The Market Map · Last reviewed August 2026
From the companies drilling the rock to the funds financing them and the hyperscalers buying the power. Geothermal power generation only; direct-use heat and ground-source heat pumps are out of scope. Profiles are compressed on purpose: enough to know who matters and why, sourced as of July 2026 and reviewed August 2026.
01 / Operators & Developers
EGS fractures hot rock and circulates fluid through it, which is the shale playbook. Geopressured systems heat fluid in place using the earth's own pressure, and can run as a subsurface battery. Closed-loop seals the working fluid in a buried radiator: no fracturing, no induced seismicity, lower power density. AI-enhanced discovery skips reservoir engineering entirely and finds conventional resources nobody could see. Superhot rock chases >374°C supercritical conditions for 5–10x energy per well. No analyst has declared a winner.
Taxonomy follows True North Institute, Next-Generation Geothermal Energy, June 2026 (PDF, 8MB)
The sector's proof case. Horizontal drilling + multi-stage fracturing from shale, plus fiber-optic reservoir management. Cape Station (Utah): 100 MW in 2026, 400 MW more by 2028. First and only EGS developer to close non-recourse project debt ($421M), then IPO'd May 2026 at $27/share, closing its debut above a $10.2B market value. In July it drilled Sawtooth 7 in 21 days, a 143% drilling-rate improvement since its first Cape Station well.
Two developments since. Cape Station had a contained well blowout on 27 May, reported without injuries or schedule impact. The share price closed at $22.54 on 31 July, below the IPO.
~658 MW binding PPAs · ~$7.2B potential revenue backlog · 3 GW Google framework
The declared fast follower. The thesis is that Fervo has absorbed the first-mover technology risk and a disciplined second entrant can now build on proven EGS viability. Salt Lake City team combining O&G operators, geothermal developers, and capital formation. Already filing comments to shape Nevada's well-stimulation rules before it has a project to permit.
~$140M Series A reported, unconfirmed · pursuing a hyperscaler PPA
Pioneered field-scale EGS stimulation at Newberry a decade before Fervo. Usually told as a cautionary tale, which is only half right: in 2023 AltaRock combined with Blade Energy Partners to form Mazama, and its Newberry work went on to set the global EGS temperature record. It had the resource and the science; what it lacked was drilling execution.
Newberry demo since ~2010 · now the JV parent of Mazama Energy
Stanford-lineage stimulation specialist turning operator. It pairs proprietary high-temperature fracturing IP with a site at Salt Cove, Utah, next door to the DOE FORGE lab, targeting a first-of-a-kind 100 MWe project at ~275°C. Its $40–60/MWh target undercuts the consensus 2035 figure for all of EGS, from a company with no operating record. Early; no third-party analysis yet.
Target LCOE $40–60/MWh · claims TRL 8 stimulation demo
Its own category, not an EGS variant, and the distinction is the point. Sage's HeatRoot architecture fractures downward using heavy fluids and the earth's lithostatic pressure, reaching higher temperatures without paying to drill to them. The same wellbore runs as a subsurface battery, charging on cheap power and discharging at peak. Houston team led by ex-Shell executive Cindy Taff.
Ormat co-led the $97M Series B and licensed the technology, which puts an incumbent with 1.2 GW of operating assets behind Sage's tech inside its own plant. Two claims still lack third-party validation: the 70–75% round-trip efficiency, and 2 km of downward fracture propagation, independently confirmed only to several hundred feet.
3 MW live on ERCOT · 150 MW Meta LOI · first Ormat-sited plant late 2026–early 2027
Formerly labelled "AGS." The rename matters: closed-loop is a specific architecture, not a catch-all for everything that is not EGS. First-of-a-kind LCOE runs $120–180/MWh, roughly 1.7–2.6x the competing target, because compensating for slow conductive heat transfer means drilling a lot of lateral.
The closed-loop proof point, and its cautionary tale in the same asset. The multilateral Eavor-Loop at Geretsried, Germany has fed the grid since December 2025, the first closed-loop electrons anywhere.
Read the capacity figures carefully. The widely quoted 8.2 MWe and 64 MWth are the design rating for a finished four-loop build, not what the site produces. Eavor's own technical update reports one loop completed, stopped at six of its twelve lateral pairs after poor cement jobs let fluid move between laterals, delivering roughly 8.5 MWth and about 0.5 MWe at a 4.3% conversion efficiency on an ORC oversized for a single loop.
