Fervo Energy: A Full Guide to Cape Station, Enhanced Geothermal, the Google Deal, Costs and Limits

On September 30, 2026, a 33-megawatt geothermal plant in Utah started selling power. Fervo Energy says it is the first greenfield enhanced geothermal development in the world to reach contractual commercial operation, one day ahead of its deadline.

Six weeks later, Fervo reported $11.9 billion of contracted revenue, up from $7.2 billion a quarter earlier, and a 15-year Google deal for 396 megawatts. In the same report, third-quarter revenue was just $116,000. That gap between a fast-growing order book and a barely started power business is the Fervo Energy story right now.

This full guide covers how enhanced geothermal works, what is actually running at Cape Station, the drilling learning curve, the contracts, the money, the Google data center angle, induced seismicity, the rivals and the honest limits. Last updated: November 14, 2026.

Key Takeaways

• Cape Station GeoBlock 1 reached commercial operation on September 30, 2026 at 33 MW, built in 23 months.
• Fervo reports 1,054 MW under binding contracts and about $11.9 billion of contracted revenue, but only 36 MW in operation.
• Cape Phase I cost about $7,000 per kilowatt; the Phase II target is $5,500 and the long-term goal is $3,000.
• Google's 396 MW take-or-pay deal delivers first power in mid-2028, with an option near 1 GW.
• The main risks are induced seismicity, unproven cost cuts and 2028 execution: three Red-level events hit in September 2026.

Fervo Energy at a glance

Let's start with what's on the record. The table below uses Fervo's own filings and releases, so treat the targets as targets.

Source: Fervo Q3 2026 results for contracts, capacity, cash, cost and land; Fervo IPO pricing release for the $27 price; SEC Form 8-K for $2.174 billion gross proceeds; commercial operation release for GeoBlock 1.

Notice the unit: megawatts, not gigawatts. Fervo's reach is far larger than its output today.

How enhanced geothermal works

Conventional geothermal needs three things at once: hot rock, water and natural permeability. Few places have all three. Enhanced geothermal systems, or EGS, create the missing pieces by engineering a reservoir in hot, dry rock.

Fervo's version borrows from shale. It drills a pair of horizontal wells through hot granite, stimulates the rock between them with hydraulic fracturing, and circulates water from one well to the other. Cold water goes down, absorbs heat in the fractures and comes back up hot to drive turbines. The Verge described the Nevada pilot as two horizontal wells with fiber-optic cables inside, where the water is reused.

The fiber matters. Distributed acoustic sensing turns each well into thousands of seismic channels, which Fervo uses to map fractures and listen for earthquakes, according to its seismicity update.

The U.S. government sees a big prize. The Congressional Research Service says EGS could enable 90 gigawatts of U.S. geothermal capacity by 2050. DOE's Enhanced Geothermal Shot targets $45 per megawatt-hour by 2035, which would open access to what it calls five terawatts of heat resource.

From Project Red to Cape Station

Fervo's path is a staircase of pilots. In 2021, Fervo and Google signed the first corporate agreement for next-generation geothermal. Project Red followed: a pilot adjacent to Nevada's Blue Mountain field, rated at 3 MW gross and delivering power to the Nevada grid since November 2023.

Google's Nevada work then scaled up. Nevada regulators later approved a first-of-its-kind Clean Transition Tariff that adds 115 MW of around-the-clock enhanced geothermal for Google's Nevada data centers, delivered by NV Energy. Fervo's Corsac project in Churchill County covers it, with 150 MW of capacity targeted for 2030.

Cape Station is the real commercial test. Fervo began drilling there in 2023, and it's now the company's flagship, with about 900 MW fully contracted with offtakers across all phases.

The drilling learning curve behind the cost story

Geothermal economics are mostly drilling economics. Fervo's pitch is that each well design drills faster, deeper and hotter than the last, the way shale operators cut costs through repetition.

The numbers are specific. Fervo's first commercial well at Project Red in 2022 reached 11,220 feet in 70 days. In February 2024, its fastest Cape well took 21 days, a 70% reduction, and drilling costs across the first four Cape horizontal wells fell from $9.4 million to $4.8 million per well.

In July 2026, Fervo said its Sawtooth 7 well, the ninth 3.0-design well for Phase II, reached 19,448 feet in 21 days in a 460 degree Fahrenheit resource.

