Fusion, and other fusion-level bets: Part II
The next great civilizational energy resources?
Reminder: This is Part 2 of a three-part series series on “fusion-level” technology pathways in the energy sector, which is the foundation of human civilization. My goal is to shed some light on the most transformational technology worth betting on for the next hundred years, or so.
Part 1 focused on nuclear energy — both fission and fusion — and made the case that fusion is, ironically, the least “fusion-level” out of all the big energy technology bets I’ve identified. (If you haven’t read Part 1 yet, I’d recommend starting there.)
Here in Part 2, I’m going to zoom in on three additional technology categories: geologic hydrogen, “crazy cheap” renewables, and geothermal “anywhere”. Just like nuclear energy, these three categories could theoretically provide us with cheap, abundant, zero-carbon energy supply at a civilizational level. While these categories are unlikely to be as flexible as nuclear, which can be deployed pretty much anywhere on the planet at varying scales, all three could make a major difference to the primary energy mix at a global level.
For each category, I’m going to discuss:
The Prize: What makes this a fusion-level bet?
The Challenge: What makes this technology so difficult and risky to try to commercialize?
The Path: What would a successful path to commercialization for this technology look like?
Let’s dive in.
Geologic Hydrogen
The Prize:
Hydrogen is, in most ways, a relatively poor choice of molecule for a fuel. I wrote a lot about hydrogen’s virtue & vices just about two years ago, during a brief period where governments and corporations all over the world were going absolutely bonkers for the stuff.
In principle, hydrogen could be an attractive net-zero carbon energy carrier for a wide range of energy end uses. After all, our current economy is founded on hydrocarbons; and in hydrocarbon molecules it’s the bonds between carbon & hydrogen that do all of the heavy lifting. All of those pesky carbon atoms are mostly just contributing density & stability, and of course they’re wrecking the climate.
Hence, transitioning to hydrogen alone seems like an elegant solution. Hydrogen could theoretically serve as a nearly universal energy “medium of exchange”, as a fuel for everything from vehicles, to power generation, to industry, to home heating. That’s why hydrogen is sometimes described as a “swiss army knife” for the energy transition.
I agree that “swiss army knife” is a pretty apt metaphor for hydrogen. But unlike hydrogen’s many admirers, I don’t view this metaphor as especially flattering.
Consider: Swiss army knives are undoubtedly fun to play with, and they’re great for camping trips when you can’t carry all of the individual tools you might need. But cramming all of those tools into a single form factor makes them all more awkward to work with. Pop quiz: How often have you seen someone on an actual job site pull out a swiss army knife?
Hydrogen is also awkward to work with… and that’s putting it mildly. In a prior post on energy storage, I referred to hydrogen as a “pain in the ass molecule”. To begin with, it has very low volumetric energy density. Even when pressurized to 700 bar, hydrogen contains less than 15% of the energy per liter as diesel, and only a little over twice as much energy per liter as a lithium-ion battery.
It’s also true that hydrogen is super lightweight, with much better gravimetric energy density than other energy carriers. This is certainly a virtue for any kind of transportation fuel. But in many such applications, volume is as much if not more of a constraint than weight.
Second, hydrogen calls for more specialized, expensive handling than any of the fuels we use today. Exposure to hydrogen embrittles the most commonly used metals for all the tanks, pipes, and other equipment needed to store & transport liquids or gases. Steel, which is the world’s go-to structural metal, is especially vulnerable to hydrogen embrittlement. Moreover, hydrogen has a way of escaping through the tiniest of gaps, so it requires components like seals, gaskets, and valves to be extra tight & robust.
So, yes, hydrogen is a lot like a swiss army knife: versatile, but unwieldy.
On the other hand, there are a few leading roles that hydrogen really could play in a lower-carbon economy. For starters, hydrogen’s most important role in the economy is always going to be HYDROGEN, the molecule, as sometimes nothing else will do. Hydrogen is a crucial ingredient in several of the world’s most critical commodities: ammonia, methanol, and petrochemical refining. Producing hydrogen feedstock for those commodity markets is mostly done by cleaving off hydrogen from methane molecules, with the leftover carbon emitted as CO2. This process accounts for roughly 2-3% of all global energy consumption.1 The world is probably always going to need some mix of these commodities, which means that decarbonizing hydrogen feedstock is a big deal. Clean hydrogen producers could spend the next twenty years trying to serve this market, alone.