Loop 2 has been pulled and Eavor is seeking new partners and investors. Precision drilling is the tradeoff for eliminating seismicity, and here it bit.
~$569M raised · BP + Chevron backing · Loop 2 timeline withdrawn (Jul 2026)
Attacks closed-loop's core physics problem directly: a proprietary Thermal Reach Enhancement material packed around the wellbore to speed conductive heat transfer, claiming a 30–50% gain. Being water-independent is a real answer to a real siting constraint in the arid West. It also holds the lowest technology-readiness rating of any developer with an up-to-150 MW hyperscaler agreement, which is the tension worth sitting with.
In July 2026 it hired Morgan Stanley to evaluate going public, which would make it the second next-generation developer to list after Fervo. Weighing a listing is not filing for one, and no S-1 exists.
150 MW Meta project (NM) in construction · ~$55M raised · 3 GW pipeline · evaluating an IPO
Retrofit-first closed loop. It reuses inactive or underperforming wellbores rather than drilling new ones, which sidesteps the drilling cost that makes ACL uncompetitive everywhere else. First field-scale closed-loop power demo at Coso, California. The ceiling is the stock of suitable existing wellbores, not the technology.
~$22.6M raised · partners: Baker Hughes, H&P, Vallourec · DoD demo
The most physically interesting variation in closed-loop: it circulates a refrigerant rather than water, working like a heat pump and claiming ~50% higher efficiency. Alberta team, out of stealth in September 2025. The fair objection is that a better working fluid does not fix drilling cost, which is what actually binds.
The Brigham City project moved forward in July 2026 on a memorandum of understanding with Fulcrum Point Holdings, which stated a long-term goal of 200 MW. An MOU is not a financed development agreement, and the 200 MW carries no timeline.
$38M Series A (TDK Ventures, Evok) · 10 MW Brigham City, UT — construction 2027 · Fulcrum Point MOU
The category most sector maps leave out. These companies are not engineering reservoirs. They are finding conventional ones nobody could see, using machine learning over gravity, magnetic, geochemical and legacy well data. Not the 10x that next-generation geothermal promises, but conventional economics (~$70/MWh) and far less technology risk.
Attacks the dry-hole problem rather than the drilling problem, and is currently the only company in the sector earning revenue from an asset it found, bought, and operates. Three discoveries in a year, capped by Big Blind in Nevada: the first blind geothermal system confirmed commercial in the U.S. in over 30 years, >100 MWe potential. Claims a >80% hit rate against 50–60% for conventional methods.
Read the hit rate with its provenance attached. It comes from the True North Institute whitepaper, and True North put $10M into this company's Series C through its All Aboard Fund. We have found no independent confirmation of the figure, and no other source for the funding totals or the Lightning Dock capacity below. That does not make any of it wrong. It does mean the strongest claim on this card and the money behind the company come from the same place.
$115M Series C · ~$180M equity + $40M development facility · Lightning Dock at 15 MW
Deliberately technology-agnostic: screen for the resource first, then pick whichever technology suits the reservoir, which the company calls "fit-for-resource." A 49-layer, 10-classifier screening engine covered 775 million acres in 2025 and produced four drill-ready assets. Nabors' backing matters more than the round size, because it points at rig access. The constraint is capitalization, a different weight class from Zanskar.
Those operational figures come from the True North Institute whitepaper and we have found no independent confirmation of any of them. The Series A is separately sourced to the company's own announcement.
$13.6M Series A (alfa8, Nabors) · Western U.S. + Indonesia · 1,233 MT stations acquired
The frontier bet. A 2023 joint venture of AltaRock and Blade Energy Partners, backed by Khosla Ventures. Set the global EGS bottomhole temperature record: 331°C at Newberry Volcano (late 2025), targeting 400°C in the 2026 campaign. Reaches superhot with conventional directional drilling rather than new physics, which is the bet that separates it from Quaise. Claims a path to sub-$0.05/kWh at terawatt scale; that remains a projection. No SHR project anywhere has generated sustained commercial power yet.
