Geothermal has split in two

In the last week of June, France's rivers ran too warm to cool its reactors. The national operator, EDF, curtailed as much as 6.4 GW of nuclear capacity, day-ahead power swung from minus €53 to €433 per MWh, and Germany averaged €111 for the month.¹

Power prices are set at the margin: when always-on supply drops out, the most expensive plant still running, usually a gas peaker, sets the price for everyone. That week put a price on supply that doesn't depend on the weather.

The money had come in the year before: across 2025, next-generation geothermal companies raised about $2.2B, up 80% on the year, from $22M as recently as 2018.² The $2.2B is the fast-growing edge of a bigger flow: conventional geothermal power drew close to $5B in 2025, roughly four times its 2018 level, and geothermal heating projects took in more than $11.5B. Meta and Google alone have contracted roughly 650 MW of geothermal power, much of it from plants that don't exist yet.³

Why now

In June the US House passed the Geothermal Energy Advancement Act, and the Department of Energy's geothermal office took a 20% budget bump this year.⁴ Germany's Geothermal Acceleration Act called for 100 new projects by 2030; the planning map already shows more than 150.⁵

Data centers need round-the-clock power close to the load, and geothermal delivers capacity factors above 90%⁶; that is what Meta and Google were buying. (The demand side of that equation, how data centers cut and flex their own consumption, is something we mapped in an earlier deep dive.)

Fervo (the current star of the geothermal world), meanwhile, drills at around 30 meters an hour in rock where 10 was long the industry standard, mostly with bits, tooling and workflows carried over from shale.⁷ That 30 is a best case, not a steady state, but it came at Cape Station, Fervo's flagship site in Utah, in granite, the hardest rock the industry drills. Oil and gas crews, rigs and subsurface software have found a second market for their expertise.

Today geothermal covers less than 1% of global energy demand. The International Energy Agency (IEA) thinks it could reach 800 GW if drilling costs keep falling; running near-constantly, that much capacity would generate some 6,000 TWh a year.⁸ The energy think tank Ember calculated in February that geothermal could replace up to 42% of the EU's fossil electricity at under €100 per MWh.⁹

Fervo's Cape Station should put its first 100 MW on the grid late this year, on its way to a 500 MW build-out, and the company's Nasdaq listing in May settled whether public markets would take a next-generation geothermal company.¹⁰ Vulcan closed €2.2B in project financing for its geothermal-and-lithium build in the Upper Rhine (a different 2.2 from the venture number in the opening, and in some ways the bigger statement: that one is debt, for a single build).¹¹ Equity's share of next-generation financing is shrinking as debt like that becomes available.²

What counts as geothermal now

What a well produces comes down to temperature. Shallow and warm gets you heat, deep and hot gets you electricity, and most systems can serve either end. The deeper you go, the more the economics tilt toward power.

At the shallow end sits the most familiar form: boreholes a few hundred meters deep, paired with a heat pump. That is the geothermal already under hundreds of thousands of houses and small commercial buildings, drawing on the near-constant temperature of the shallow ground to heat in winter and cool in summer.¹² No deep resource required, and drilling is easy where the ground stays soft.

Hydrothermal is the bankable baseline wherever the resource exists (Figure 1 sorts the full map by depth). Enhanced geothermal systems (EGS) fracture rock to create a reservoir where none exists naturally; that is how the US is expanding the resource, and where the seismicity risk comes from.

Around those two sit single-well co-axial retrofits of old oil and gas wells, closed-loop designs that circulate fluid through sealed bores and give up some heat transfer for control, and superhot rock, still at pilot depth in Iceland but promising several times the electricity per well.¹³

The IEA has mapped what EGS could reach by depth. Within 8 kilometers of the surface, some 600 TW is technically recoverable, roughly 2,000 times the technical potential of conventional geothermal. By 7 kilometers, almost every region on earth is sitting on a usable resource to generate power.⁸

“By 7 kilometers, almost every region on earth is sitting on a usable resource to generate power.”