Additionally, I think that “clean” hydrogen is probably the leading candidate for decarbonizing iron & steel production, which accounts for another 8% of global energy demand. Around the world, the industry has already begun a very gradual transition from coal-fired blast furnaces to natural-gas-fired “direct reduced iron” furnaces. And the leading manufacturers of gas-fired furnaces have already demonstrated variants which can run on blends of natural gas and hydrogen, all the way up to 100% hydrogen. (I’d also suggest taking a look at Hertha Metals, a startup which has developed a distinctive hybrid-electric take on this kind of flexible furnace design, which is also capable of processing iron ore into steel without the need for an additional electric arc furnace.)
So, no, clean hydrogen is not going to take over the world. But the problem for clean hydrogen is not that there isn’t a big enough market. The problem is cost.
Without getting too deep into the weeds: There is nowhere in the world today that producing low-carbon hydrogen is cheaper than producing regular-carbon hydrogen. That’s true whether your low-carbon hydrogen production method is “green” — produced by splitting water molecules with renewable electricity — or “blue” — produced by capturing the CO2 from old fashioned methane splitting — or any other slice of the hydrogen “color wheel” that technologists have dreamed up. Even in best case scenarios for these approaches, I don’t see the cost of lower-carbon hydrogen coming within striking distance of conventional “grey” hydrogen.
There is one way, however, in which we might be able to produce hydrogen with zero carbon emissions, as cheaply as we produce it today.
Just like fossil fuel, pump it out of the ground.
This is the idea of “geologic hydrogen”, and it is a very new idea, indeed. Three years ago, there were just a handful of academics and wild-eyed scientists who were talking about the idea that there could possibly be significant, naturally-occurring deposits of hydrogen underground, just waiting to be extracted. We’re lucky at Energy Impact Partners that one of those wild-eyed scientists was Dr. Michael Webber, our former CTO and a Professor at UT-Austin. As I wrote two years ago…
In a handful of places around the world, hydrogen actually seeps naturally out of the earth and leaves a telltale sign: a dearth of vegetation in big circular patterns which are, adorably, the basis for ancient stories about “fairy rings”. So, we’ve known for a long time that at least small amounts of geologic hydrogen exist; but frankly, we’ve never had to go exploring for it in earnest because fossil fuel is so much more valuable. Why go looking for a “pain in the ass” molecule?
The reason to go looking for such a “pain in the ass” molecule is that it can probably be extracted for costs that are at or below the current cost of grey hydrogen. This is The Prize for geologic hydrogen. At EIP, we’re investors in one of the first companies to begin exploring for this prize: Koloma, which we believe has a significant data and first-mover advantage. But now that the idea has entered the public conversation, they’re not the only ones searching.
The Challenge
The Challenge for geologic hydrogen is much easier to explain than The Prize.
Essentially: we still don’t have a good grasp on how much geologic hydrogen is out there, and in what concentration, or how easy it will be to extract. Estimates of the total global resource potential are all over the map. For example, see the results of a US Geological Survey Assessment from late 2024, which modeled the total global resource between about a billion tons, which is about ten years of supply for the current commodity hydrogen market, and about ten trillion tons, which is enough energy to satisfy the entire world for thousands of years.
Those are some pretty big error bars. The Challenge is narrowing them, and the only way to do that is to dig more holes in the ground. That brings me to The Path, which is even simpler.
The Path
Explore for geologic hydrogen. That’s it. The more we explore, the more we’ll be able to characterize this resource. In an optimistic scenario, companies like Koloma dig more holes in the ground, find a big natural deposit of hydrogen at a high enough concentration to extract it economically (say, less than $ per kilogram), and then use the proceeds from selling that hydrogen to dig more holes. As they go, they’ll get better at identifying where to dig, and they’ll learn how to extract hydrogen at a lower cost. There’s even a possibility that geologic hydrogen can be stimulated, enabling humans to produce even more hydrogen than we find naturally occurring in certain types of geologic formations.
And crucially, because the oil & gas industry already contains essentially all of the technical skills needed for both exploration and production, it’s easy to envision this new industry ramping up very quickly.
So: Dig.
“Crazy cheap” renewables
The Prize
Solar and wind power are already pretty darn cheap. If you’re reading this blog, then I imagine you’re already familiar with the story of how solar & wind costs plummeted over the past two decades. In case not…
(note that the charts below exclude subsidies — i.e. federal tax credits)
Here’s solar in the US:
And here’s wind:
Today, pretty much anywhere in the world where you can find plentiful sun or wind, plenty of open space to build on, and transmission capacity to deliver power to population centers, solar and wind energy are now fairly attractive resources. Yet even in the very best locations in the United States, they’re not quite competitive with the marginal cost of burning fossil fuel to generate electricity, today, without relying on federal tax credits. And unfortunately, due to a confluence of factors I wrote about in “Adolescence for renewables”, the cost of solar & wind power in the US has been trending up for the past five years.
