XGS Energy: What's Actually Built vs. the 3 GW on Paper

XGS Energy has one demonstrated result and a long list of announcements, and the two are easy to mix up because the company's own materials present them side by side. The demonstrated result: a water-independent geothermal system that ran for more than 3,000 hours in a workover of a long-idle well at California's Coso geothermal field, holding a temperature difference between injection and production that XGS reports as "greater than 158˚F (70˚C)." The announcements: 150 megawatts planned in New Mexico with Meta and PNM, 115 MW under an agreement with a California power buyer, an engineering order from Baker Hughes, a supply deal with Vallourec, and what XGS describes as a three-gigawatt project pipeline.

None of those megawatts exist yet. No source we reached establishes any XGS-owned or XGS-operated generating capacity, or any power delivered under the Meta, PNM or CC Power agreements.

What XGS Energy Is

XGS Energy is a privately held geothermal developer, headquartered in Houston since 2026, whose pitch is aimed at the places conventional geothermal cannot go. Its own description: "Our proprietary solid-state geothermal system uses thermally conductive materials to deliver affordable energy anywhere there is hot rock." Its stated mission: "We are decoupling geothermal from dependence on water and geology."

Strip the framing and the business case is this: conventional geothermal needs naturally occurring hot water underground, and even newer approaches typically depend on how permeable the rock is or how much water a site can spare. XGS builds a sealed system that, in the company's words, involves "no consumptive water use in operations" and "no hydraulic stimulation or fracking processes required." If that holds up at commercial scale — which has not yet been shown — it would widen the map of viable geothermal sites considerably.

How the Technology Works

A pipe inside a pipe, in a single vertical well

Per XGS's own description, its "smallest modular unit is a single well, pipe-in-pipe heat exchanger, in which working fluid, typically water, is circulated inside a steel casing, warms up, and then is returned to the surface through an inner insulated tube." XGS does not publish the detailed flow path or heat-transfer geometry beyond that description.

That water never touches the rock. As XGS puts it, "water is continuously recycled in zero-loss closed-loop." The working fluid is sealed inside steel for its entire journey — which is what the company means by "water-independent." A site does not need an aquifer, and operations do not consume water.

Thermal Reach Enhancement: the differentiator

A sealed pipe has an obvious physics problem: rock is a poor conductor, so a closed well can only pull heat from the rock immediately around it. XGS's answer is what it calls Thermal Reach Enhancement, or TRE. The Journal of Petroleum Technology, published by the Society of Petroleum Engineers, describes it as "a highly-conductive material that is pumped downhole as a liquid slurry through perforations in the well casing to contact hot rocks." Heatmap News describes the same thing from the outside: XGS "fills the gap between the metal and the rock with a patented slurry that conducts heat."

The slurry acts as a thermal bridge, extending the well's effective reach into the surrounding hot rock. Neither XGS nor any independent source we found discloses what the material is made of — "highly-conductive" and "patented" is as specific as the public record gets. XGS labels the system "solid-state" and says it "uses thermally conductive materials"; the company does not spell out the connection further, but the mechanism it describes is heat moving to the well by conduction through solid material rather than by hot fluid flowing through the rock.

Why "not dependent on permeability" matters

XGS lists as a design feature that "performance is not dependent on rock permeability and porosity." Conventional plants need rock that naturally lets hot fluid through; enhanced geothermal systems fracture the rock to create that permeability. A conductive closed loop needs neither — just hot rock. XGS's boilerplate claims the approach "decouples geothermal energy production from its historical dependence on natural water resources and geological conditions, unlocking low-risk power project development, geographic flexibility, easier permitting, and faster deployment." That is the company's language; every clause of it is a forward-looking claim.

Where XGS Fits in the Geothermal Family

A 2024 document submitted to a U.S. congressional hearing sorts next-generation geothermal into three buckets, and it is the most authoritative sorting we found:

• Conventional geothermal — heat extraction by "production of naturally occuring hydrothermal fluid." This is the category today's conventional geothermal power plants fall into: find hot water underground, bring it up, run a turbine.
• Enhanced geothermal systems (EGS) — "circulating water through fractures (either existing, created, or enhanced through stimulation) in hot rock." The document's named examples: Fervo, Utah FORGE, Mazama. (Our EGS explainer covers this category in depth.)
• Closed-loop geothermal — "circulating water through continuous boreholes/underground heat exchanger in hot rock." The document's named examples: Eavor, GreenFire, XGS.

So XGS sits in the closed-loop bucket, alongside Eavor and GreenFire — three separate companies the congressional document lists as distinct examples of the same category. Beyond that shared classification, the public record we verified does not support a detailed architecture-by-architecture comparison between XGS and Eavor.

The contrast with Fervo is better documented. Heatmap News, reporting on the Coso demonstration in September 2025, drew it directly: "Unlike Fervo, XGS doesn't use fracking technology to drill horizontal wells in pursuit of hot, dry rocks from which to harvest energy. Instead, XGS drills vertical wells and inserts a closed steel pipe with water and fills the gap between the metal and the rock with a patented slurry that conducts heat." And where "technology like Fervo's requires pumping cold water over the fractured hot rocks to harvest heat," the XGS working fluid stays sealed in its pipe. Vertical wells, no fracking, no fluid-rock contact — that is the dividing line between the two approaches.

What Has Actually Been Built

XGS's field record, as of this writing: one demonstration, at one well, at the Coso geothermal field in California's Western Mojave Desert region.

Coso is an established geothermal field owned by Atlantica Sustainable Infrastructure and operated by Coso Operating Company, which — per Heatmap News — "runs a 270-megawatt geothermal power plant on the land." That 270 MW is Coso Operating Company's plant and has nothing to do with XGS's capacity. What XGS did, in its own words, was design and execute "a workover program on a well that had been idle in the Coso geothermal field for over two decades." An idle well on someone else's producing field, brought back into service as a test bed.

