to make commercial EGS a widespread renewable energy option by reducing its costs to $45 per MWh by 2035. The IEA predicts next-generation geothermal technologies could achieve costs of around $50 per MWh by 2035.

If this is achieved, firm, low-carbon power from EGS could be competitive with grid electricity in some locations, particularly where the existing grid relies on a high-carbon or less reliable sources. That might encourage operators in those areas to invest in the first commercial EGS projects, paying a premium for low-carbon, firm power.

Even with favorable output prices, EGS installations will remain substantial projects that require long-term investments. They will most likely be built by energy operators, with data centers accessing power either through the grid or via on-site energy partners. Even behind-the-meter installations are likely to include grid connections.

Although EGS can open up previously inaccessible geothermal resources, it will not be viable everywhere. US research company Rhodium Group has matched potential EGS developments to data center distribution, estimating that EGS could unlock exploitable geothermal resources in 13 of the top 15 existing US data center hubs. According to the analysis, this could support up to 64% of the new US data center capacity predicted to be in operation by the early 2030s. However, the resource potential is largely concentrated in the Midwest and the West. By contrast, Northern Virginia — the largest data center market in the world — has very low geothermal potential. As a result, many applications may not be practical to host in areas where EGS power is available.

Commercial activity

Globally, governments invested more than $400 million in EGS and other advanced geothermal energy initiatives in 2024, with the US, the Netherlands, Poland and Germany among the largest supporters.

Venture capital investments in advanced geothermal attracted around $790 million in the same year, predominantly in the US. Oil and gas companies are also significant investors; for example, Shell supports Fervo, and Chevron has invested in various geothermal companies via Baseload Capital.

Some data center operators are also expected to play a role in financing advanced geothermal projects, acting as early customers willing to commit to long-term contracts for large amounts of firm low-carbon power.

Drivers and barriers

Drivers:

  • EGS offers firm, low-carbon power that aligns with large data centers’ demand.
  • The technology has low emissions compared with natural gas.
  • Geothermal energy produces much less waste and emissions than other thermal power technologies such as natural gas or nuclear power.
  • EGS uses mature drilling technologies and workforce expertise from the petrochemical industry.

Barriers:

  • EGS remains relatively expensive, and further technological development is still required.
  • Environmental groups have raised concerns over several issues associated with oil and gas fracking, such as water pollution and seismic activity. EGS will likely raise similar concerns, although it reportedly uses fewer chemical additives and has a lower risk of pollution.
  • Geothermal projects require permitting from both local and national governments. The timeline for a project typically ranges from 4 to 8 years, with the process varying between regions.
  • While some systems recirculate water in a closed loop, EGS can consume large amounts of water.
  • Even with EGS, accessible geothermal energy remains localized. Some data center hubs, such as Northern Virginia, have little or no potential, which means that many projects will not be able to use EGS.

The Uptime Intelligence View

EGS is a promising source of firm, low-carbon power. In regions with suitable geology, data center operators have an opportunity to become early customers and invest in the technology as it becomes commercially viable over the coming decade.

Data center operators considering small modular reactors (SMRs) should also investigate the possibility of EGS in their area: EGS projects are likely to reach fruition sooner and face fewer permitting hurdles than SMRs.

While EGS projects are inherently limited to locations where subsurface heat sources can be exploited, advances in the technology are likely to expand the number of locations. Over the next 5 to 10 years, data center operators planning new projects may consider siting them in areas where medium-term EGS power is possible.

Other related reports published by Uptime Institute include:
Enhanced geothermal: long-term clean power — for some 
Small modular reactors: building critical mass

 

About the Author

Peter Judge

Peter Judge

Peter is a Senior Research Analyst at Uptime Intelligence. His expertise includes sustainability, energy efficiency, power and cooling in data centers. He has been a technology journalist for 30 years and has specialized in data centers for the past 10 years.

Posting comments is not available for Network Guests