Zanskar just bagged a massive $115 million Series C to do what the entire geothermal industry has been getting wrong for decades: finding untapped conventional geothermal resources that experts think could generate terawatt-scale power. Using AI and modern drilling techniques, the startup has already resurrected a dying power plant in New Mexico and discovered two new sites worth over 100 megawatts. Here's why that matters for the grid.
The startup has a radical thesis: we've been underestimating conventional geothermal by somewhere between one and two orders of magnitude. Zanskar co-founder and CEO Carl Holland told TechCrunch that while the Department of Energy projects 60 gigawatts of geothermal capacity by 2050, that number assumes most gains come from enhanced geothermal - the fracking-based approach pursued by startups like Fervo and Sage Geosystems. But conventional geothermal, which taps naturally fractured hotspots, has been stuck at just 4 gigawatts in the U.S., barely budging a gigawatt higher over the past decade.
The reason? Outdated assumptions about where geothermal resources actually exist. "They underestimated how many undiscovered systems there are, maybe by an order of magnitude or more," Holland said. With modern drilling techniques and AI-powered discovery, "you can get a lot more out of each of them, maybe even an order of magnitude or more. All of a sudden the number goes from tens of gigawatts to what could be a terawatt-scale opportunity."
That bullish outlook just got institutional backing. Zanskar closed a $115 million Series C led by Spring Lane Capital with participation from a deep bench of climate and venture investors including the All Aboard Fund, Carica Sustainable Investments, Clearvision Ventures, GVP Climate, Lowercarbon Capital, Munich Re Ventures, Obvious Ventures, and Union Square Ventures, among others.
The capital came after Zanskar proved its core thesis works. In their previous funding round, the startup explored three geothermal sites - and succeeded at all three. They revived a flagging power plant in New Mexico and discovered two entirely new sites capable of generating over 100 megawatts combined. "Three of three," said Zanskar CTO Joel Edwards. "What does it look like when you try 10?"
The secret sauce is AI applied in two distinct ways. First, Zanskar uses supervised machine learning models trained on historical accidental discoveries - because here's the kicker: about 95% of all geothermal systems have no visible surface indicators like hot springs or volcanoes. We've only found most of them by accident. "We just keep finding them by accident. That is a great application for AI," Holland noted.
Once a promising site is identified, the company's team validates it on the ground. Then comes the hard part: actually developing the resource. That's where Zanskar deployed a more sophisticated AI technique called Bayesian evidential learning, or BEL. It takes existing data, builds assumptions about how the geothermal system works, then stress-tests those hypotheses to spit out probabilities. Where gaps remain, the team fills them with a custom geothermal simulator they've built.
The approach is working because it's combining brute-force machine learning with domain expertise. But the real test comes next. Holland wants to find and confirm at least 10 geothermal sites to unlock project finance - the kind of patient capital that flows to infrastructure, not startups. That's the pathway out of what's been called the "valley of death" that's swallowed countless climate tech companies trying to scale from pilot to production.
With enough MW in the pipeline to support at least a gigawatt of generating capacity, Zanskar is positioned to make that leap. Holland acknowledged the company hasn't solved every challenge in geothermal exploration, but he's betting the AI-first approach changes the game. "We now know this is the future of exploration," he said. "This is going to change geothermal in very short order." The startup is focusing initially on the U.S. West, where the best conventional geothermal resources are concentrated.
Zanskar's $115 million Series C bet signals a major shift in how the energy industry thinks about conventional geothermal power. For years, the sector treated it as a mature, slow-growth business overlooked by venture capital. But Holland and his team are arguing that AI can unlock a trillion-watt opportunity hiding in plain sight - untapped resources scattered across the U.S. West that nobody bothered to look for because they lack obvious surface clues. If Zanskar can deliver on that promise across a dozen sites, it won't just reshape the geothermal sector. It could become the blueprint for how AI accelerates the entire transition to cleaner energy infrastructure.