The absolute cheapest renewable energy in the world, today, comes from massive solar projects in the Middle East — specifically, Saudi Arabia. The Kingdom recently inked twelve gigawatts of solar power purchase agreements at prices that would undercut fossil fuel power generation anywhere in the world.
Hence, it’s fair to say that solar power is already the cheapest source of electricity in the world… in a region with the best solar resources on the planet, practically endless open space, low-cost labor largely imported from developing countries, and a strong (ahem…) central government which has made solar a priority. The big prize for solar is achieving Saudi prices in less perfectly ideal circumstances. Think: Arizona, Australia, Portugal, etc.
Achieving this goal would indeed require “crazy cheap” renewables — at least 50% cheaper than we have today. So let’s use that as our benchmark.
What about storage? (I hear you asking.) Yes, solar and wind are inherently intermittent, weather-dependent energy resources. But I am extremely confident that storage technology is rising to this challenge. We can address at least half of the intermittency problem with lithium-ion battery technology that is already proven in the field today. Fundamentally cheaper, safer sodium-ion battery technology is just around the riverbend. And a combination of Form Energy plus a dash of distributed natural gas generation can take care of the rest.
Sidenote: Form Energy’s multi-day, iron-air battery technology is especially valuable in scenarios with cheap, abundant renewable energy. In a study earlier this year, Form collaborated with the economic consulting firm Charles River Associates to identify the optimal low-carbon resource mix for a Northwestern US utility. They found that the ideal portfolio is about 50% lower cost when Form batteries are included in the mix.
But in a scenario in which we achieve crazy cheap solar, there’s an even bigger prize than displacing combustion at existing power plants. In the absolute best places to deploy solar (like Saudi), we could see solar becoming competitive with the embodied “thermal energy content” of natural gas or coal. This benchmark is especially important for decarbonizing industrial energy demand, which accounts for more than a third of all global energy consumption. Because most of the energy in industry is used to generate high temperature heat, rather than electric power or motion, we really do need to compete with the cost of setting a lump of coal on fire. (In these settings, I believe we can very easily manage the intermittency of solar generation with thermal energy storage — see, for example, the Rondo Energy “heat battery” — at an ultra low cost.)
If you take the long view — the civilizational view, as I’m trying to do in this series — then I believe the cost of energy embodied in a lump of coal is the ultimate target for renewables, or any other low-carbon primary energy resource. I believe that solar, in particular, is the only option we have with a fighting chance of accomplishing this mission.
The Challenge:
First of all, a brief substantial aside: Why not wind?
My short answer is entirely empirical. I simply haven’t encountered any ideas with a realistic hope of reducing the cost of wind energy by 50%.
Modern horizontal axis, three bladed wind turbines are now a mature technology. The foundational design has been pretty well optimized over the past 25 years. (Actually, the concept goes back much longer than that.) For the past decade, the wind power industry’s primary strategy to reduce the cost of energy has been to increase the size and power of each individual turbine. But that strategy faces clear, practical limitations. (For example, there are now specialized aircraft being developed to transport the largest turbine blades, because they can no longer be transported by trucks on regular roadways.)
Hence, achieving much lower cost wind energy production probably requires an entirely new technology paradigm. There is, in fact, one interesting theoretical paradigm which has been explored in fits and starts for the past twenty years, which is generally referred to as “high altitude” or “airborne” wind energy.
The basic idea of airborne wind energy calls for positioning some sort of aerodynamic device at a much higher altitude than conventional tower-mounted wind turbines could ever reach. Energy can theoretically be generated by turbine blades held aloft in various ways, then transmitted back to the ground via a combined tether & electrical wire. Or, energy can be generated by the force that the wind applies to a tether, which cranks a ground-based generator.
The theoretical benefits of this strategy are fairly obvious: high altitude wind currents are much stronger and more consistent than they are closer to the ground. If we could harness these currents, we could effectively tap into an entirely new class of renewable energy resource. But unfortunately, nobody has figured out how to do this practically, and cost-effectively.
That’s not for lack of trying. Around twenty years ago, there was a small crop of companies with a range of technical visions for airborne wind power. The most prominent of these was “Makani Power”, which was acquired by Google in 2013 as part of the company’s “moonshot” program. Makani spent a total of thirteen years iterating on multiple designs, before ultimately winding down operations in 2020. Generously, Google decided to share practically all of the concepts and data they gathered over this period with the world, which has left us with a unique window into the daunting technical challenges in this domain. So far, these challenges have scared off most entrepreneurs and investors from returning to the space, although there is another small crop of scrappy companies giving it a shot — e.g. Kitepower, Skysails, and Kitemill (which was actually started around the same time as Makani).
