23 months. Fervo Energy went from a graded pad in Beaver County, Utah to selling electricity, and on 30 September 2026 its first block began generating revenue one day before its power purchase agreement said it had to.

Enhanced geothermal stopped being a physics question years ago. It has been an accounting one: what a kilowatt costs to build, what a well costs to drill, and whether a lender will fund the difference. Cape Station answers all three at once.

The block synchronised to the grid on 24 September, reached 33 MW of net production, cleared the threshold written into its PPA, and cleared it in six days. Phase I is roughly 100 MW across three such blocks, the other two due by 1 January 2027. A further 400 MW is already under construction for a 2028 date, and contracted offtake at the site stands at about 900 MW.

The 23-month plant

Fervo broke ground at Cape Station in June 2023. TechCrunch put the distance between that date and commercial operation at 23 months — the shortest build anyone has published for a utility-scale enhanced geothermal system, or EGS, and the first greenfield EGS anywhere to reach contractual commercial operations.

The unit that reached that status is called a GeoBlock, and it is deliberately small. Thirty-three megawatts net, from a single production and injection loop, running under a PPA signed before the plant existed. GeoBlocks 2 and 3 have until 1 January 2027 to reach their own dates.

900 MW of contracted offtake is roughly the annual consumption of a million American homes. That figure is the pipeline talking. 33 MW is the plant talking. The distance between the two is this entire industry.

33 MW net output, GeoBlock 1

First commercial EGS block

Net production clearing the PPA threshold six days after grid synchronisation · Fervo Energy, 2026

Nothing about the 30 September date was ceremonial. Fervo declared commercial operation one day ahead of its contractual COD, which means the plant passed its performance test before the clock governing its revenue expired. A one-day margin is small in engineering terms and large in financing terms.

🎯
Cape Station is a cost milestone, not a physics milestone. The disclosed curve runs from $7,000 per kilowatt installed today toward a stated $3,000 per kilowatt, and the curve only matters if it survives a second and third block.

Non-recourse project debt closed at 100 MW scale in March 2026 — $421 million — which is the specific obstacle that keeps first-of-a-kind infrastructure unbuilt.

Drilling productivity, measured against Fervo's own first well, is up 143%. One well is a data point. Thirty is a business.

What changed between the pad and the plug

EGS is not a new idea. Hot dry rock has been the target for decades; what was missing was a way to make the reservoir behave like a reservoir. Fervo's answer borrows wholesale from oil and gas — long horizontal laterals, perforated and stimulated, treating rock as a production interval instead of a geological curiosity.

The disclosure trail shows the method compounding. In July 2026 the company reported third-generation well design lifting drilling rates 143% against its first Cape Station well, with Sawtooth 7 reaching 19,448 feet of measured depth including a 7,500-foot lateral in 21 days.

Then the part that decides whether any of it scales: money. On 19 March 2026 Fervo closed $421 million of non-recourse project debt for Phase I — $309 million construction-to-term, $61 million tax-credit bridge, $51 million letter of credit. Non-recourse means lenders have no claim on the rest of the company if the project disappoints. It is the reason first-of-a-kind plants usually never get built.

Non-recourse financing has historically been considered out of reach for first-of-a-kind projects. Cape Station disrupts that narrative.— David Ulrey, Chief Financial Officer, Fervo Energy

On 22 September, two days before grid synchronisation, Fervo announced two US Department of Energy awards for next-generation geothermal development in Idaho and Nevada. Public money is now underwriting the geology at three sites, and the third is not in Utah.

The cost curve is the whole thesis

Here is the figure that outranks the ribbon. Fervo says Cape Station generates electricity at about $7,000 per kilowatt of installed capacity, and names $3,000 per kilowatt as its target — the level at which it claims parity with natural gas.

Run that against the other direction of travel: new gas plant prices are up 66% in two years. A technology with a falling cost curve is now competing against fuel whose plant cost is rising, and the crossover Fervo named sits inside the range gas has already moved through.

Conventional geothermal has the opposite problem. It is cheap where the resource is exceptional — shallow, hot, wet, already proved — and it is not repeatable, because the good rock is finite and largely mapped. EGS trades a natural resource for an industrial process. That is a worse bet on geology and a better bet on manufacturing.

So the comparison worth making is not EGS against geothermal. It is EGS against everything else that has to be built, permitted, interconnected and financed to serve load that keeps growing. On installed cost, Fervo is not competitive today. On the slope of that curve, it is the only geothermal route with a published path to parity.

We wrote in September about Mazama Energy's 15-day well at Newberry. The comparison that matters is the shape of the number rather than the depth of the hole: both companies are selling drilling speed as the product, and both will be judged on well cost per kilowatt delivered.

The commercial layer now matches the engineering layer. Fervo listed on Nasdaq on 14 May 2026 under FRVO, selling 80.5 million Class A shares at $27.00 including the full over-allotment — about $2.2 billion gross, on a base raise Bloomberg put at $1.89 billion. A company that raises that much does not have to convince anyone that geothermal is interesting. It has to deliver the next four gigawatts on schedule.

Which puts the burden back on the rig floor. Every megawatt of EGS is a vertical slice through hot rock, and well cost falls with rig count, crew reuse and lateral length. Sawtooth 7 is a template. 900 MW is roughly thirty times GeoBlock 1. Plant engineering that produced one block in 23 months has to produce thirty without a fresh miracle each time.