The split: the US validates power, Europe builds heat

The US has about 4 GWe of geothermal power installed, and two companies, Ormat and Calpine, operate roughly 69% of it.¹⁴

Europe shows a similar headline number, around 3.5 GWe. Only about 1 GWe of that sits inside the EU, and Italy's 0.9 is most of it. The rest is elsewhere: Türkiye runs 1.8, Iceland 0.8.¹⁵

Europe's lead is in heat, and it is decades old. The suburbs of Paris have drawn on the Dogger aquifer since the 1970s; some 50 networks warm roughly 250,000 households from limestone 2 kilometers down.¹⁶ More than 400 district heating and cooling plants now run across the continent, and the European Geothermal Energy Council counts another 500 in development: whole neighborhoods, heated from beneath their own streets.¹⁷

In the US the demand is already under contract; utilities and hyperscalers have signed for firm capacity, and data centers are siting next to wells. A signed offtake is what makes a project bankable: it turns a drilling program into something a lender will finance against. In Europe most of it is still a government target. Germany wants 10 TWh of deep geothermal heat by 2030 under the Acceleration Act; France is aiming for 16 TWh of combined geothermal heat by 2030, and around 25 by 2035; the Netherlands, 4 TWh by 2030.¹⁸

In Utah, Fervo's Cape Station test wells peaked at flow rates worth more than 10 MW of electricity per well pair, triple what the company managed at its first project, and drilling costs per foot roughly halved too.¹⁹ The financing was ordinary power purchase agreements and project debt. For a European team, that is a repeatable path from pilot to hundreds of megawatts, built with tools it can license and improve on.

The US also starts with an advantage Europe can't legislate: a century of onshore oil and gas. Idle rigs, drilling crews between contracts, subsurface data on tens of thousands of legacy wells, and a permitting culture that lets a company start drilling a test well quickly and learn from it. "Drill, baby, drill" is grid policy now, whatever else it is.

Underground Ventures, in Copenhagen, invests only in geothermal; its data sits behind the financing figures at the top of this piece.² We put the Europe question to Thomas Lambaa Giehm, an investment manager there.

"Where does Europe realistically play in next-generation geothermal over the next decade? Will we see experienced U.S. players in Europe soon?"

"Europe will leverage its engineering talent to innovate across drilling technologies and power conversion, in particular. Our supply chain is well positioned for closed-loop geothermal, which is an obvious way to replace imported gas for heating and has the potential to produce baseload power in the future. Close collaboration with North America can accelerate geothermal development via technology and talent transfer, which will advance both closed-loop and enhanced geothermal systems. In the next decade I expect we will see U.S. developers coming to Europe, especially in Central and Eastern Europe where Romania and Hungary are probable destinations."

The granite problem

Approvals in Europe fragment across regions and agencies, and a project can spend years assembling signatures that a Texas operator collects in months. The companies building here either carry that navigation in-house or partner with the utilities and municipalities who already own the permit relationships.

Geothermal also carries resource risk from the start: roughly half of first wells in a new field miss, and the World Bank's IFC puts success across the first 5 exploration wells at around 60%.²⁰

Basel, 2006: a stimulation project, high-pressure fluid pumped down to crack the rock open, triggered a magnitude 3.4 event. The project was terminated and claims ran to about $7M. St. Gallen followed in 2013 at magnitude 3.5.²¹

What happened in Switzerland is still written into Germany's rulebook. Stimulation is permitted only in limestone and sandstone, and the chemical additives pumped down with the water sit in the same regulatory red zone as fracking. So the European route around US-style EGS runs through closed-loop designs and, further out, superhot rock. One exception to this is Romania, with established onshore oil-and-gas infrastructure and rules that still permit stimulation. US EGS developers are starting to look there for fresh opportunities.

But both of those still chase what power needs: high temperature, and high temperature means depth. In most of Europe the deepest, hottest resources sit in hard crystalline basement, in Germany mostly granite. With stimulation off the table, there is no shortcut through it.

Even where the permits clear, Europe is short of rigs, and the shortage is worst for the super-large rigs that deep wells need. Drilling stays the largest single line in the budget, 30 to 57% of capex by the US National Renewable Energy Laboratory's count, and the “rig days” concentrate in the hardest section of the hole.²²

That hardest section is the granite. Conventional bits slow to a crawl and cost per meter climbs fastest exactly there. Drilling costs have fallen nearly everywhere else; in granite they haven't broken yet.

For deep European geothermal, the whole case narrows to a single lever: how fast, and how cheaply, you can make hole through hard rock, kilometers below the surface.

“The whole case narrows to a single lever: how fast, and how cheaply, you can make hole through hard rock.”

(Europe's heat opportunity sits shallower, and it doesn't run through granite.)

What has to mature

Cut the cost of drilling hard rock and the rest of the case follows. Granite-hosted resources become economic and the deep power case comes into range. Closed-loop designs get back the margin their weaker heat transfer eats up. (How the contending methods actually work, and the cost and failure data behind them, is a second piece of its own.)

