How AI Is Speeding Up Geothermal by Weeks

Fervo Energy’s AI-powered drilling slashes costs and time, advancing geothermal as a scalable, 24/7 clean energy source.

13 Jun 2025

At a remote site in Utah, geothermal energy may be getting a long-awaited upgrade. Fervo Energy, a Houston-based start-up, has drilled an ultradeep well in just 16 days, an operation that once took months. Using artificial intelligence to guide its drill through hard rock, Fervo has cut costs nearly in half. Each well now costs $4.8m, down from $9.4m, according to the firm.

This shift in speed and efficiency changes the economics of geothermal energy. Unlike solar and wind, geothermal provides a constant source of power. That makes it attractive to industries and utilities in need of reliable, low-emissions electricity. But until now, high drilling costs and slow development have limited its reach.

Fervo’s breakthrough came at Cape Station, a commercial site under development in Utah. The company’s AI tools process real-time sensor data to adjust drill paths mid-operation, improving accuracy and speed. “We’ve proven we can do it faster, and cheaper,” says Tim Latimer, Fervo’s chief executive. The approach builds on earlier work with X, Alphabet’s moonshot division, though the Utah result is the first to show commercial viability at this scale.

The technology is promising, but not yet a panacea. AI-driven systems rely on large volumes of high-quality data and may perform less consistently in unfamiliar geological settings. Surface infrastructure, especially power conversion and grid integration, still needs investment. Yet Fervo’s success suggests geothermal could finally become more than a niche.

In theory, geothermal energy could supply vast amounts of baseload power in regions with deep underground heat. In practice, it has languished, supplying less than 1% of America’s electricity. But as policymakers look for clean, round-the-clock energy sources, AI may help tip the balance.