Fervo's Cape Station: Who Gets the Geothermal, Who Pays for the Grid?

Fervo Energy synchronized the first 33-megawatt block of its Cape Station geothermal project to the grid on September 24, 2026, marking what the company calls the first utility-scale enhanced geothermal system (EGS) to reach commercial operation anywhere in the world.^[1] The milestone is real. The technology is sound. But the energy narrative now being written omits the structural question that will determine whether this becomes a template for clean private power or another mechanism to extract ratepayer subsidy.

Start with what is clear: Fervo's Cape Station Phase 1 will generate roughly 100 megawatts.^[4] Three 33-megawatt blocks are scheduled to synchronize by January 1, 2027.^[4] The company has announced a deal to supply almost 400 megawatts, a future phase, to Alphabet Inc. for a prospective Utah data center.^[5] Fervo went public in May 2026 with backing from Bill Gates.^[1] Those facts are in the record. What is not in the public record is what comes next: the tariff architecture, the collateral, the minimum-take obligations, and above all, who bears the cost if that data center load does not materialize or if the grid upgrades required to integrate it exceed what the customer is contractually bound to pay.

The promotional frame is geopolitically appealing: next-gen geothermal offers baseload without intermittency, and it is being deployed precisely where AI compute wants it, Utah, near fiber, near data centers.^[1] But that frame is incomplete. Enhanced geothermal systems drilled on the oil-and-gas model may be scalable and deployable anywhere.^[4] That scalability does not exempt them from the tariff structure that determines whether the public or the paying customer absorbs stranded capacity. Fervo is a private company. The grid is a regulated monopoly in Utah and California. When a startup signs a confidential power-purchase agreement with a utility or a large end-use customer, the deal's economics, term length, demand ratchet, exit fees, and cost allocation for dedicated network upgrades, determine whether ratepayers end up funding a subsidy to a venture-backed energy company. Fervo's announcement does not disclose any of those terms for the Google deal or for the Phase 1 offtake to Southern California Edison or other regional utilities.

The pattern is now familiar from data-center tariff work by researchers at Harvard's Electricity Law Initiative and from Grid Strategies' interconnection queue analysis. A large industrial load (whether a data center or, now, a geothermal supplier signing a fixed-price export contract) announces a multi-phase capacity build. The utility's integrated resource plan is amended to reflect the new load, justifying transmission and generation capex that ratepayers fund. The customer contract is sealed (redacted prices, term, ratchet). If the announced load does not materialize, if Google changes its Utah strategy, or if Fervo cannot fulfill its 400-megawatt promise, the capacity that was socialized into rates remains on ratepayers' books for decades. That is the unasked question: Does Fervo's contract with its future Google customer require a 10-year, 85-percent minimum-take demand charge, matched to the life of any dedicated transmission upgrades? Or is it a looser arrangement that leaves utilities and ratepayers holding the reserve margin risk?

The honest case for EGS is that it can be financed and deployed by the company that benefits from it, not by a regulated monopoly. Fervo itself went to the capital markets successfully in May 2026. If the Cape Station build is economically sound, and if Google or other data-center customers truly value reliable, sited-to-location geothermal, the venture should be able to finance generation, transmission, and integration costs itself or through project finance, not by securing utility approval of special tariffs that socialize cost risk. That model is available now: bring-your-own-generation, with the customer bearing 100 percent of dedicated network costs and a collateral structure that ensures the utility is made whole if the offtake fails. Several states, including Virginia and Ohio, have approved large-load tariffs with those protections embedded, high minimum-take ratchets, long terms, and cost isolation so the customer class, not residential households, carries the risk. Utah does not yet have such a framework. Utah regulators have a window to demand one before Fervo's full Cape Station expansion and any follow-on agreements are locked in.

The geothermal technology itself deserves deployment. The grid integration should not. Ask the Public Service Commission of Utah: Will Fervo's power-purchase agreements for Cape Station include a minimum-take ratchet of at least 85 percent of contracted generation and transmission capacity over the contract term? Will the term match the asset life of any dedicated network upgrades? Will collateral cover unamortized investment if the customer exits? Will all network costs be assigned to the customer class driving the load, not socialized system-wide? If the answer to any of these is no, the public is funding Fervo's build while a private company and its venture backers capture the returns.