Grid stress and utility shifts: is geothermal ready?

Power lines in California near the Nevada state line (source: flickr/ Bill & Vicki T, creative commons)

Alexander Richter 20 Jun 2025

The structure of the U.S. power sector is going through a fundamental shift. The rapid growth of AI-driven data centers is leading to a surge in electricity demand—demand that is now starting to outpace grid capacity and is putting pressure on traditional utility business models. It is therefore not surprising to see an increasing number of articles and studies addressing this development.

According to the latest Horizons report by Wood Mackenzie, more than 130 GW of proposed data center capacity has now entered the development pipeline. This is more than double the capacity reported just one year ago. What is particularly noteworthy—based on Wood Mackenzie’s analysis—is that this growth is not limited to established tech hubs. It is now also expanding to other parts of the country, including Pennsylvania, Indiana, and Ohio. These are regions where local power systems are often less prepared for this scale of new load.

For utilities, keeping pace with this trend is becoming increasingly difficult. They are caught between multiple pressures: ensuring grid reliability, avoiding stranded investments, and dealing with regulatory uncertainty. In regulated markets, some utilities are starting to respond with new large-load tariffs and faster generation planning. But in deregulated areas, where market prices are expected to drive investment, forward price signals are currently not sufficient to trigger new capacity. As a result, the gap between what data centers are planning and what the grid can provide continues to widen—raising concerns around system stability, financial risk, and who will ultimately pay for the necessary infrastructure.

In this context, geothermal energy – particularly next-generation solutions like enhanced geothermal systems (EGS) – is receiving new attention as a possible clean and firm power source. A March 2025 study by the Rhodium Group, which we covered earlier this year, suggests that geothermal could supply a significant share of future electricity demand from data centers. In a scenario where these facilities continue to cluster around traditional infrastructure hubs, geothermal could meet up to 64% of the new load by the early 2030s. If data centers were to locate closer to strong geothermal resources, this figure could rise to 100%, with significantly lower electricity costs compared to grid-connected supply.

Geothermal is especially well-suited for this type of application due to several key characteristics: high capacity factors, constant availability, and the possibility to operate behind the meter. This last point is important – especially as delays in grid interconnection and regional congestion are becoming more frequent. Being able to bypass the grid and deliver electricity directly to data center operations presents a clear strategic advantage. At the same time, geothermal plants have a relatively small land footprint, which is attractive for developers seeking reliable and space-efficient solutions.

However, there are still barriers to overcome. Permitting processes – particularly on federal land – remain long and complex. How recent changes by the new U.S. administration will indeed on a broader scale help speed up those processes will have to be seen, yet recent news are proimising in that direction. In many states, legal clarity around subsurface heat rights is still lacking. And while investment into geothermal drilling and development is increasing, the supply chain and workforce are only beginning to scale. Furthermore, the independent power producer (IPP) model that most geothermal developers follow may face regulatory and contractual challenges, especially when entering procurement systems designed with other technologies in mind.

A further question that often comes up in discussions is whether data center companies will actually invest in power generation directly – or if they will remain purely electricity buyers. In the past, most preferred to rely on utility partnerships and renewable energy credits. But this could be changing. Companies such as Google are engaging beyond power purchase agreements, but how and if they are actively investing in clean energy supply, including geothermal will have to be seen. In select cases, there are early signals that hyperscale developers may enter into direct contracts with IPPs, especially if the projects can deliver firm, hourly-matched clean power that aligns with corporate climate targets.

This could open new opportunities for geothermal developers – particularly those able to offer behind-the-meter or co-located solutions. It would, however, also require new forms of collaboration and risk-sharing between project developers, investors, and corporate buyers – something that will not happen automatically, but needs to be designed and built.

As Wood Mackenzie points out, this is not only a matter of energy supply – it is a question of how the electricity market is structured. If utilities and competitive power markets are not able to deliver at the speed and scale that is now required, other models will have to emerge. Geothermal may not have been developed with AI infrastructure in mind, but its characteristics make it a compelling option for addressing the combined challenges of reliability, emissions, and load concentration. Whether the sector can take on this role will depend on more than technology – it will require faster permitting, clearer regulation, willingness to partner with new types of buyers, and a much more ambitious approach to scaling deployment in the years ahead.

Sources: Wood Mackenzie Horizons White Paper – June 2025, Rhodium Group Report – March 2025