ItemFigureDirection
Installed cost, Cape Station ≈$7,000 per kW ✗ above gas today
Stated end-state cost $3,000 per kW ✔ claimed gas parity
Drilling productivity 19,448 ft MD, 7,500 ft lateral in 21 days ✔ rates +143% vs first well
Project debt available $421M non-recourse, Mar 2026 ✔ first-of-a-kind scale
Competing plant prices New gas plants +66% in two years ✗ moving away from parity

Company disclosures and press reporting, 2026. Parity figures are Fervo's own stated targets, not third-party estimates.

What still has to be true

Induced seismicity first. Cape Station sits in seismically active Utah, and an EGS field fractures rock at depth on purpose. Fervo published an independent technical review of its mitigation protocol in August 2026 reporting positive learnings, and has published seismicity updates through the build. Positive is the right word. It is not the same as settled.

Second, well cost at depth. The 143% improvement is measured against Fervo's own first well — the easiest available baseline, and the one a sceptic reaches for. Deeper resource, more wells, and a $3,000-per-kilowatt target all pull the same direction.

Third, water and interconnection. A plant that runs around the clock is a large water user and a large load to connect, and neither problem is exotic. Both are slow. Slow is what breaks schedules.

Does $3,000 per kilowatt actually close the gap with gas?

Close enough to matter, and not yet close enough to bank. $3,000 per kW covers the capital cost of the plant; it does not include the levelised cost of the electricity it produces, which depends on capacity factor, parasitic load and how much reinjection the rock tolerates.

Confirmation criterion: a published all-in cost per kWh at Phase I scale, not an installed-cost figure. Until that number exists, parity is a target line, not a result.

Who signs for 900 megawatts

The offtake is the least speculative part of the story. Fervo signed 396 MW with Google on 1 September 2026 — the largest enhanced geothermal PPA on record — and in March 2026 agreed a development framework with Google covering up to 3 GW through 2033, with 1 GW proposed in the first two years. Earlier agreements cover Southern California Edison, Shell Energy and community-choice aggregators.

A framework is not a plant. The 2028 COD for the next 400 MW and the 2033 horizon on the Google framework are both dates Fervo has put in public filings, and both will be tested against capital markets rather than against rock physics.

Will Fervo's installed cost reach $3,000 per kilowatt before 2030?

🔮
By 2028 Fervo will have published an all-in cost per kilowatt-hour for Cape Station Phase I, and it will land above today's $7,000 per kW figure rather than at the $3,000 target.

Probability: 70% — disclosed drilling gains compound quarterly while the company keeps capital commitments to contracted revenue; the binding constraint is well cost at depth, not plant design.

✅ Arguments for

$421M of non-recourse debt already closed at 100 MW scale, with a lender appetite that has no precedent for first-of-a-kind geothermal

396 MW contracted with Google plus a 3 GW framework, which funds drilling without waiting on equity markets

Confirmation criteria: a Phase II non-recourse close and a fourth-generation well design published with a per-well cost

❌ Arguments against

Every productivity gain so far is measured against a first well drilled at a single site

Wells get deeper as the easy rock is consumed, and $3,000 per kW leaves little room for that curve

Refutation criterion: a published per-well cost above $9M, or an induced-seismic event above the protocol threshold at Cape Station
📊
Signals worth tracking

Commercial operation dates for GeoBlocks 2 and 3 against the 1 January 2027 commitment

The next published well cost per well, and whether a fourth-generation design appears at all

A non-recourse financing close on the next 400 MW, or an equity raise instead

Any seismic event above the threshold in Fervo's own mitigation protocol

Development scenarios

🟢 Optimistic (30%)

Both remaining GeoBlocks reach commercial operation on 1 January 2027, and the next 400 MW is financed on non-recourse terms by mid-2027.

Consequence: $3,000 per kW becomes a schedule item rather than a target, and the 3 GW Google framework turns into a build plan.

🟡 Base (50%)

Phase I completes with one block slipping a quarter, well costs fall without reaching the level that makes $3,000 per kW credible before 2030, and the next 400 MW carries more sponsor equity.

Consequence: Fervo builds gigawatts slowly and profitably, and the sector still waits on a second operator to prove the model is not company-specific.

🔴 Pessimistic (20%)

A seismic event above the protocol threshold triggers a Utah curtailment, or well costs rise with depth faster than drilling rates improve.

Consequence: the 2028 COD moves right, and $3,000 per kW moves out of reach for the sector as well as the company.
Fervo declares commercial operation at Cape Station, one day early
The primary record: grid synchronisation on 24 September, commercial operation on 30 September, 33 MW net from the first block, and revenue starting under the PPA.
Fervo Energy via GlobeNewswire
A company press release on its own milestone, and still the only source precise enough to check the dates against. Worth reading for the six-day gap between synchronisation and commercial operation.
A first-of-a-kind geothermal plant finished in 23 months
Independent confirmation of the build window, plus the $7,000-per-kW installed cost and the $3,000-per-kW parity target Fervo attaches to it.
TechCrunch
The cost figures come from the company, but the schedule is checkable against a groundbreaking date anyone can verify. That combination is what makes the milestone worth analysing.
Cape Station, the data-centre power squeeze, and the scale test
Written a month before first power, framing the open question: whether enhanced geothermal can reach cost parity at scale while demand for firm carbon-free power accelerates.
CNBC
Useful as a record of what was still uncertain before the milestone — the sequencing here is the opposite of a launch story.