The Geothermal Surprise: Clean Baseload Energy Has Been Hiding Under Your Feet

There is a quiet revolution happening in energy that almost nobody is paying attention to.

It is not fusion. Fusion is 15 years away, as it has been for the past 50 years. It is not small modular nuclear reactors, which I write about extensively and believe in deeply, but which face a decade-long licensing gauntlet in every jurisdiction. It is not a new battery chemistry or a breakthrough in solar efficiency.

It is drilling. Specifically, it is a group of former oil and gas engineers in Houston who realized that the techniques perfected over 15 years of the shale revolution - horizontal drilling, multistage completions, fiber-optic monitoring, pad drilling - could be pointed downward into hot rock instead of sideways into shale. And that doing so would unlock clean, 24/7, baseload energy virtually anywhere on Earth.

The company is called Fervo Energy. Their first commercial-scale project delivers power in October 2026. And the numbers are so good that they deserve far more attention than they are getting.

The constraint that limited geothermal for 50 years

Conventional geothermal energy requires a geological miracle: hot rock, natural underground water, and permeable pathways between them, all co-located near enough to the surface to drill economically. This trifecta exists along tectonic plate boundaries and volcanic hotspots - Iceland, parts of California, Indonesia, Kenya's Rift Valley - and almost nowhere else.

The result: after five decades of development, global geothermal capacity stands at roughly 17 GW. The United States, despite having the largest geothermal resources on the planet, has 4 GW installed. For context, total US generation capacity is 1,300 GW. Geothermal is a rounding error.

This was never a physics problem. The Earth's thermal gradient averages 25-30C per kilometer of depth. At 3-4 km, rock temperatures reach 150C almost everywhere. At 5-7 km, you are above 200C across most continental crust. The heat is universal. What was missing was a way to access it without geological luck.

What enhanced geothermal actually is

Enhanced Geothermal Systems eliminate two of the three constraints. You only need the heat. You supply the water and create the permeability yourself.

The process:

1. Drill a vertical well to 2-5+ km depth, where rock temperatures reach 150-300C
2. Kick off horizontally using directional drilling - the same polycrystalline diamond compact bits, measurement-while-drilling tools, and rotary steerable systems used in shale
3. Hydraulic stimulation - pump fluid under controlled pressure to re-open pre-existing natural fractures in the hot dry rock (this is hydro-shearing, distinct from fracking: the goal is to reactivate existing fractures, not create new ones in sedimentary rock)
4. Drill a second well into the fractured zone
5. Inject cold water down one well, let it heat as it flows through fractured hot rock, extract hot water from the other well
6. Generate electricity at the surface via binary cycle or flash steam
7. Reinject the cooled water - closed loop at surface, open loop underground

None of this is exotic. Every technique was developed and perfected by the oil and gas industry at a cost of over $1 trillion in R&D and operational learning over the past two decades. The shale revolution was, in hindsight, a trillion-dollar training program for geothermal.

Fervo: the numbers

Fervo Energy was founded in 2017 by Tim Latimer, a former BHP drilling engineer who left to study at Stanford, and Jack Norbeck. They are based in Houston. They hire from the oil and gas talent pool. They use the same rigs, service companies, and equipment as shale operators.

They have raised $1.35 billion across 19 rounds. Their investors include Breakthrough Energy Ventures, Devon Energy (the largest pure-play US E&P company), Google, CPP Investments, Mitsubishi Heavy Industries, and JB Straubel (Tesla co-founder).

Here is what they have built:

Project Red (Nevada, 2023) - Proof of Concept

• First-ever horizontal wells in an EGS system
• Depth: 2,400 m, resource temperature: 185C
• 30-day well test: 63 liters/sec flow rate, 3.5 MW per well
• Came online November 2023, delivering 24/7 carbon-free power to the Nevada grid
• 12 months of operation: zero thermal decline
• Powers Google data centers

Cape Station (Utah, 2024-2026) - Commercial Scale

• Depth: 4,805 m, resource temperature: 204-260C
• 15 wells drilled
• Production test: 107 kg/s flow rate, 10+ MW per well - triple the output of Project Red
• Called "the most productive enhanced geothermal system in history"
• BLM approval for up to 2 GW at full build-out
• Phase I: 100 MW, commercial operation October 2026
• Phase II: 400 MW by 2028
• 400 MW fully contracted

Project Blanford (Utah, 2026) - Next Generation

• Appraisal well drilled in under 11 days
• Resource temperature: exceeding 290C at 3,400 m depth - hottest well in company history
• Independent assessment: multi-gigawatt resource potential
• Sedimentary reservoir (easier to drill than granite, expanding global applicability)

The drilling performance improvement is what makes this credible:

Fervo's drilling performance follows a 35% learning rate - every doubling of wells drilled cuts drilling time by 35%. They are already exceeding NREL's 2035 "moderate technology improvement" projections. In 2024. A decade early.

The PPA that should have made headlines

In June 2024, Fervo signed two 15-year power purchase agreements with Southern California Edison totaling 320 MW. It is the largest geothermal PPA in history.

Combined with their Google/NV Energy contract, Cape Station has 373 MW fully contracted from investment-grade buyers before it has delivered a single commercial megawatt-hour.

This is not a science experiment. This is a power plant with signed, bankable contracts.

How it compares

The relevant comparison for EGS is not solar or wind - those are intermittent sources serving a different grid function. The comparison is other firm, 24/7 baseload power:

EGS is already roughly half the cost of new nuclear and competitive with natural gas. The DOE target of $45/MWh by 2035 would make it the cheapest firm baseload power on the grid. And unlike nuclear, it does not require a decade of licensing and regulatory navigation.

