Particle.news
Download on the App Store

Fervo, NVIDIA and PNNL Announce AI Digital Twin to Model Geothermal Reservoirs

The tool aims to cut drilling risk by blending Fervo field data with PNNL physics models plus NVIDIA GPU acceleration using DOE supercomputers for a target rollout in 2029.

Overview

  • A deal announced Monday between Fervo Energy, NVIDIA and Pacific Northwest National Laboratory creates EGS‑Twin, a digital twin platform for enhanced geothermal systems.
  • PNNL will begin immediate training of scalable AI models on Fervo’s proprietary Nevada and Utah field data using NVIDIA AI infrastructure and U.S. Department of Energy supercomputing resources.
  • EGS‑Twin will pair high‑resolution field telemetry with high‑fidelity physics simulations and AI forecasting to infer subsurface fractures and forecast reservoir responses in near real time.
  • The partners say the platform is designed to reduce subsurface uncertainty, lower drilling and capital risk, and speed deployment of geothermal power by improving site decisions and reservoir management.
  • The project is a multi‑year, data‑ and compute‑intensive effort that leans on Fervo’s recent operational data from Project Red and Cape Station, and it faces technical, regulatory and commercial risks before planned implementation in 2029.