Innovative AI Methods Save Geothermal Power Plant in New Mexico from Closure

As MIT Technology Review reports (https://www.technologyreview.com/2026/07/29/1140896/geothermal-second-chance/), a small company named Zanskar took over a geothermal power plant in New Mexico in June 2024 that was on the verge of economic failure due to declining underground water temperatures. Within two years, they managed to bring the plant back to full capacity using advanced Artificial Intelligence (AI).

Challenge: Cooling of the Geothermal Reservoir

The main problem of the plant was that the water from the underground reservoir was becoming increasingly colder. This drastically reduced the efficiency of energy production and made operation uneconomical. Traditional methods to revive such plants are often expensive, time-consuming, and carry high risks.

AI-Supported Optimization of Resource Use

Zanskar relied on innovative AI algorithms to better understand and control the complex geological and thermodynamic processes. The AI continuously analyzed data from drilling, temperature measurements, and pressure sensors to develop optimal extraction strategies. This allowed the plant to be precisely managed to maximize heat efficiency and sustainably preserve the reservoir.

Integration of Machine Learning and Simulations

By using machine learning combined with physical simulations, the company was able to make predictions about the reservoir’s development and make early adjustments. This prevented further cooling and enabled stable energy production.

Importance for the Energy Transition

The successful revival of the plant demonstrates how AI technologies can help make renewable energy sources more efficient and economical. Geothermal energy is considered a particularly sustainable energy source because it can provide constant electricity regardless of weather conditions. The combination of AI and geothermal energy could thus make an important contribution to reducing CO2 emissions.

Outlook and Potential

The example from New Mexico could serve as a model for similar plants worldwide that have been shut down or are close to shutdown due to technical or natural challenges. Integrating AI into the control and optimization of geothermal power plants opens new possibilities to tap previously unused resources and extend the lifespan of existing plants.

Conclusion

The recommissioning of the geothermal power plant by Zanskar underscores the potential of AI applications in the field of renewable energies. Data-driven optimization can better align ecological and economic goals, which is of great importance given the global climate crisis.