331°C record · $20–25M DOE support · 15+ MW pilot targeted 2026
A drilling-hardware company that became a developer to prove its own hardware. Project Obsidian in Central Oregon is designed as a 250 MW plant built in phases, starting at 50 MW, using gyrotron millimeter-wave drilling to reach superhot rock anywhere regardless of local gradient. The gap between 118 m demonstrated and 10+ km required is the whole story. Also profiled under Drilling Tech →
$134M Series B first close (Jul 2026) · JERA + Idemitsu · commercial ~2030
The largest listed pure-play at ~1.2 GW operated, bolting EGS onto sites that already have permits, interconnection, and a plant. That is a structurally lower-risk posture than the venture pure-plays. It is also the sector's most active validator: co-led Sage's Series B and licensed the technology, allied with SLB on EGS, partnered with Zanskar.
Worth noting the dissent from the operator with the most data. CEO Doron Blachar calls next-generation geothermal a ~$200/MWh product today, well above the $45–90 consensus.
Up to 150 MW Google PPA via NV Energy (PUCN pending) · EGS pilot with SLB at Desert Peak
Constellation (acquired Calpine and The Geysers' 725 MW for $16.4B, Jan 2026) · Chevron (GeoZone EGS pilot, Sonoma County; 20 MW toward 200) · Cyrq (20 MW Eavor pilots) · Controlled Thermal Resources (Salton Sea lithium + power, SPAC at ~$4.7B). Early stage: Exceed Geo Energy (supercritical-CO2 closed loop with sequestration; 110 MW Presidio figure is a term sheet, 9.9 MW first phase targeted late 2026) · Geothermal Technologies (Johns Hopkins spinout; 5 MW GenaSys-1 planned in Colorado's D-J Basin, announced 2023, no well drilled, no company news since Sep 2025) · Criterion. Thin public data, so treat this as a watch list rather than a diligence-ready set.
02 / Picks & Shovels — Subsurface
Legacy oilfield MWD tools top out around 175°C; geothermal wells routinely exceed 200°C and superhot targets pass 300°C. Drilling and reservoir creation consume 60–80% of project capex, split across four problems: breaking rock, surviving the heat, creating permeability, and knowing where to point the bit. Conventional rigs and consumables are the other 65% of well cost and are incumbent territory. The venture opening is everything below.
The most ambitious bet in the sector: below a certain depth, replace the drill bit with a gyrotron energy beam that vaporizes rock. MIT spinout, ~$225M raised. The $134M Series B (Jul 2026) included Japanese energy majors JERA and Idemitsu, which is corporate validation for a technology still years out. Also a developer; see Project Obsidian under SHR ↑
Approaching 1 km at its Texas field site · Project Obsidian (Central Oregon) 250 MW phased
The company trade press credits with "breaking the temperature barrier" in directional drilling: its Pandora210 MWD system is rated to 210°C, demonstrated 200+ hours above that mark. Attacks the single most-cited technical bottleneck in the sector, on ~$22M of total funding.
210°C rating vs. ~175°C legacy ceiling · $17.8M Series A (2026)
Often misread as having abandoned plasma. It has not. NexTitan Pulse still uses electrical discharge to pre-fracture rock ahead of the mechanical cutters, repositioned from replacing mechanical drilling to assisting it. That is a smarter commercial posture rather than a retreat: the Anchorbit bolts onto conventional rigs, so operators adopt it incrementally. Real-well validation at Norway's NORCE Ullrigg (Feb 2026); ~$110M raised.
$82.7M Series C (Mar 2026) · Nabors partnership gives it a global rig fleet as its route to market
Long-running, lower-profile laser-mechanical hybrid drilling play with ARPA-E roots, Petrobras history, and Chevron on the cap table. Still filing patents in 2026, but commercial milestones are sparse next to Quaise and GA Drilling. A slow burn worth checking on, not a momentum story.
~$9.1M ARPA-E era funding · active 2026 patent filings
Electrical Reservoir Stimulation: high-voltage pulses create self-propped fracture networks with no water, no proppant, and directional control, at roughly $1M per well against $5M for a conventional frac. It attacks the two things that have killed geothermal programs outright, induced seismicity and water use. It is already earning O&G restimulation revenue while geothermal scales, which is the dual-revenue model True North Institute argues this layer needs.