Source: Fervo, July 8, 2026. Longer laterals, bigger casing and hotter rock are meant to raise megawatts per well.

The tools came from oil and gas. Fervo used polycrystalline diamond compact bits from shale basins and mud coolers to handle heat, and sustained an average 70 feet per hour on its fourth Cape horizontal well, per its February 2024 release.

There's a catch. These are company-reported drilling results, not audited costs. Faster wells lower cost per kilowatt only if each well also produces more power, and the proof of that comes from the plants, not the rigs.

Cape Station status: what's built and what isn't

GeoBlock 1 is the only commercial unit so far. Fervo synchronized it to the grid on September 24 and declared commercial operation six days later, after building and commissioning the first GeoBlock in 23 months. The company says it can reach an 18-month timeline for future GeoBlocks.

The rest of Phase I is close. As of November 9, GeoBlock 2 had reached grid sync and was exporting test energy, and GeoBlock 3 was in cold commissioning. Both are expected to reach commercial operation by January 1, 2027.

Phase II, branded Cape Google 400, covers GeoBlocks 4 through 11. Fervo says the first set of 50-megawatt turbines passed factory acceptance testing, more than 60 percent of the air-cooled condensers are on site, and drilling is about 18 percent complete.

One more number is easy to miss. A new independent heat-in-place study by DeGolyer and MacNaughton supports about 8.4 gigawatts of capacity potential at Cape Station, almost double the earlier 4.3 gigawatts. That's a resource estimate, not a reserve or a construction plan.

Who's buying Fervo's power

Demand is not Fervo's problem. California's mid-term reliability mandate, which requires utilities to procure 1,000 MW of firm clean power, created the first big wave of buyers, as Fervo's Shell Energy release explains.

Source: Fervo Q3 2026 results for the Google Cape deal; Fervo SCE release for the 320 MW; Google for the 115 MW tariff; Fervo Shell release for the 31 MW. Fervo's total of 1,054 MW under binding contracts also includes smaller deals.

Pricing isn't public for most of these contracts. Fervo has said the Google agreement has fixed terms, which is the right question to keep asking: how many dollars per megawatt-hour, and for how long?

The money: IPO, cash burn and the revenue gap

Fervo went public on May 13, 2026. It priced its IPO at $27 per share, and with the overallotment sold 80.5 million shares for about $2.174 billion gross.

The Q3 report shows a company spending heavily before revenue arrives. Q3 capital spending was $256.7 million, with $450 million to $500 million expected in Q4, and the operating loss was $43.7 million on revenue of $116,000. Fervo ended the quarter with $1.9 billion in cash and $301.6 million in debt.

Net income was positive at $11.7 million, but only because of a one-time $29.9 million tax benefit and an $11.8 million gain on interest rate swaps, according to the same release. Read the operating line, not the headline.

Fervo said Q4 will be its first full quarter of Cape revenue. That's the first number worth tracking after January.

Cost per kilowatt: $7,000 today, $3,000 someday

Capital cost is the number that decides whether geothermal competes with gas, nuclear and solar plus storage. Fervo says GeoBlocks 1 to 3 will cost about $7,000 per kilowatt in total.

For Cape Google 400, Fervo targets $5,500 per kilowatt, and its long-term goal is $3,000 per kilowatt. Fervo's Phase II numbers rest on the 3.0 well design producing substantially more megawatts per well, which hasn't been proven at plant scale yet.

One caution on the $514.3 million Fervo moved into service for GeoBlock 1. The company says that isn't the standalone cost of one block, because it includes shared wellfield and infrastructure for all three Phase I blocks plus financing costs. Don't divide it by 33 MW and call it a price.

The Google deal and the data center angle

Google's September 2026 agreement is the largest enhanced geothermal contract signed so far, by Fervo's account. It covers 396 MW across eight GeoBlocks, with first power in mid-2028 and full capacity targeted by year-end 2028.

The structure is unusual. On Fervo's earnings call, Google agreed to pay over 15 years even if it never builds the proposed Beaver County data center. Fervo also offered Google an option to add about 600 MW, for nearly 1 GW in total, which the companies plan to deliver behind the meter under Utah's SB132 framework.

That makes the Utah project a live test of a trend Inside Deep Tech has tracked across AI power: when interconnection queues stretch for years, buyers try to place generation next to the load. See Inside Deep Tech's behind-the-meter generation guide for the gas, fuel cell and battery options geothermal now competes with.

