One resource, many energy forms: why integration changes geothermal's economics.

A geothermal project is almost always judged as a power project: how many megawatts, at what cost per megawatt-hour. It is a reasonable question, and an incomplete one. Electricity is only the highest-temperature use of a resource that stays useful long after it is too cool to spin a turbine.

That distinction is not a marketing frame. It is the physical structure of the resource — and it is where the economics are decided.

The cascade

Geothermal fluid gives up its energy in stages, and each stage has a customer. Above roughly 175 °C it drives conventional steam generation; from about 100 °C a binary cycle can still make power. Below that the fluid is no longer useful for electricity, but it is very useful for something else: industrial process heat and cooling from roughly 100–150 °C, space and district heating from 60–100 °C, greenhouse and aquaculture heat from 30–60 °C.

The engineering term for using one fluid down that ladder is cascading, and it is established practice rather than a proposal. As Lund, Huttrer and Toth document, low-to-moderate temperature resources are run through a binary power plant and then "cascaded for direct uses, before being injected back into the aquifer" — in Iceland, Austria, Germany and at the Oregon Institute of Technology.

The consequence is easy to state and easy to underrate. A project optimised only for electricity discards most of what it produced. A project designed around a demand profile does not.

Heat is the larger business

It is worth being blunt about the relative sizes here, because the industry's own framing obscures it. Global installed geothermal power capacity was about 16 GW at the end of 2025, on IRENA's count. Global geothermal direct use — heat, delivered — was an estimated 126,000 MWt across more than 88 countries, supplying roughly 1,280,000 TJ a year.

Geothermal is already, by a wide margin, more a heat business than a power business. What it has rarely been is a business that sells both from the same asset, to the same customer, under one contract.

The leg nobody expects: cooling

The least intuitive output is cold. An absorption chiller uses heat, not electricity, to drive the refrigeration cycle: geothermal fluid at roughly 90–120 °C boils refrigerant out of an absorbent, and the cycle delivers chilled water. In a cascade it can sit after the power plant, using fluid on its way to reinjection.

One number needs care. A single-effect absorption chiller has a coefficient of performance around 0.6–0.8 — a peer-reviewed study of geothermal-driven systems at 80–105 °C measured a maximum COP of 0.795 for a lithium-bromide cycle. That looks poor beside an electric chiller's COP of 3–6, and the comparison is meaningless: an electric chiller's COP counts purchased electricity in, while an absorption chiller's counts heat that the plant already has and would otherwise reinject. The relevant question is not efficiency per unit of heat. It is whether the heat had any other buyer.

In summer, in a district-heating system, it does not. Cooling converts a seasonal heat surplus into a seasonal product — which is precisely why it raises the utilisation of the asset rather than the efficiency of a component.

Honesty requires the scale, too: geothermal cooling and snow-melting together accounted for about 435 MWt globally in 2020 — well under one percent of direct use. This is a technically proven application and a commercially negligible one. That is a description of an opening, not of a market.

Why the demand side is moving

Two things have changed at once.

The first is load. The IEA's April 2026 assessment puts data-centre electricity consumption at roughly 485 TWh in 2025 — up about 17% in a year, against roughly 3% growth in electricity demand overall — and projects it to roughly double to about 950 TWh by 2030, near 3% of global electricity. These are loads that want power and cooling, continuously, in one place. The IEA goes further and names the consequence: AI deployment is "a major source of momentum for the nuclear and advanced geothermal industries." That momentum is now contractual, not rhetorical — in March 2026 Fervo Energy and Google signed a framework agreement contemplating up to 3 GW of geothermal through 2033.

The second is policy. The European Commission's AccelerateEU communication of April 2026 states that geothermal "can also replace natural gas in district heating and cooling networks", and commits the Commission to explore geothermal de-risking and insurance schemes with public financiers. The inclusion of cooling in that sentence is not incidental; nor is the acknowledgement that the sector's binding constraint is risk, not resource.

What integration actually buys

Set the pieces together and the argument is not that geothermal can do many things. It is that doing several of them for one customer changes the numbers.

Geothermal's structural advantage is that it runs. On the US EIA's 2025 annual figures, geothermal plants ran at a 65.9% capacity factor, against 34.2% for wind and 24.4% for utility-scale solar photovoltaics. An asset that is available roughly twice as often as wind, and nearly three times as often as solar, is a different financial instrument — and an asset that also sells heat and cooling from the same fluid is monetising output that a power-only project throws away.

That is the whole of the integration case, and it is worth stating without embellishment: the resource is not scarce, and the technology is not exotic. What is scarce is the capability to design a project around a customer's actual demand — electricity, heat, cooling, and the storage to shape them — and then to finance, build and operate it.

What we are not claiming

Two of the five forms deserve restraint.

Hydrogen is at pilot scale. The pairing is logical — high-temperature electrolysis wants exactly what geothermal supplies, firm power and high-grade heat — but what is running today are non-commercial demonstrations, explicitly framed as such by the operators building them. That is a pilot programme, not an industry.

In-reservoir storage is more advanced than most people assume and less proven than its advocates imply. The idea that an engineered reservoir can itself store energy for hours to days has peer-reviewed support — Ricks, Norbeck and Jenkins set it out in Applied Energy in 2022, and Nature Energy returned to flexible geothermal operation in 2024 — and it has been field-tested. It has not been demonstrated over many charge–discharge cycles at commercial scale. It is a credible engineering direction, not a settled capability, and we will describe it that way until it is one.

Nor is the sector large. Global geothermal capacity grew by about 0.3 GW in 2025; solar added 511 GW. Anyone arguing that geothermal is about to be the centre of the energy system is not paying attention.

The point

The case for geothermal does not rest on it being the biggest source of energy. It rests on it being one of the few that can deliver several forms of energy, continuously, from one resource, next to the customer that needs them.

Catoxy Energy is built to develop, operate and optimise projects on that basis — designing around the demand rather than around the turbine. We put that forward as the capability the company is being built to hold, not as a record it has yet established. The argument here is about the shape of the opportunity; the work of proving it is the company's to do.

---

Sources: International Energy Agency, Key Questions on Energy and AI (April 2026); IRENA, Renewable Capacity Statistics 2026 (March 2026); US Energy Information Administration, Electric Power Monthly, Table 6.07.B (2025 annual data); Lund, Huttrer & Toth, "Geothermal Development and Use, 1995–2024", GRC Transactions Vol. 48 (2024) — direct-use figures are the authors' 2023 estimate; Zhao et al., International Journal of Low-Carbon Technologies 19:1689–1698 (2024); European Commission, AccelerateEU — Energy Union, COM(2026) 370 final (22 April 2026); Fervo Energy first-quarter 2026 results (June 2026), for the Google framework agreement; W. Ricks, J. Norbeck & J. Jenkins, "The value of in-reservoir energy storage for flexible dispatch of geothermal power", Applied Energy 313:118807 (2022), and J. Jenkins et al., "The role of flexible geothermal power in decarbonized electricity systems", Nature Energy (2024).