Enhanced Geothermal and the Rupiah: Firm Heat, Imported Capex, and the Grid-Confidence Ledger

Rupiah Stability Watch · 2026-10-13

The premise

Fervo Energy’s Cape Station in Utah is not an Indonesia currency event. It is a threshold signal in a ledger Indonesia already lives with: imported fuel, diesel backup, cooling load, data-center reliability, and the confidence cost of a grid that must serve an archipelago.

IEEE Spectrum reported that Fervo is now selling electricity from Cape Station, with an initial 33-megawatt unit synchronized to the grid on 24 September and commercial operation beginning six days later. The report describes the plant as the first utility-scale “next-generation” geothermal facility to reach that point, and notes a planned 500 MW buildout, including two more 33 MW blocks by 1 January 2027 and a further 400 MW phase supported by a Google agreement for 2028. Fervo’s own releases say the first GeoBlock reached 33 MW net output, Phase I is roughly 100 MW, and Cape Station has around 900 MW of contracted offtake across the larger development.

That matters for Rupiah Stability Watch because it sharpens a theme from our earlier work. In “Duration Is the New FX Variable,” we argued that the value of clean power changes when it can displace diesel and stand through long reliability gaps, not only when it adds nameplate capacity. In “The Data-Center Trip Test” and “Data-Center Power Demand and the Rupiah,” the problem was not simply more megawatts for compute; it was whether Indonesia can promise firm, trusted power without importing a new burden of equipment, fuel, and dollar finance. “Neglected Energy Buffers and the Rupiah” made the same point from the other side: small, local, reliable buffers matter when they avoid imported fuel at the edge. “Hourly Heat Load and the Rupiah” tied cooling demand to peak-load stress. “Nuclear Readiness and the Rupiah” warned that clean-energy ambition can either reduce external pressure or create new hard-currency, fiscal, and safety liabilities.

Enhanced geothermal belongs in that same ledger. It should be read as a comparison case, not a ready-made Indonesian answer.

What the evidence supports

The Cape Station evidence supports a limited claim: engineered geothermal has crossed from pilot demonstration toward utility-scale grid supply.

IEEE Spectrum says Cape Station’s initial 33 MW unit is selling power to the grid and will supply Shell Energy, a California retail power provider. It also reports that Cape Station is a $2 billion project planned to reach 500 MW, and that Fervo uses oil-and-gas drilling techniques to create reservoirs in hot, dry rock rather than relying only on naturally occurring hot water or steam. Fervo’s September release gives the technical framing: horizontal drilling adapted from the oil and gas industry, multiple wells from a single pad, and engineered access to geothermal resources previously inaccessible by traditional methods. Its October release adds that the first GeoBlock reached contractual commercial operation, generated revenue under a power purchase agreement, and produced 33 MW net.

The cost evidence is less settled. IEEE Spectrum reports that Fervo had told Canary Media Phase I was set to deliver power at about $7,000 per kilowatt and Phase II at about $5,500 per kilowatt, while comparing that with an average 2025 natural-gas plant construction cost of $2,157 per kilowatt. Fervo has not disclosed the terms of its power purchase agreements. So the signal is not “cheap geothermal has arrived.” It is narrower: a firm clean-power technology with high upfront drilling and construction cost has reached commercial operation at a scale relevant to utilities and large loads.

Indonesia’s geothermal baseline is stronger than most countries’, but still underused. Indonesia’s energy ministry reported at the 2026 Indonesia International Geothermal Convention & Exhibition that national geothermal potential is 23,202.77 MW, while installed capacity is 2,776.39 MW — roughly 11.6 percent of the potential. The same ministry release says about 88.4 percent remains undeveloped and frames geothermal as part of energy sovereignty and lower dependence on imported energy. In the 2025–2034 PLN electricity supply plan, the ministry said Indonesia planned 69.5 GW of added generation capacity by 2034, with renewables and storage making up about 76 percent of the total. The listed renewable additions include 5.2 GW of geothermal, alongside solar, wind, hydro, bioenergy, storage, and early small modular nuclear units.

This is the useful contrast. The United States signal is about opening engineered geothermal in places where conventional geothermal was not expected to work. Indonesia’s starting point is different: it already has large volcanic geothermal resources, operating plants, and institutional experience. Its bottleneck is less “does heat exist?” and more “can projects be permitted, drilled, financed, connected, operated, and trusted at the pace implied by the plan?”

Where the rupiah benefit could appear

The rupiah channel is indirect. It runs through the operating ledger.

First, firm geothermal can reduce fuel-import exposure when it displaces diesel or gas-fired backup rather than merely adding capacity to a congested grid. Indonesia’s ministry was explicit that geothermal is part of reducing dependence on imported energy in a period of geopolitical and energy-price uncertainty. For the rupiah, this matters most where avoided fuel is real, measurable, and recurrent: island grids, remote industrial sites, cold chains, public-service loads, ports, hospitals, and data centers that would otherwise protect themselves with imported fuel and equipment.

Second, geothermal has a grid-confidence value that solar nameplate alone does not provide. It runs day and night, subject to reservoir performance and maintenance. That is why Cape Station is relevant to the data-center problem even though it is in Utah. IEEE Spectrum ties next-generation geothermal demand to data centers and the need to balance intermittent solar and wind. Indonesia’s version of that problem is not whether an AI campus can claim a clean-power contract. It is whether PLN and local grids can support compute, cooling, manufacturing, and public infrastructure without turning reliability into a hidden import bill.

