PNNL, Fervo Build 2029 Geothermal Digital Twin Using NVIDIA AI
Updated
Updated · Renewable Energy Magazine · Jun 24
PNNL, Fervo Build 2029 Geothermal Digital Twin Using NVIDIA AI
3 articles · Updated · Renewable Energy Magazine · Jun 24
Summary
PNNL and Fervo said the EGS-Twin platform will start training immediately on proprietary field data from Nevada and Utah, with implementation scheduled by 2029.
High-resolution field data will be combined with physics-based modeling and AI forecasting on NVIDIA AI infrastructure, then integrated into NVIDIA Omniverse libraries.
PNNL is building the workflows and data pipelines with high-performance computing and U.S. Department of Energy supercomputing resources to run large-scale simulations.
Fervo said the digital twin should help operators detect subsurface changes faster, optimize power generation, and improve reservoir management, heat recovery, and system reliability for enhanced geothermal systems.
With its digital twin, can Fervo's new Utah plant prove geothermal is profitable?
Will this digital twin make geothermal energy a reliable backbone for the US grid?
Can AI finally tame the unpredictable underground to unlock vast geothermal power?
Revolutionizing Geothermal: EGS-Twin’s AI Digital Twin Platform Aims for 500 MW by 2029
Overview
Announced on June 24, 2026, the EGS-Twin initiative is a groundbreaking collaboration between Pacific Northwest National Laboratory, Fervo Energy, and NVIDIA. Together, they are developing a cutting-edge digital twin platform designed to accelerate geothermal energy development. This platform gives operators advanced tools to manage and optimize their operations, enabling them to identify subsurface changes more quickly, improve power output, and scale geothermal projects more efficiently. By leveraging advanced computing and real-time data, EGS-Twin aims to transform geothermal energy into a more reliable and scalable clean power source for the future.