Snapshot

Geothermal power plants tap heat from below the earth's surface to spin turbines and generate electricity. Global installed capacity was 16,873 MW at end-2024, spread across 35 countries — roughly the output of 17 large nuclear reactors. The United States leads with 3,937 MW (23% of global), almost all in Nevada and California. The entire world added only 389 MW in 2024, a 2.4% growth rate. Unlike solar or wind, geothermal runs 24/7 at a 90%+ capacity factor, making it one of the few clean sources that can serve as genuine baseload. That property is now attracting data-center buyers: Google, Switch, and Meta have all signed geothermal PPAs since 2024. But the resource is geology-constrained, capital-heavy, and tiny relative to the power demand it is being asked to fill.

The world has 16.9 GW of geothermal capacity and adds roughly 400 MW a year. US data-center operators have signed PPAs for over 1.6 GW of new geothermal — more than four years of the entire world's recent build rate. The supply chain and geology cannot deliver that on the timeline demand is requesting. The open question is whether enhanced geothermal (EGS) can break the bottleneck, or whether these contracts slip years.

Global capacity: ThinkGeoEnergy 2024 Top 10 report. US capacity and PPA data: NREL 2025 US Geothermal Market Report. LCOE: NREL 2025 US Geothermal Market Report (2022$). Capacity factor: US DOE Office of Geothermal.

The product & how money is made

A geothermal power plant drills wells into hot rock or hydrothermal reservoirs (typically 150–350°C), brings hot water or steam to the surface, and uses it to spin a turbine-generator. Two main technologies exist:

• Flash steam: High-pressure hot water (>180°C) is "flashed" into steam in a separator. Used at The Geysers in California (the world's largest complex). Lower cost per MWh ($63–$74) but requires hotter, rarer reservoirs.
• Binary cycle: Moderate-temperature fluid (120–180°C) heats a secondary working fluid (isobutane or isopentane) that boils at lower temperature and drives the turbine. More broadly deployable but higher cost ($90–$110/MWh). This is what Ormat primarily builds and operates.

A third technology, Enhanced Geothermal Systems (EGS), uses horizontal drilling and hydraulic stimulation (techniques from oil & gas fracking) to create artificial reservoirs in hot dry rock. EGS dramatically expands the geographic footprint beyond volcanic regions. Fervo Energy (private) is the leader; Ormat has early-stage EGS pilots with SLB and Sage targeting 2–4 MW each, expected in 2027.

Revenue models

• IPP (independent power producer): Own the plant, sell electricity under long-term PPAs (15–25 years). This is Ormat's core business — 80% of revenue. Ormat's CEO disclosed in Q4 2024 that data-center PPA negotiations are at rates exceeding $100/MWh.
• Equipment / EPC: Design, manufacture, and install geothermal power plants for others. Lumpy revenue tied to project milestones. Ormat's product segment — ~16% of FY2024 revenue, $314M backlog as of May 2025.
• Energy storage: Co-located battery storage and ancillary grid services. Ormat's fastest-growing segment (up 153% YoY in Q1 2026).

Technology descriptions: US DOE Office of Geothermal. LCOE ranges: NREL 2025 US Geothermal Market Report. Data center PPA pricing: ORA Q4 2024 earnings release (Feb 2025). Product backlog: ORA Q1 2025 earnings (May 2025).

Demand

Contracted demand contracted

Since 2021, 26 new US geothermal PPAs have been signed for over 1,642 MW of new capacity. The acceleration is driven by a 2021 California PUC order requiring utilities to procure 1,000 MW of "non-weather-dependent, non-battery, zero-emission energy," and hyperscaler demand for 24/7 clean baseload for data centers.

PPAs: Ormat investor releases (Jan 2026, 2026); DataCenterDynamics (Google/Fervo); ThinkGeoEnergy (Fervo/SCE); NREL 2025 US Geothermal Market Report (total PPA count + MW). Meta/Sage: NREL report.

Ormat also has ~190 MW of existing contracts expiring 2031–2034 at legacy rates averaging the "mid-$80s" per MWh that management plans to renegotiate at higher prices. The CD4 (Casa Diablo) contract was already extended through 2037 with a ~27% price increase.

Forecast demand est.

Data-center power demand in the US is projected to roughly double from ~25 GW today to 50–70 GW by 2030 est.. Geothermal's share will remain small in absolute terms but is growing from near-zero. US next-gen geothermal companies alone have attracted $1.5 billion+ in private capital since 2021, led by Fervo ($642M equity + $331M debt) and Eavor ($387M equity + $142M debt). The US DOE targets 2–5 GW of next-gen geothermal deployed across 4–6 states by ~2030 est., requiring $20–$25 billion in investment est., and envisions 90+ GW by 2050 est..

Supply

Installed capacity

ThinkGeoEnergy 2024 Top 10 Geothermal Countries report.

