Last year, the Lightning Dock geothermal power plant in New Mexico was on sale, citing underperformance issues. So in May 2024, geothermal exploration and development company Zanskar swooped in and acquired it, betting that it could turn it into one of the most productive pumped geothermal wells in the U.S.
Carl Hoiland, Zanskar co-founder and CEO, told Latitude Media that, like many geothermal sites across the country, Lighting Dock was initially discovered by accident: A farmer was drilling a well for water, and got steam instead.
“[But] they’d never fully explored and characterized the deeper part of that system, and they were producing only from that zone that was found by accident,” Hoiland said. “And as a result of not fully understanding the system, it had underperformed and was declining over the years.”
When it was put on sale, both the small number of potential bidders and the price that was paid made it clear that no one else in the industry was willing to take on the risk of acquiring it. Zanskar, on the other hand, was ready to do so.
And last month, around one year after the acquisition, the bet seemingly paid off. Zanskar connected the restored 15-megawatt power plant to the grid — and tripled its production.
Relying on AI tools
The reason Zanskar was confident in the plant’s potential was because it had more information than its competitors. The company uses artificial intelligence to streamline the exploration and development process for geothermal, making the discovery, de-risking, and operation of geothermal sites more efficient.
“With geothermal, you have this unique challenge of understanding what’s in the subsurface,” Hoiland said. “Historically, this has been a very bespoke process. Every project might hire some geologists, and they’ll have to make unique decisions.”
Zanskar makes that process faster and cheaper by augmenting or automating these human decisions with AI tools trained on vast amounts of geophysical and geologic data sets, such as the Earth’s gravitational and magnetic fields, the rock types, the fault lines, the location of previously discovered sites, and drilling data. (Of these, drilling data is the hardest and most expensive to get.)
These tools allow Zanskar to look at a large area and estimate the location of geothermal resources before drilling, significantly reducing the risk and cost of the overall operation. Once the estimates are in, Zanskar goes out into the field to drill, collecting more data to improve the accuracy of its AI tools along the way.
“We’ve gone out and tested enough of these sites to show that this is repeatable, and we’ve now discovered more of these new hidden geothermal resources than any company in history,” Hoiland said.
In Lightning Dock’s case, rather than facilitating the discovery of a new site, the tools allowed Zanskar to evaluate the site quickly, estimate that it had a larger geothermal resource potential than previously tapped into, and drill a new well that generates over 100 megawatts of thermal energy.
“That’s about the size of some small SMRs, from a single well that you can wrap your arms around” Hoiland said. (After efficiency conversion, the well produces about 15 MW of net electric generation.)
For more on how geothermal power plants get built, listen to Carl Hoiland’s recent interview on Catalyst with host Shayle Kann:
Looking for project finance
Lightning Dock’s grid connection showcases Zanskar’s AI tools potential and is a big milestone for the company. But going forward, the startup will mostly focus on discovering and developing new sites, rather than acquiring old ones, because supply is limited; there are only about 36 operating geothermal fields in the entire U.S.
Despite having been around for decades, and having enormous potential, up until recently geothermal was a stalling technology, with “a long history of failures that has really plagued the industry,” Hoiland said.
But in recent years, thanks both to the skyrocketing demand for baseload power spurred by AI and to the introduction of next-gen technologies like enhanced geothermal systems, it’s attracting investments and attention once again.
Zanskar conducted a stealth seed round led by Azolla Ventures’ Prime Impact Fund in 2021, and has since raised a Series A in 2022 and a Series B in 2024. But the company hopes that by showing it can de-risk a proven technology like conventional geothermal, it will soon mature to the point that it can attract project finance to develop its new power plants.
“Banks and financial groups want to see a new track record of successes, even for conventional geothermal,” Hoiland said, “And we’ve been building this track record for several years on the early stage discovery. These new tools are making it significantly cheaper and repeatable. Lightning Dock is not just a one-off, it’s repeatable. It’s predictable. And this is the type of data that we’re taking to the banks for the financing of future projects.”