I’m glad to know that there are still some bold adventurers trying to make progress in airborne wind, but the odds of a breakthrough seem to be low. Meanwhile, there are a few other potential paradigm shifts out there to keep an eye on, but not nearly as much activity as there is in solar.
The Challenge, for solar:
At a basic level, there are three variables that determine the cost of solar energy: A) The cost to manufacture solar panels; B) The efficiency with which those panels convert solar energy into electrical energy; and C) The cost to deploy those panels in the field. These are the three big levers we have to work with, in our quest to bring down the cost of solar energy by 50% or more.
Lever A — the cost of manufacturing solar panels — has already been pulled about as far as it can. If you’re reading Steel For Fuel, I’m sure you’re already familiar with this story: Solar manufacturing made a tremendous leap forward in just a few short years around 2010, propelled by the combination of supportive policies in Europe, and California, and the Chinese government’s commitment to supply chain scale.
Remarkably, global prices have continued to fall ever since, even during the inflationary period we’ve been living through for the past five years. However, this trend is pretty clearly unsustainable, because it depends on Chinese manufacturers operating at a loss. There’s a good chance we see solar prices reverse course a bit as China revises its industrial policy.
In the US, there are also (entirely justifiable) solar tariffs to contend with.
This begs the question: How on earth could solar become cheaper? If anything, I’m convinced that solar module prices need to increase, for the sake of economic sustainability and geopolitical comity. Especially if solar is going to be made in America, or in any other Western nation, we’re going to need to see a tremendous amount of innovation simply to approach the price levels that China and other Southeast Asian countries have already achieved. (See, for example: Leap Photovoltaics.)
Moreover, because solar panels have become so dang cheap, they now constitute a diminished fraction of the total installed cost of complete solar power projects. Before 2015, solar panels accounted for nearly half the cost of a “utility scale” project; now they’re not even a third of the budget. This change has been even more pronounced for rooftop solar projects, which require a lot more sales, marketing, financing, and installation labor for every kilowatt deployed.
Lever B — solar panel efficiency — has much greater potential. In large part, that’s because efficiency improvements ripple throughout every aspect of a solar project. Hence, Lever B is directly connected to Lever C: the cost of deploying solar in the field.
Holding all else equal, increasing photovoltaic efficiency means that you can produce more kilowatts from the same sized panel. This means that you need less of everything else throughout the supply chain for a large solar project: space to transport and store panels in ships, trucks, and warehouses; land to build on; steel racks to support all those solar panels in the field; and labor to install them.
On the other hand, the industry has been hard at work on photovoltaic efficiency for decades, and it’s been a grind.
Don’t get me wrong: 30% improvement in less than fifteen years is pretty good. But in order to cut the cost of solar by half again, roughly speaking, we’d need to double efficiency from where it stands today.
This presents a fairly obvious problem: The maximum theoretical efficiency of a single junction polysilicon solar panel — the kind we know how to manufacture cheaply — is about 33.7%. That means it is physically impossible to double the efficiency of today’s solar cells, without adding more “junctions”, which means more layers of photovoltaic material which absorb different frequencies of light, which pretty much inherently adds more cost. (I’ll return to this point in a moment.)
The Path, Part 1
Solar researchers have a possible answer to this efficiency conundrum. It’s a word you’ll encounter before too long if you spend significant amounts of time in solar tech circles: “perovskites”.
Perovskites are a broad class of materials with a particular crystalline structure, and a myriad of molecular compositions. Much like silicon, perovskites can in some cases be employed as semiconductors to capture solar energy. But unlike silicon, perovskite based photovoltaics can be flexible, rather than rigid. Some variants can be applied onto solar panels as an ultra-thin, lightweight film — a technique which the leading American solar manufacturer First Solar pioneered for Cadmium-Telluride photovoltaics. Because thin film solar manufacturing can be made into an extremely high throughput process, First Solar’s modules are currently the only “alternative” solar technology still afloat in a sea of Chinese polysilicon.
Perovskites have one additional feature which has made them especially intriguing for the solar industry: Their “bandgaps”, which determine the frequency of solar energy they absorb, can be tuned by tweaking their particular chemical formulas. Some perovskite bandgaps are very complementary to silicon — meaning they capture the wavelengths that silicon doesn’t, and vice versa.
