Who's Really Buying Up the West's Geothermal Land - and What It Means for Next-Gen Power

Something unusual is happening at federal geothermal lease auctions across the West. Bidders are showing up with deep pockets, targeting only the best acreage, and refusing to say who they work for.

According to analysis by market intelligence firm Currence, anonymous land agents now account for nearly half of all federal geothermal lease spending over the last decade^[1]. They don't bid at the minimum. They don't buy marginal acres. And they pay, on average, roughly twice the per-acre price of any other bidder category^[1]. That kind of discipline - and that kind of money - doesn't come from startups.

The June New Mexico Sale Is the Clearest Signal Yet

The mid-June Bureau of Land Management sale in New Mexico put the trend in sharp relief. The auction netted more than $16.5 million, making it the second-highest-grossing federal geothermal sale in BLM history, with a record top bid of $701 per acre. Forty-seven parcels sold at an average of $107 per acre - more than triple the $32-per-acre average recorded in Nevada's October 2025 sale. And only 17% of parcels sold at the $2-per-acre minimum bid, down from 76% in 2019, as competition intensified for sites with genuine development potential.

Two names landed at the top of that auction that no one in the geothermal industry had seen before. Rock Canyon Resources and Novo Mesa - both first-time bidders - spent $5.3 million on 21,000 acres and $3.7 million on 44,000 acres, respectively^[1]. Rock Canyon Resources also set a single-tract record, paying $3.14 million for one 4,479-acre parcel.

Neither company has a public project history. Neither has announced a development plan. That's the point.

Who's Behind the Curtain

Individual land agents can represent a range of clients, and Currence senior research associate Guy Cohen is careful about the uncertainty^[1]. But the bidding behavior - premium targeting, no minimum bids, consistent above-market pricing - points toward players with long time horizons and large balance sheets.

Buffalo River Minerals, which has been bidding in federal geothermal auctions since 2024, is widely believed in the industry to be purchasing on behalf of Chevron^[1]. That would make sense: oil and gas majors have the drilling expertise, the capital, and the strategic incentive to hold geothermal options as the energy transition accelerates. EGS technologies adapt and use processes originally developed for the oil and gas industry to access deeper geothermal resources, which means the technical overlap is real, not just theoretical.

Data center developers are another plausible client category^[1]. The geography lines up: geothermal development is heavily concentrated in Nevada, Utah, Texas, and New Mexico - the same region seeing rapid data center campus growth. Hyperscalers that have already signed geothermal PPAs understand the value of controlling land position before it becomes competitive.

The Established Players Are Moving Too

The anonymous agents aren't operating in a vacuum. Established renewables developers have been staking out enhanced geothermal positions alongside them - and the scale of some of those positions is striking.

Hexagon Energy holds more than 430,000 acres across state, federal, and private land, making it second only to Fervo in total geothermal leasing footprint^[1]. Invenergy ranks fourth on that list, behind geothermal giant Ormat^[1]. These are not geothermal specialists. They are large-scale renewables developers who have decided that next-gen geothermal is worth holding an option on.

Cohen's read on why is direct: federal opposition to wind and solar has pushed developers to look elsewhere for growth^[1]. "Next-gen geothermal looks like it might be there, and they want the option in case it does, and so both have been buying up a lot of land," he said^[1].

Fervo, the company that has done the most to prove enhanced geothermal at commercial scale, sits at the top of the leasing table. Its IPO filing disclosed 595,900 acres of leased geothermal acreage across seven Western states, assembled at a weighted-average cost of approximately $4 per acre between 2019 and 2021 - a figure that looks very different against today's auction prices. Fervo's 500 MW Cape Station project in Utah is under construction, fully contracted, with power slated for delivery to Southern California Edison, and is expected to begin delivering electricity to the grid in 2026.

The Startup Problem

Trailing behind the agents and the big renewables developers is a long tail of earlier-stage geothermal companies. Cohen's assessment of their land position is blunt: "A lot of the early-stage geothermal companies...they're buying crap, they're buying stuff no-one else wants," he said^[1].

That's partly a cash problem. Startups simply don't have the capital to compete for premium acreage at auction, or to carry the ongoing lease costs that come with it^[1]. But it's not entirely a disadvantage. For companies pursuing very deep, ultra-hot rock approaches - where the surface geology matters less than what's miles underground - lower-quality surface acreage may work just as well^[1].

The more interesting hedge is AI. Zanskar, which raised a $115 million Series C in January 2026, has built its entire exploration model around AI-driven surface analysis. Zanskar uses custom AI tools and geoscience models trained on subsurface datasets, satellite imagery, and remote sensing data to pinpoint geothermal resources before drilling - including sites with no surface indicators whatsoever. In December 2025, the company announced the discovery of "Big Blind," the first commercial blind geothermal well discovered in 30 years, found entirely through AI modeling without exploratory drilling.

The logic is straightforward: if AI modeling can reliably identify viable resources from surface data alone, the cost of exploration drops dramatically, and the premium on pre-validated federal acreage shrinks. That's the bet Cohen is tracking when he describes "claimed improvements in AI modeling for the surface" as an emerging trend^[1].

The word "claimed" is doing real work there. Zanskar has validated its approach at Pumpernickel in Nevada - drilling a 137°C well at 760 meters depth that confirmed the AI-predicted resource - but the approach is still being proved at scale. A six-plant execution plan over three to four years is a pipeline, not a track record.

What the Leasing Data Actually Tells Grid Planners

The bidding patterns, taken together, point to a few conclusions that matter for anyone trying to forecast near-term geothermal capacity.

First, demand for next-gen geothermal is stronger than public project announcements suggest^[1]. The lease auction is a leading indicator. Projects get announced years after land is secured.

Second, a growing share of that capacity is likely to be developed by incumbents - oil and gas majors, large renewables developers, and potentially hyperscalers - rather than by the geothermal startups that have driven most of the technology narrative^[1]. Deep-pocketed players can hold land, wait for the technology to mature, and move when the economics are right. Startups generally cannot.

Third, the auction price signal is real. The DOE estimates total domestic EGS resources at 5,157 gigawatts of electrical capacity - roughly 450% of current U.S. generating capacity. The land rush happening in BLM auctions right now is a bet that a meaningful fraction of that resource will be developed commercially. The bidders paying $701 per acre think the odds are good enough to act on.

What they don't have yet - none of them - is a signed PPA, a construction permit, or a financing close on the acreage they're buying. That's the gap between the leasing story and the capacity story. The land rush is real. The projects are not yet.