Third, geothermal can act as a geographic stabilizer. A plant near a load center or island system can reduce dependence on long fuel supply chains and fragile backup arrangements. This is where the earlier “Neglected Energy Buffers” frame matters: the rupiah benefit is not only national fuel statistics; it is the avoidance of repeated small hard-currency leaks when local systems must buy diesel, spare parts, and emergency services because reliability was not built into the grid.

Fourth, geothermal capability has an industrial-policy dimension. If Indonesia deepens local maintenance, reservoir management, drilling services, environmental monitoring, and operations capability, geothermal spending can circulate domestically. If it does not, the clean-power asset may still reduce fuel imports, but part of the pressure returns through imported rigs, turbines, control systems, specialist services, and foreign-currency debt.

Where the rupiah risk re-enters

The risk is that Indonesia treats geothermal as a clean-megawatt headline while the hard-currency liabilities sit off to the side.

The first liability is drilling risk. The World Bank’s Indonesia geothermal brief puts the problem plainly: exploration and drilling can be costly and risky, with resource exploration at a potential site costing $30 million to $50 million and no way to recoup losses if the resource is insufficient; later development of a 100 MW site can cost up to $400 million. That is the geothermal version of a currency trap. The country avoids imported fuel only after it has absorbed expensive upfront uncertainty, often with foreign capital, imported equipment, and long payback periods.

The second liability is finance. Deloitte’s review of Indonesia’s geothermal sector argues that high upfront costs, exploration and drilling risk, long payback periods, tariff gaps, regulatory uncertainty, and land-acquisition complexity remain central barriers. It also notes that geothermal costs are sensitive to financing costs because projects are capital intensive. ADB’s 2025 Muara Laboh Stage 2 package shows both progress and dependence: $92.6 million in arranged finance for an approximately 83 MW plant, including ordinary capital, a syndicated B loan from Sumitomo Mitsui Banking Corporation, and concessional climate finance from the Australian Climate Finance Partnership. That blended structure can be constructive. It also shows that the currency composition, tenor, guarantees, and risk allocation matter as much as the megawatt count.

The third liability is grid buildout. The 2025–2034 RUPTL announcement itself says that renewable construction depends on networks: if the grid is not ready, PLN can be left paying take-or-pay obligations for plants that cannot be fully absorbed. A geothermal plant that is physically firm but transmission-constrained is not a rupiah stabilizer. It is a stranded-cost risk.

The fourth liability is governance and social confidence. Geothermal projects touch land, forests, water, seismic concerns, and local consent. Fervo’s own forward-looking statements list permitting, environmental regulation, operational disruption, supply-chain dependence, interest rates, and capital availability as risk factors. Indonesia’s risks are different in detail but similar in kind: if communities see geothermal as imposed extraction rather than reliable local benefit, the confidence ledger weakens before the electricity arrives.

The least-harm path

Indonesia should not copy Cape Station. It should use Cape Station to improve the questions it asks of its own geothermal plan.

The least-harm path is a geothermal capability ledger with six measures.

1. Avoided diesel and fuel imports, not only added megawatts. Each project should state what imported fuel use it is expected to displace, where, and under what grid-dispatch assumptions.

2. Capacity factor and dispatch value. Firm power should be valued for the hours it actually protects: evening peaks, cooling loads, island reliability, industrial continuity, and data-center uptime.

3. Local operating depth. Track how much of drilling support, reservoir monitoring, plant maintenance, environmental monitoring, and emergency response can be done domestically over time.

4. Financing currency and risk-sharing. A project financed in dollars but paid in rupiah can still be valuable, but the exchange-rate exposure should be named at approval, not discovered under stress.

5. Grid bottlenecks. No geothermal target should be counted as a rupiah buffer until the transmission and offtake path is credible.

6. Transparent failure accounting. Failed wells, delayed permits, local conflict, and underperforming reservoirs should be reported as system-learning costs, not hidden until they become tariff disputes.

This is the same discipline we asked for in “Duration Is the New FX Variable” and “Nuclear Readiness and the Rupiah”: count the operating benefit and the imported tail together. Firm clean power is valuable precisely because it can reduce the need for fuel-backed confidence. But if Indonesia imports the drilling capability, borrows in hard currency, delays the grid, and socializes failed-well risk without clear disclosure, the pressure does not disappear. It changes form.

What I am uncertain about

The largest uncertainty is cost transferability. Cape Station’s reported cost trajectory is useful, but it does not tell us what enhanced geothermal would cost in Indonesian geology, regulation, procurement, and financing conditions.

The second uncertainty is whether enhanced geothermal is even the right frontier for Indonesia near-term. Indonesia’s conventional geothermal base is already large and underdeveloped. The immediate rupiah gain may come less from adopting Fervo-style engineered reservoirs and more from de-risking known fields, improving transmission, and using geothermal heat directly where it avoids diesel or process fuel.

The third uncertainty is local capability depth. Indonesia can turn geothermal into a domestic operating advantage only if it measures the supply chain below the headline plant: rigs, pumps, turbines, controls, reservoir models, maintenance crews, and environmental monitoring. Without that, firm clean power can still help the rupiah — but less than the megawatt announcement implies.

The threshold signal from Utah is real. The Indonesian conclusion should be sober: firm heat is valuable, but only when the ledger includes the imported capex, the financing currency, the grid path, and the public trust required to keep the power firm.