Behind-the-Meter Generation for AI Data Centers: A Full Guide

Inside Deep Tech's full guide to behind-the-meter power for AI campuses, interconnection queues and honest 2026 to 2028 timelines.

Inside Deep Tech

The scale is the honest limit. Google's Cape deal is hundreds of megawatts. The gigawatt campuses covered in Inside Deep Tech's Meta Hyperion guide need far more, which is why hyperscalers sign with several firms at once.

Induced seismicity: the risk Fervo can't hide from

Fracturing rock to build a reservoir can trigger earthquakes. Fervo uses a traffic-light protocol: events under magnitude 2.0 are Green, 2.0 to 3.0 are Yellow and pause operations for at least six hours, and 3.0 or above are Red and pause them for at least 24 hours.

The record is public. Since February 2024, Fervo's system has detected 167 Yellow events and three Red events. All three Red events came in September 2026 within about two weeks, from stimulation of one newer well pad, measuring roughly M3.0 to M3.3 on the vendor network and M2.6 to M3.1 in the USGS and University of Utah catalogs.

Ground motion in nearby Milford was very weak or not felt, and Fervo reports no injuries or damage. After the events, it cut injection rates near the heel of the wells and skipped intervals closest to a nearby fault.

Two older projects show why regulators care. Basel's 2006 stimulation triggered a magnitude 3.4 quake and the project was terminated in 2009. In Pohang, South Korea, an EGS project was confirmed as the cause of a magnitude 5.5 earthquake that injured over 100 people and caused $300 million in damage.

Fervo Energy vs the other next-generation geothermal players

Fervo leads on built capacity, but it isn't alone. The table compares the main U.S. next-generation developers on approach and stage, using company and trade-press sources.

Source: Fervo Q3 2026 results; Renewability.net on Sage's Project Vector at Ormat's Blue Mountain plant, its 150 MW Meta deal and Quaise's Oregon plans; XGS Energy on its 150 MW New Mexico project with Meta.

Sage has a different bet: supply heat to an existing Ormat power plant, so no new turbines or grid connection are needed. Quaise aims much hotter, with a long-term bet on millimeter-wave drilling that is still unproven.

Compared with the other firm clean power options Inside Deep Tech covers, geothermal's pitch is timing. Nuclear is a long haul: Inside Deep Tech's small modular reactor guide explains why most of that electricity won't arrive this decade, and its X-energy Xe-100 guide shows the licensing path. Fusion is earlier still, as the Commonwealth Fusion Systems SPARC guide lays out. Geothermal is the one already selling electrons.

Inside Deep Tech's take

Honest limits of what is known

• Tiny operating base. Only 36 MW is running, against more than 1,000 MW contracted, per Fervo's Q3 report.
• Costs are targets. $5,500 and $3,000 per kilowatt are goals; the delivered Phase I figure is about $7,000.
• Company-reported drilling data. Well times and costs come from Fervo releases, not independent audits.
• Long-term thermal output is unproven. Fervo says Project Red shows no premature thermal decline, but that is a 3 MW pilot, not a decades-long record.
• Seismic risk is live. Three Red-level events in September 2026 show the traffic-light system works and that the hazard is real.
• Resource estimates aren't reserves. The 8.4 gigawatt Cape figure is a capacity potential, not planned construction.

FAQ

What to watch next

• GeoBlocks 2 and 3 reaching commercial operation by January 1, 2027, which would complete Cape Phase I.
• Fervo's first full quarter of Cape revenue in Q4 2026, plus the $450 million to $500 million capital spending guide.
• Output per well and installed cost on Cape Google 400, against the $5,500 per kilowatt target.
• Whether Google exercises its roughly 600 MW option, which reporting says runs to June 2030.
• Sage's first Blue Mountain well, due to start drilling in Q4 2026 per Renewability.net.
• Fervo's next seismicity update and the formation of its independent advisory panel.

Fervo has turned enhanced geothermal from a pilot into a power plant. The next test is repetition: if the second and third GeoBlocks deliver on time in January 2027 and Phase II keeps its cost promise, geothermal becomes a serious answer to AI's power problem. If not, it stays a promising niche.

Cover background: Fervo Energy Cape Station Phase 1 construction, August 2025, photo by Braddah n8, CC0 via Wikimedia Commons, cropped, darkened and blurred.