Cheaper holes aren't the end of it, though, says Giehm again:

"Everyone in this market talks about drilling costs. What do you see as the next constraints to overcome? Where does our existing technology hit its limits?"

"Once the bottom temperature of a well exceeds 200°C, standard oil & gas tools no longer suffice, and options at 300°C are extremely limited. We need better high-temperature completion tools that can secure well integrity and ensure optimal fluid flow for long-term heat extraction. Achieving this requires innovations in material science, reservoir engineering, and hardware design that go beyond technology transfer from the oil & gas industry."

Sealed bores sidestep most of the seismicity problem, and real-time monitoring picks up what's left. Superhot rock is the long bet; the Icelandic pilots suggest several times the electricity per well.¹³

Heat pays back sooner than power, and compact, low-footprint rigs, with increasingly robotic drilling to lower cost further, would open up the dense urban blocks district heating can't reach today. A borehole drilled for heat can also store it: pump surplus warmth underground in summer, draw it back out in winter.

For a measure of how hard all of this is, look at Geretsried, a town south of Munich. Eavor's flagship project there, Germany's marquee closed-loop build, came in between 0.5 and 1 MWe gross against an 8.2 MWe design target, with 1 of 4 loops drilled and, on that loop, 6 of 12 lateral pairs. Eavor has since stepped back from the operator role toward licensing its technology.²³

The reality is that most novel drilling approaches will not survive their first field campaign. And the benchmark itself keeps moving; Fervo's 30 meters an hour came from optimized versions of the industry's standard diamond-cutter bits, while a generic program in the same rock has historically managed 8 to 15.⁷ (We're looking at a lot of the state of the art right now, and some of what's coming is genuinely exciting. A claim of beating the benchmark tells me little on its own. The decks I keep re-reading explain what breaks at 6,000 meters, and why their tool breaks later.)

Where we play

We back the technology layer at Seed and Series A: the companies making wells cheaper, deeper, and lower-footprint. That means hard-rock drilling for the power case, and heat for the city blocks the networks don't reach.

Our current watchlist is the market map below. Who are we missing?