MIT spinout · Mazama 330°C demo · FORGE 2027 → Newberry 2028 → Salton Sea 2029
The cheapest line item is also the highest-return one, because a mischaracterised resource wastes the other 95%. Terra AI ($20M Series A, Khosla + BHP Ventures, Jun 2026) · New Mantle (physics-plus-AI characterisation) · Teverra (DoD contracts, Fort Wainwright) · TLS Geothermics (first exclusive licences in Martinique) · STRYDE (bp spinout, employee-owned since 2024; seismic nodes across 33 geothermal projects) · Geothermal Radar (global thermal model, 40+ geospatial layers, freeware tier) · EnviTrace (physics-informed AI; small team, no published accuracy metrics) · Adrok (proprietary ADR scanning; performance claims not independently validated). The incumbent to beat is Bentley's Seequent.
~1 in 3 geothermal wells is a dry hole industry-wide · the moat is data, not the model
03 / Oilfield Services & Materials
All four oilfield-services majors now run dedicated geothermal lines, which is the clearest institutional validation that oil & gas expertise transfers. They are converging on integrated subsurface-to-surface partnerships rather than arms-length equipment sales.
Full-lifecycle geothermal under its New Energy unit, anchored by the decades-old GeothermEx consulting practice. Alliance with Ormat pairs SLB's subsurface and well construction with Ormat's plant design and operations; EGS pilot together at Desert Peak, plus a Star Energy (Indonesia) partnership.
New Energy unit targeting ~$3B revenue by 2030 (geothermal one of five lines)
Vulcanix geothermal-rated bits plus integrated development deals: 500 MW over five years with EnCap-backed Mantle Reach Power, a drilling collaboration with Helmerich & Payne, and exploration/engineering work on XGS's New Mexico project. Trade coverage puts cumulative geothermal-linked capacity commitments at roughly 800 MW as of 2025.
500 MW Mantle Reach agreement (Jun 2026) · ~800 MW cumulative
Launched RangeStar Geothermal (Feb 2026), a magnetic-ranging well spacing and intercept service aimed directly at multilateral EGS and closed-loop architectures. Patent filings on real-time geothermal gradient characterization show sustained internal R&D rather than rebadged O&G product.
~70 years of geothermal history · 20–30 yr well-integrity design horizon
The "arms dealer" of the group: rigs, the Phoenix geothermal bit portfolio, and thermal coatings. Hardware sales without exploration or market risk. Less headline partnership activity than SLB or Baker Hughes, but it captures value as the sector's specialized-hardware demand grows.
Equipment-first positioning · no resource or offtake risk taken
Premium casing and vacuum-insulated tubing (THERMOCASE) for geothermal wells, with an $800M five-year agreement making it Fervo's exclusive supplier of U.S.-made tubulars and VAM connections, and a disclosed target of 10–15% of group EBITDA from New Energies by 2030.
Because casing demand scales directly with well count, Vallourec's quarterly geothermal bookings are one of the most quantifiable proxies for sector build-out pace. Bookings more than tripled in 2025 versus the prior two-year average.
The wellhead is the adjacent piece of the same interface. Stream-Flo Industries (Edmonton, 60+ years of wellheads and valves) markets a geothermal line, GeoFlo, covering casing suspension through stimulation-readiness, though it publishes no named geothermal deployment.
Bookings 3x in 2025 · ~400,000-tonne addressable tubular market
04 / Power Equipment
Roughly half of plant cost is surface hardware, and nearly all geothermal turbomachinery is made outside the U.S. Fervo ordering 1.7 GW of turbine capacity years before construction tells you where the market expects the constraint to bind next.
The incumbent ORC turbine OEM at scale, with MHI's manufacturing muscle behind it. Supplies Cape Station's 50 MW "GeoBlocks," then signed a 1.7 GW, up-to-35-block framework with Fervo (Apr 2026), likely the largest ORC equipment commitment in sector history.
1.7 GW Fervo framework · 3-year horizon
The counter-bet: factory-built, container-sized turbine units trucked to site in weeks, or "power plant as a product." SpaceX-alumnus founder, rocket-turbomachinery lineage, $22M seed raised since 2024. Directly targets the turbine-shortage thesis Rhodium flagged as the sector's emerging bottleneck.
LA pilot · first commercial 2.5 MW targeted 2027
Italian ORC maker differentiated on its Radial Outflow Turbine design, claiming 600 MWe installed worldwide across geothermal, waste heat, and biomass. Competes with Turboden on turbine efficiency rather than manufacturing scale.