What happened in that well, per the company's September 30, 2025 announcement:

• The system operated "at commercial scale for over 3000 hours" — a bit over four months of continuous run time.
• Flow testing sustained "a difference between injection and production temperatures greater than 158˚F (70˚C)" — the water returned hotter than it went down, and held that gap for the duration.
• "The extensive operating history was used to verify XGS's production models, with actual performance falling within 2% of predicted values."
• "Placement of TRE in the well was completed in less than 24 hours" — a first data point suggesting the slurry step doesn't blow up the drilling schedule.

The Coso work was not a power plant, and no source reports it adding generating capacity to anyone's grid or selling electricity. It was a demonstration — a long one, with models that held — in a reworked well on a field that belongs to somebody else.

Announced, But Not Built

Everything else attached to XGS's name is an agreement, an order, or a projection.

Per the Baker Hughes announcement, the New Mexico project "will support the delivery of clean, round-the-clock power to the Public Service Company of New Mexico's (PNM) grid in support of Meta's data center operations in the state." Canary Media reported in August 2026 that New Mexico "is backing the commercial deployment of a novel geothermal system through its $75 billion sovereign wealth fund," supporting XGS through the fund's venture-capital investments as the company "develops a 150-MW closed-loop system in New Mexico along with Meta."

The California counterparty, CC Power, is "a joint powers authority comprised of nine California CCAs" whose members "represent 2.7 million customers across 112 municipalities."

And the "three gigawatt pipeline" deserves its scare quotes. A pipeline, in developer language, is a list of prospects at every stage from signed agreement to hopeful spreadsheet. Three gigawatts is twenty times the announced New Mexico project. Nothing in the public record breaks down what stage those prospective gigawatts are at, so the figure tells you about XGS's ambition, not its order book.

There is one recent signal that the announced column may be starting to move: a July 22, 2026 XGS release on new operations executives states it comes "as XGS begins construction of multiple projects across" its footprint. Construction beginning is not power flowing — but it is the next checkable step.

Why the Market Cares

The first demand story is California's arithmetic, laid out in XGS's own CC Power announcement: "California sits atop more than 35 gigawatts (GW) of untapped geothermal potential as identified by the Clean Air Task Force in a 2025 report. Yet, geothermal plants represent only 2.7 GW of over 89 GW total installed electric generation in the state today." Those are XGS's figures, quoting a third party, in a document arguing its own case — but the gap between estimated potential and built capacity is what the whole next-generation sector is pitching against. Whether closed-loop systems can run without consumptive water use at scale, in a state where water politics can stall a project, is still the open question.

The second is data centers. The anchor customer for XGS's largest announced project is Meta, and the promised product is round-the-clock power for data center operations — the same always-on demand profile behind the data-centre interest in geothermal. We cover that broader collision of compute demand and geothermal supply in our geothermal and data centers explainer; XGS's New Mexico agreement is one entry in that pattern, distinguished mainly by its closed-loop technology and its 2030 target date.

Who Runs XGS

As of September 2026, XGS's team page lists the following C-suite:

• Josh Prueher — Chief Executive Officer
• Ghazal Izadi, Ph.D. — Chief Growth Officer
• Axel-Pierre Bois, Ph.D. — Chief Technology Officer
• Lucy Darago, Ph.D. — Chief Commercial Officer
• Al Vickers — Chief Operations Officer
• Kurt Fricker — Chief Procurement Officer
• Richard Chong — Chief Financial Officer

Earlier company press releases quote Ghazal Izadi as "XGS Chief Operating Officer" in both the September 2025 Coso announcement and the March 2026 Baker Hughes announcement. That is not an error in either document — the titles changed in mid-2026. A Houston Business Journal item dated July 23, 2026, carried on XGS's own site, reports that XGS "expanded its C-suite with a new COO, chief procurement officer and chief growth officer just months after adding a CFO," following the company's headquarters move to Houston. Izadi now holds the growth title; Al Vickers holds operations.

The board includes Martin S. Craighead, former chairman and CEO of Baker Hughes.

Can You Invest in XGS Energy?

Not on any stock exchange. XGS is privately held; there is no ticker and no publicly traded share.

Its capital comes from venture and strategic investors. The company's own boilerplate names them: Anzu Partners, Aligned Climate Capital, BlueScopeX, B Current Impact Investment, ClearSky, ClimateIC, Constellation Technology Ventures, H&P, MIH Capital, Thin Line Capital, Valo Ventures, VoLo Earth Ventures, and WovenEarth Ventures. No reliable cumulative funding total is public, and no valuation is public, so we publish neither.

The company's capital raising surfaces as Form D filings on SEC EDGAR.

What to Watch Next

The near-term signals to track, in rough order of importance:

1. Steel in the ground in New Mexico. The July 2026 release says construction of multiple projects is beginning. Watch for site-specific confirmation — drilling underway on the Meta/PNM project — and for whether the first, smaller phase of the two-phase 150 MW plan holds its "operational by 2030" projection.
2. First commercial megawatt-hours, anywhere. The single number that moves XGS from the announced column to the built column. No such figure has been published.
3. CC Power's GEODE agreement advancing from exploration and development toward a definitive offtake — public power authorities document their steps, so this will be visible.
4. A second demonstration or third-party validation. The Coso demonstration was one reworked well on one field. A repeat performance in different rock would say more about "anywhere there is hot rock" than any press release.
5. New Form D filings on EDGAR, which would signal fresh capital for the construction phase.