Bottlenecks

• Geology: Conventional hydrothermal resources are concentrated in volcanic/tectonic zones — Western US, Indonesia, East Africa, Iceland. EGS aims to fix this, but commercial-scale EGS does not yet exist outside pilot projects.
• Drilling risk and cost: Each well costs $5–$13 million and can fail if it doesn't hit the right temperature or permeability. At Utah FORGE, drilling speed improved 500%+ since 2017 and costs dropped from $13M to $5M per well, but this is one site under lab-like conditions.
• Timeline: A new conventional geothermal plant takes 5–7 years from exploration to first power est.. EGS projects target 3–5 years but have no track record at scale. Fervo's first 70 MW phase at Cape Station (started drilling 2024) targets 2026 delivery.
• Grid interconnection: US grid interconnection queues average 55 months. Geothermal projects face the same queue as everyone else.
• Workforce: The oil & gas drilling talent pool is being tapped by EGS developers, but rig availability and specialized geothermal expertise remain limited.

Ormat targets 2.6–2.8 GW total portfolio by end-2028, up from 1.8 GW today. That includes ~216 MW of geothermal/solar under construction or development and ~1.5 GWh of additional energy storage.

The gap

Pricing direction

Geothermal PPA pricing is moving up. Ormat's CEO stated that data-center PPA negotiations are at rates exceeding $100/MWh, compared to legacy contract rates in the mid-$80s/MWh and a conventional LCOE of $63–$74/MWh for flash. The DOE projects next-gen geothermal LCOE can reach $60–$70/MWh by 2030 est. and the "Geothermal Shot" target is $45/MWh by 2035 est..

PPA pricing: ORA Q4 2024 earnings call. LCOE projections: US DOE Geothermal Liftoff Report (Sep 2024). Flash LCOE: NREL 2025 US Geothermal Market Report.

The players

Note on FCEL: FuelCell Energy is not a geothermal company. It makes molten-carbonate fuel cells — electrochemical devices that convert natural gas (or hydrogen) to electricity without combustion. It is included here because it competes for the same buyer (data centers wanting on-site, always-on, clean-ish baseload power).

ORA financials: Q4 2024 and Q1 2026 earnings releases; StockAnalysis.com (cash flow). FCEL financials: 10-K FY2025 (Oct 2025); StockAnalysis.com (income statement, statistics). Stock prices: Jun 2, 2026. Capacity: ORA corporate profile (1.8 GW); FCEL 2025 annual report (approaching 1 GW cumulative, ~100 MW in South Korea).

Key private competitors

• Fervo Energy — EGS technology leader. Cape Station (Utah): 400 MW planned, first 70 MW targeting 2026. PPAs with Google (115 MW) and SCE (320 MW). $642M equity + $331M debt raised.
• Sage Geosystems — Next-gen geothermal. 150 MW PPA with Meta. Early stage.
• Eavor Technologies — Closed-loop geothermal (no fracking needed). $387M equity + $142M debt raised. European focus.

The price of exposure

Ormat Technologies (ORA) — $144.48/share

At $144, the enterprise value is ~$11.5B for a business guiding to ~$630M adjusted EBITDA in 2026 and targeting ~$800M+ by 2028 if capacity grows to 2.6–2.8 GW est.. The electricity segment generates $700M+ in steady recurring revenue with ~37% gross margins under long-term contracts. Free cash flow has been negative because capex ($488M in 2024) exceeds operating cash flow ($411M), though the gap is narrowing as capacity comes online.

FuelCell Energy (FCEL) — $24.64/share

FCEL trades at $1.3B market cap with a $188M/yr net loss and negative gross margins. The $1.19B backlog provides a demand signal, and the 12.5 MW data-center power block (announced March 2026) could shift unit economics if costs come down at scale. Manufacturing expansion from 100 MW to 350 MW is planned but unfunded from operations. At current burn, cash lasts roughly two years without further dilution est..

Stock prices and shares: StockAnalysis.com (Jun 2, 2026). ORA financials: Q1 2026 and FY2024 earnings releases. FCEL financials: 10-K (FY ending Oct 2025), StockAnalysis.com income statement + statistics. Reverse split: FCEL press release Nov 7, 2024.

What to deep-dive next

• Ormat's 10-K (FY2025): Full cash flow statement, breakdown of generation by plant, PPA expiration schedule, and details on the 2.6–2.8 GW 2028 target. How much of the new capacity is already funded vs. requiring incremental capital.
• Ormat's ~200 MW data-center PPAs: Full list, pricing, and delivery timeline to separate contracted cash from aspirational guidance.
• FCEL 10-K (FY Oct 2025): Backlog breakdown by segment and counterparty creditworthiness. How much of the $1.19B backlog converts to cash in the next 2–3 years? Unit economics of a 12.5 MW data-center power block vs. a geothermal PPA.
• Fervo Energy IPO watch: Track Cape Station's first 70 MW delivery (targeting 2026) as the make-or-break milestone for EGS commercialization.
• PPA pricing transparency: Ormat's "exceeding $100/MWh" data-center quote is the best data point. Compare to nuclear PPA rates ($90–$120/MWh for new-build) and gas peaker rates.

Sources & confidence