600 MWe installed · 3 MWe UK plant with Geothermal Engineering Ltd
Climeon (Stockholm, public micro-cap): modular HeatPower units for 75–105°C resources most binary plants can't touch; geothermal is one of several verticals. Orcan Energy (Munich, TUM spinout; modular ORC, its named geothermal units are 125–640 kW, two orders below utility scale) · Gradient Geothermal (patented HXC sled: heat exchanger, ORC and dry cooler on one mobile skid) · Hybrixcel and Heat Recovery Innovations (both private and early; minimal public milestone data). Echogen: supercritical-CO2 power cycles, an emerging candidate for higher-temperature EGS/SHR plants, with geothermal relevance still forming.
Iceland HeatPower deployments · DOE STEP program ecosystem
05 / Capital
The capital base is broad and mainstream, not a climate niche: patient climate VC, generalist growth equity underwriting the AI-power thesis, shale-style PE platforms, corporate strategics, and sovereign and pension money.
Venture Capital
PE, Infra & Strategic
Public Markets
06 / Buyers & Offtake
The structural shift since 2024: tech companies negotiating direct large-volume PPAs, often routed through a utility via NV Energy's Clean Transition Tariff, where the hyperscaler pays the green premium so ordinary ratepayers do not.
The dominant relationship in the sector: 3 MW pilot (2021) → 115 MW PPA → up to 150 MW more via Ormat → the 3 GW Fervo framework with right of first refusal. Also an equity investor in both Fervo and Baseload Capital.
3 GW framework through 2033
Entered later but at scale, and deliberately diversified: 150 MW with Sage (Texas) and 150 MW with XGS (New Mexico). Two developers, two different sub-technologies. A different risk posture than Google's Fervo concentration.
~300 MW across two counterparties
Southern California Edison: 320 MW with Fervo across two 15-year PPAs, the largest traditional-utility commitment. NV Energy: the critical intermediary via its Clean Transition Tariff. Clean Power Alliance and Shell Energy round out the board.
320 MW SCE · CTT = the replicable template
Compliance-driven demand: 99.9% energy resilience mandated for critical missions by 2030. Six companies (Fervo, Eavor, Sage, GreenFire, Zanskar, Teverra) across 7+ installations, with a standardized ~5 MW solution shortlisted for up to ~50 sites.
Legally mandated, appropriations-dependent
Iceland: Landsvirkjun and ON Power negotiating data-center geothermal supply (~150 MW in live talks). Indonesia: Barito Renewables targeting 1 GW with $365M committed, a listed power company expanding on its own balance sheet.
Barito: 1 GW target · $365M committed
Data-center operator Switch (13 MW, Ormat) said the quiet part out loud: the move to firm geothermal is a response to "operational and financial risks posed by soaring AI-driven electricity demand and increasing grid instability." That is a reliability hedge rather than a sustainability preference.
Firmness is the product
07 / Land & Leasing
The best-documented land trade in the sector came from patience rather than a flip. Fervo assembled ~595,900 acres at roughly $4/acre in 2019–2021, years before EGS was bankable. That position now sits inside a public company that debuted above $10B.
Since then, premium basins have re-rated hard: Utah's record sale averaged $111/acre, California's $117/acre, New Mexico's June 2026 sale ~$109/acre. The pattern is that price follows proximity to a credible de-risking catalyst such as Cape Station, FORGE or a Meta-anchored project, rather than raw geology. Generic acreage still clears at $4–12/acre.
Under 0.5% of the leasable federal base is currently under lease, and BLM now mandates annual competitive sales. The running room is real; so is the bifurcation.
| Sale | $/acre |
|---|---|
| Fervo assembly, 2019–21 | ~$4 |
| Nevada (typical range) | $8–36 |
| Oregon record (Malheur Co.) | $82 |
| New Mexico, Jun 2026 | ~$109 |
| Utah all-time record | $111 |
| California record, Sep 2025 | $117 |
Source: BLM sale results · average bonus bids
Federal acreage is not the only channel. In July 2026 the Colorado State Land Board opened a competitive process on 6,597 acres of state trust land across 11 parcels in La Plata, Saguache, Las Animas, Alamosa and Rio Grande counties, with proposals due 20 September 2026 and many parcels sited near existing transmission.
Watch the scoring rather than the acreage. Colorado weighs community engagement and a water plan alongside the financial bid, where a BLM sale is decided on price. Where state land boards run geothermal processes, the winner is not necessarily the highest bidder.