AI Powers Geothermal Future
Fervo, Nvidia, PNNL Partner to Launch AI Digital Twin for Enhanced Geothermal Systems
Key Points
EGS-Twin will integrate the trained AI models into NVIDIA’s Omniverse libraries, aiming to help geothermal operators “more quickly identify and respond to subsurface changes” and improve scalability of enhanced geothermal systems.
Fervo CTO and co-founder Jack Norbeck said: “We believe that digital twins will expedite the learning curve for geothermal development as we build and operate our GeoBlock assets,” adding that “Integrating high-fidelity physics-based models with AI-driven forecasting has the potential to reshape reservoir management, improve heat recovery, and enhance system reliability.”
Fervo’s Project Red has already delivered 3.5 MW of enhanced geothermal power (since November 2023), supplying clean energy to Google data centers—providing real-world operating data behind the initiative.
The article notes Nvidia is a founding member of the Enhanced Geothermal Systems Deployment Coalition, which counts Fervo among its members—placing EGS-Twin within a broader industry-government push for geothermal scale-up.
Fervo Energy, Nvidia, and the Pacific Northwest National Laboratory have announced EGS-Twin, an AI-powered digital twin platform for enhanced geothermal systems. The platform will simulate underground reservoir conditions in real time, helping operators drill smarter and generate more power. MarketScreener reported the deal on June 22, 2026.
Fervo Energy shares jumped more than 11% after the news broke, according to Investing.com. The platform is targeted for full deployment by 2029 and will draw on Fervo's existing field data from sites in Nevada and Utah.
What EGS-Twin Does and How It Works
Enhanced geothermal systems work by drilling into hot rock, then pumping water through to create artificial underground reservoirs. The problem has always been visibility — operators can't easily see what's happening a mile underground. EGS-Twin aims to fix that by creating a virtual copy of the reservoir that updates in real time.
PNNL will build the data pipelines and run large-scale simulations using U.S. Department of Energy supercomputing resources. Nvidia will supply GPU and AI infrastructure and integrate the models into its Omniverse platform. The goal, according to MarketScreener, is to help operators "more quickly identify and respond to subsurface changes" and improve the scalability of geothermal energy.
Fervo's Real-World Data Gives the AI a Head Start
PNNL will begin training the AI models using Fervo's proprietary field data collected at Project Red in Nevada. That site has delivered 3.5 MW of enhanced geothermal power to the grid since November 2023, supplying clean energy to Google data centers. It reached drilling temperatures of 191°C, setting a U.S. record at the time.
Fervo CTO and co-founder Jack Norbeck said the platform will "expedite the learning curve for geothermal development" as the company builds out its GeoBlock assets. He added that combining physics-based models with AI forecasting "has the potential to reshape reservoir management, improve heat recovery, and enhance system reliability."
Nvidia and PNNL Bring AI Muscle to the Project
Nvidia is not new to geothermal. The chip company is a founding member of the Enhanced Geothermal Systems Deployment Coalition, which also counts Fervo as a member, according to Yahoo Finance. That coalition brings together energy companies and the U.S. government to push geothermal energy to scale.
PNNL earth scientist Maruti Mudunuru, the project's principal investigator, said EGS-Twin will allow operators to "understand, in real time, the dynamics of their reservoir and act quickly to maximize the power generation potential." The lab will refine its models continuously as more production data comes in from Fervo's sites.
Big Targets: 500 MW and a 90% Cost Drop by 2035
The stakes are high. The DOE's "Enhanced Geothermal Shot" aims to cut the cost of geothermal power by 90% to $45 per megawatt-hour by 2035. Fervo's Cape Station project in Utah is targeting 100 MW of capacity by late 2026 or 2027, with an ultimate goal of 500 MW. The DOE estimates the U.S. could reach 120 GW of geothermal capacity by 2050.
Fervo carries a market cap of roughly $10.32 billion as of June 22, according to Yahoo Finance. Analysts project nearly 50% revenue growth as the company moves from pilot projects to utility-scale generation. However, some analysts caution the stock may already be trading above fair value given its early-stage finances.
Sentiment
Mostly PositivePublishers
16
Articles
38
Reach
54