The "geothermal anywhere" thesis

This is the part that changes the strategic calculus.

The Earth's heat is not localized. It is everywhere. The question is only how deep you have to drill:

• At 3-4 km: viable in volcanic and high-gradient regions (western US, Iceland, East Africa)
• At 5-7 km: viable across most of the continental US, much of Europe, China
• Beyond 7 km: viable almost everywhere on continental crust

The DOE estimates the US alone has 70+ TW of technical EGS potential. For context, total global electricity generation capacity is 8 TW. The US has enough heat underground to power the entire world eight times over.

Land footprint: EGS needs 7.5 km2 per TWh of production - 5x less than solar, 10x less than wind.

Water usage: lower per MWh than fossil fuel generation, and the DOE estimates 90%+ of future EGS can use degraded or non-potable water.

This is not a niche technology for lucky geographies. This is a universal baseload energy source.

The workforce argument nobody is making

There are approximately 300,000 people in the United States with the skills needed for geothermal work - directional drilling, completions engineering, reservoir management, well integrity. The current geothermal workforce is under 9,000.

These are oil and gas workers. They do not need to retrain. They do not need new certifications. They need new employers. The skills are directly transferable because the technology is directly transferred.

Devon Energy, the largest pure-play US exploration and production company, is one of Fervo's biggest investors. Fervo is headquartered in Houston. The DOE launched GEODE, a program specifically to bridge oil and gas expertise into geothermal, in September 2024.

This is one of the only clean energy technologies that is politically viable in oil states - Texas, Oklahoma, North Dakota, Louisiana - because it does not ask the existing workforce to disappear. It asks them to point the drill bit at a different target.

The competition

Fervo is not alone. The next-generation geothermal space is filling up:

Quaise Energy (MIT spinout) is developing millimeter-wave drilling - using a gyrotron to vaporize rock with electromagnetic energy instead of a physical drill bit. In September 2025, they demonstrated live drilling through granite at 5 m/hour. Their target is superhot rock at 400C+, which could yield 5-10x more power per well. Still early - they have drilled to 100 m; commercial depth is 5-20 km. But if it works, it makes the economics absurd.

Eavor Technologies (Calgary) has built a closed-loop system - no hydraulic stimulation, no water loss, no seismicity risk. Their Eavor-Loop circulates fluid in sealed pipes, harvesting heat by conduction. In December 2025, they delivered first-ever closed-loop geothermal power to the commercial grid in Bavaria, Germany. Lower output than Fervo's approach, but zero seismicity risk makes it viable in earthquake-sensitive regions.

Sage Geosystems (Houston) is building what amounts to an underground thermal battery - geothermal power generation plus grid-scale energy storage in one system. They have a 150 MW contract with Meta and a $25M investment from Ormat Technologies.

Zanskar Geothermal uses AI-driven exploration to find overlooked conventional geothermal resources - their thesis is that 1 TW of geothermal is hiding in plain sight. They raised $115M in January 2026 and have six 20 MW plants in their pipeline.

The challenges are real

I would be dishonest if I did not address the risks.

Induced seismicity is the most politically sensitive. Hydraulic stimulation can trigger felt earthquakes. Basel, Switzerland (2006) saw a M3.4 earthquake from a deep geothermal project, causing $9M in insurance claims. Pohang, South Korea (2017) triggered a M5.5 earthquake that injured 82 people and damaged 2,000 houses. Both projects were terminated.

Fervo's hydro-shearing approach is lower risk than the techniques used at Basel and Pohang, but the risk is not zero. Public perception of any earthquake, regardless of magnitude, can kill a project.

Well longevity is unproven at scale. Thermal drawdown of 1-4% annually at production flow rates means a well may only sustain economic output for 10-20 years before the rock cools below threshold. Fervo's Project Red showed zero decline over 12 months - promising but early.

Scaling requires a massive expansion of the drilling fleet. Even leveraging oil and gas equipment and crews, building out gigawatts of EGS means thousands of wells drilled per year. That is achievable - the US drills roughly 40,000 oil and gas wells annually - but it requires deliberate redirection of industrial capacity.

Why this matters more than you think

I have written extensively about thorium molten salt reactors as the long-term answer to civilization-scale energy. I still believe that. But thorium is a 2035-2040 technology in the best case - the physics is proven, but the licensing, regulatory, and manufacturing challenges are real and substantial.

EGS is a 2026 technology. It works now. It scales on existing industrial infrastructure. It employs existing workers. It does not require new regulatory frameworks. The first 100 MW commercial plant comes online in seven months.

The DOE's target trajectory: 5 GW by 2030, 38 GW by 2035, 90 GW by 2050. For reference, the current US nuclear fleet is 95 GW and took 40 years to build.

Here is the timeline I am watching:

• October 2026: Cape Station Phase I delivers 100 MW. If Fervo hits this on time and on budget, it is the proof point that unlocks the next wave of capital. This is the Falcon 9 moment for geothermal.
• 2028: Cape Station reaches 400 MW. Blanford development begins. Multiple other EGS companies reach commercial scale.
• 2030: 5 GW installed. EGS becomes a recognizable grid resource.
• 2035: 38 GW. EGS is roughly 40% of current nuclear capacity, built in a fraction of the time. LCOE approaches $45/MWh.

The most significant clean energy technology of the next decade might not be the one that wins the most headlines. It might be the one that quietly borrows the shale industry's playbook, points it at hot rock, and delivers clean baseload power to a grid that desperately needs it.

The heat has always been there. We just learned how to reach it.