Download the full market map

By Leo Allgoewer, with support from Hélène Dupré

Sources

1. June 2026 heatwave: EDF river-temperature curtailments up to ~6.4 GW; French day-ahead range −€53 to €433/MWh; German June average €111/MWh.
2. IEA commentary, Jan 2026 (data via Underground Ventures): next-gen geothermal financing $2.2B in 2025, +80% YoY, from $22M in 2018; conventional geothermal power ~$5B (~4× 2018); heating projects >$11.5B; equity's share shrinking as project debt becomes available (e.g. Fervo's $421M non-recourse debt, Mar 2026). https://www.iea.org/commentaries/investment-in-next-generation-geothermal-is-surging-policies-are-key-to-further-growth
3. ~415 MW hyperscaler PPAs: Meta–XGS 150 MW, Jun 2025 (https://www.businesswire.com/news/home/20250612778008/en/XGS-Energy-and-Meta-to-Partner-on-150-MW-Advanced-Geothermal-Project); Meta–Sage 150 MW, Aug 2024 (https://about.fb.com/news/2024/08/new-geothermal-energy-project-to-support-our-data-centers/); Google–Fervo/NV Energy 115 MW (https://www.datacenterdynamics.com/en/news/google-buys-115mw-of-geothermal-energy-to-power-nevada-data-centers/)
4. US policy: Geothermal Energy Advancement Act (H.R.5631) passed House 2 Jun 2026; FY2026 appropriations $150M to DOE Geothermal Technologies Office, +20% YoY (Congress.gov).
5. Germany: Geothermal Acceleration Act, 100 new projects by 2030; 150+ in planning as of mid-2026 per the updated project map (BMWK via ThinkGeoEnergy). https://www.thinkgeoenergy.com/germany-aims-for-100-new-geothermal-projects-by-2030/
6. 90%+ capacity factors: well-established for conventional hydrothermal (US DOE EERE; UMich CSS factsheet); not yet proven for next-gen EGS at commercial scale. https://css.umich.edu/publications/factsheets/energy/geothermal-energy-factsheet
7. Fervo drilling rates: Sugarloaf appraisal well (a Cape Station well, granitoid basement) max-avg ~29 m/h ≈ "around 30" (Fervo PR: https://fervoenergy.com/fervo-energy-drilling-results-show-rapid-advancement-of-geothermal-performance/; Power Magazine: https://www.powermag.com/partner-content/geothermal-groundbreakers-the-projects-redefining-firm-clean-power/). Hard-rock generic baseline 8–15 m/h: FORGE early campaigns ~6–8 m/h rising to ~15 by 2024 (same sources + NREL 2025 drilling cost curves).
8. IEA, The Future of Geothermal Energy: geothermal meets <1% of global energy demand today; up to 800 GW / ~6,000 TWh a year with cost declines; EGS-by-depth ~600 TW technically recoverable within 8 km below $300/MWh; by ~7 km almost every region has a usable resource. https://www.iea.org/reports/the-future-of-geothermal-energy/executive-summary
9. Ember via Euronews (19 Feb 2026): up to 42% of EU coal- and gas-fired generation replaceable below €100/MWh. https://www.euronews.com/2026/02/19/geothermal-energy-could-replace-42-of-eus-fossil-electricity-which-nation-has-the-most-pot
10. Fervo Cape Station build-out: Phase I ~100 MW to the grid from late 2026 (https://fervoenergy.com/fervo-energy-secures-421-million-in-non-recourse-project-financing-for-cape-station/); full build-out 500 MW by 2028, upsized from 400 MW and fully contracted (https://fervoenergy.com/fervo-secures-new-financing-to-accelerate-development/); Nasdaq listing May 2026.
11. Vulcan: €2.2B project financing closed Dec 2025 for the Lionheart geothermal-lithium build, Upper Rhine.
12. Shallow boreholes + heat pumps under "hundreds of thousands of houses": conservative — Germany alone has 400,000+ geothermal heat pumps installed, ~20,000 added per year (Fraunhofer IEG via Clean Energy Wire). https://www.cleanenergywire.org/factsheets/geothermal-energy-germanys-largely-untapped-renewable-heat-source
13. Superhot rock, several times the electricity per well in principle: IDDP pilot data (36–50 MW potential per superhot well vs ~5–8 MW typical) via CATF, Jan 2025. https://www.catf.us/2025/01/introduction-next-clean-energy-frontier-superhot-rock-geothermal/
14. US capacity: 3.97 GWe end-2024; Ormat + Calpine ~69% of installed capacity (NREL 2025 US Geothermal Market Report). https://www.nlr.gov/geothermal/2025-us-geothermal-market-report
15. Europe capacity: ~3.5 GWe total, ~1 GWe inside the EU, Italy ~0.9 GWe; Türkiye 1.8, Iceland 0.8 (EGEC market reports 2023/2024). https://www.egec.org/media-publications/egec-market-report-2024-key-findings/
16. Paris basin: ~50 geothermal district-heating networks, ~250,000 household-equivalents, Dogger aquifer at 1,500–2,000 m, operating since the 1970s (BRGM; ThinkGeoEnergy; DBDH).
17. District heating and cooling: 412 GeoDHC plants operating across Europe, ~500 more in development (EGEC 2024). https://www.renewableenergymagazine.com/geothermal/egec-report-forecasts-resurgence-in-deployment-of-20250711
18. National heat targets: Germany ~10 TWh deep geothermal heat by 2030 (Geothermal Acceleration Act / BMWK, see 5); France ~16 TWh combined geothermal heat by 2030, ~25 by 2035 (PPE3: https://www.economie.gouv.fr/files/files/2026/ppe3.pdf); Netherlands ~4 TWh by 2030
19. Fervo Cape Station well test: peak 107 kg/s in 30-day 2024 test = >10 MWe per well pair, ~3× Project Red (Fervo PR, Sept 2024: https://fervoenergy.com/fervo-energys-record-breaking-production-results-showcase-rapid-scale-up-of-enhanced-geothermal/); drilling cost per foot roughly halved vs Project Red ($450–550/ft average vs ~$1,050/ft; best wells ~$350/ft; Power Magazine, see 7)
20. Well success: first exploration well ~50%; ~60% success across the first 5 (IFC global study via ThinkGeoEnergy). https://www.thinkgeoenergy.com/global-study-on-success-of-drilling-geothermal-wells-by-ifc/
21. Induced seismicity: Basel 2006 (ML 3.4, ~$7M in claims, project terminated); St. Gallen 2013 (ML 3.5).
22. NREL: drilling 30–57% of capex (NREL 2025 drilling cost curves).
23. Eavor Geretsried: between 0.5 and <1 MWe gross vs 8.2 MWe design target; 1 of 4 loops drilled, 6 of 12 lateral pairs on that loop; operator-role exit toward technology licensing. Technical update, May 2026: https://eavor.de/wp-content/uploads/2026/05/Technisches-Update-Geretsried-260529.pdf