AI’s 24/7 power sprint just found a missing supply-chain map: geothermal’s drill-to-ORC-to-PPA chain

AI load growth has forced every power plan to answer the same question: not “can the grid run the average day,” but “can it run the worst 3 hours—every day.” Batteries and demand response help, but they don’t replace the procurement discipline utilities and large customers increasingly need: round-the-clock, clean output contracted in megawatts—not just energy.

Geothermal—specifically enhanced geothermal systems (EGS)—is being pulled into that procurement slot because it can be engineered and financed as dispatchable “baseload-like” capacity rather than a weather-dependent profile. The investment case gets real only when you can map the end-to-end chain that turns heat in rock into power at the meter.

Fervo Energy’s Cape Station program is designed to turn geothermal heat into contracted output with specific delivery milestones. In its SEC filings, Fervo describes expected first power from a 500 MW Cape Station project starting in Q4 2026, aiming to reach about 100 MW of operating capacity in Q1 2027 (subject to performance and contractual terms). Fervo’s PPA-backed build plan converts “geothermal potential” into scheduled, contractual megawatts.

On the offtake economics, Fervo’s Cape Station data-center-oriented PPA/offtake documentation cites a $7.2B potential revenue backlog and names Southern California Edison (SCE) and Shell Energy as offtakers for the fully contracted 400 MW output context.

In geothermal, “firm” is not a technology slogan; it’s execution risk priced into contracts. Fervo’s SEC filing shows how that execution is operationalized:

• expected delivery timing tied to PPA COD (commercial operation deadline)
• liquidated damages and termination rights if key milestones aren’t met
• a high level of permit/approval progress for Cape Station phases

This is why the supply chain is more than an ORC story. If drilling, completion, and permitting slip, the grid doesn’t get capacity—and the turbine OEM scope becomes stranded capital rather than recurring revenue. In other words: the critical path starts upstream at wells and approvals, then flows downstream into ORC equipment and ultimately into offtake revenue.

A load-following gas plant can substitute in fuel and dispatch; geothermal must substitute in equipment procurement and commissioning. For Fervo’s Cape Station Phase II, Baker Hughes was selected to deliver key surface power generation equipment for five ORC units.

Baker Hughes’ project description specifies the ORC configuration as five 60-MWe ORC units (total ~300 MW once operational under that Phase II scope) and identifies its surface equipment scope including turboexpanders and BRUSH generators.

That matters for investors because ORC turbine/generator OEM demand tends to scale with: (1) executed PPA MW, (2) COD schedules, (3) the number of standardized unit blocks.

Put simply: ORC OEM scope scales in unit blocks—so PPA MW you can underwrite becomes turbine/generator revenue you can model.

Gas often “wins” on speed until you hit fuel and transmission constraints and run into procurement caps or decarbonization policies. Nuclear often “wins” on baseload ambition but loses on schedule risk and long lead-time construction.

Geothermal’s bet is different: it is trying to compress the path from contracted MW to energization by focusing on repeatable drill-permit-commission cycles. When the upstream chain is executed, geothermal becomes a firm-power procurement tool that can be “arranged” around AI load growth without relying on dispatch improvisation.

The investor takeaway is not that geothermal is always faster than everything; it’s that the chain is built to meet a contract deadline the same way large-scale industrial projects do—meaning you can compare timing via disclosed COD targets and equipment scope, not just by headline technology timelines.

To ground the listed-side exposure, consider how ORC and surface generation suppliers look financially today and what their cash-generation profile implies. Baker Hughes reports FY2025 revenue of $27.7B with net income of $2.6B and EBITDA of $4.3B (from its FY2025 financial statements), giving it a base to absorb project-cycle volatility while it scales incremental geothermal work.

This doesn’t mean geothermal is immaterial—rather, geothermal order momentum can act like an execution multiplier inside a diversified industrial portfolio. For Ormat Technologies, the market is already pricing geothermal exposure at higher multiples; the key is whether additional 24/7 AI-driven offtake contracts convert into durable, repeatable unit deployments.

The investor job is to watch whether the upstream geothermal chain can deliver the CODs that unlock backlog and reduce termination risk—because that’s what transforms a supplier’s engineering line item into sustained revenue.

Short term (days to quarters): The earliest signal is whether Cape Station’s milestones move from “expectations” toward evidenced progress (operating capacity ramp and approvals). Fervo’s filing emphasizes expected delivery windows and the consequences of missing them, so investors should treat updates as risk repricing rather than marketing.

Long term (1–3 years): If standardized ORC unit blocks keep getting procured as PPA MW are executed, geothermal can become a bankable “firm clean” procurement template that AI customers and utilities repeat. That template could reduce reliance on gas’s dispatch improvisation and delay nuclear’s construction timelines.

The central thesis is not that geothermal is the only 24/7 solution; it’s that geothermal is the clean firm-power option most exposed to a contract-milestone execution feedback loop, which directly links upstream drilling success to downstream offtake revenue and supplier utilization.