Hence, one possible pathway to much more efficient solar panels is to layer cheap perovskite films on top of either silicon cells, or other perovskites, or both — in order to capture a much greater share of the solar spectrum. One good example of a company pursuing this strategy is Tandem PV, which is publicly targeting somewhere in the range of 30-40% aggregate panel efficiency.
Of course, perovskites also have a number of weaknesses to contend with, the most important of which is rapid degradation. Historically, perovskites have not stood up very well to… umm… sunlight, unfortunately. Or moisture. Or other common environmental conditions you would typically find in a big, open field.
Fortunately, researchers are making good progress on making perovskites more robust. My EIP colleague Melissa Ball, a materials scientist who did her doctoral work in solar materials, has been following this space for many years, and she is convinced that we’ve reached an inflection point for perovskite longevity. Melissa also works very closely with a perovskite company which we’ve made a bet on at EIP, but we have not yet announced publicly. (Hence, unfortunately, I need to wait for a future post to share more. For now, here’s a quick preview: Imagine you had a much lighter-weight, more flexible perovskite panel — actually, “panel” isn’t quite the right word — which was significantly cheaper to manufacture, AND which didn’t require so much rigid steel and aluminum to hold up to the sun. Hmm…)
The Path, Part 2
Aside from photovoltaic efficiency, there are plenty of other levers for reducing the cost of deploying solar in the field. Most of these are strategies for cutting the three major inputs to any solar project besides photovoltaic panels and power electronics: land, labor, and structural materials (i.e. Lever C). In my view, these are the best places to hunt for a possible paradigm shift.
My simple rationale is that the combined cost of land, labor, and materials has swelled to nearly half the total cost of a typical solar project. (“Why do you rob banks?” they asked Willie Sutton. “Because that’s where the money is”.)
Installing a solar project typically entails a number of highly repetitive, physically demanding tasks, performed over thousands of acres, out in the hot sun, often far from population centers. Hence, labor costs in particular have become a growing cost driver, which means that construction automation is emerging as one of the most promising strategies. For example, Charge Robotics, one of our portfolio companies at EIP, has developed a process which can automate about a third of the labor required for most solar projects. And Built Robotics, which automates another big share of solar construction labor, is also gaining traction.
I’m confident that these kinds of solutions can make a big difference for the cost of solar. However, I must admit, I’m not so sure that they can quite sum up to the 50% cut that I’m looking for. In order to achieve that level of savings, I expect we’re also going to need a boost from higher photovoltaic efficiency.
To sum up: I think the path to another 50% cost reduction for solar is going to require several overlapping strategies. Perovskites could well be a part of the solution, but perovskites alone are not going to lead us to the promised land. We need to consider what we can do with perovskites that will have the biggest impact on the cost of solar, at the full system level. And robots are always a good bet for cutting the cost of repetitive, labor-intensive tasks.
Geothermal anywhere
The Prize
Geothermal is perennially “the other” renewable energy resource — a junior partner to solar, wind, and hydropower.
Actually, “junior partner” is vastly overselling geothermal’s role in the energy system. Out of a total of about 9,000 gigawatts of power generation capacity installed globally — including about 3,000 gigawatts of solar & wind power, combined — just 16 gigawatts are geothermal.
So, the first question to ask is: How could geothermal become a resource that matters… at all…? Even though, theoretically, we have the makings of a fusion-level resource.
Geothermal energy originates deep within the earth, resulting from pressure and radioactive decay, combined with heat dating back to the origin of the planet. This heat rises gradually through the earth’s crust, towards our civilization up here on the surface. The resource is effectively limitless, power-dense, and available 24/7. It can be used as a substitute for fossil fuel combustion across power generation, industrial heat, and even district heat applications.
In other words: If we could find a way to tap into geothermal anywhere in the world, we would secure a resource akin to fusion energy, but with zero fuel supply chain risks, and zero radioactivity. (As I discussed in Part 1, most fusion approaches need to contend with both of those complications.)
In short: The promise of “Geothermal anywhere” is effectively fusion… but better.
Unfortunately, it turns out that unlocking geothermal energy anywhere is probably just as difficult as fusion. There’s a good chance it’s even harder.
The Challenge
Capturing geothermal energy requires three critical elements to come together, just so. Naturally, this tends to be an uncommon occurrence, because none of these three variables are equally distributed throughout the earth’s crust.
High grade heat: Geothermal heat rises towards the surface everywhere in the world, but it doesn’t actually reach the surface at the same volume, or temperature. The flow of heat depends on the specific geology at each location on the planet. In some places, heat rises naturally practically all the way to the surface at sufficiently high temperatures to generate power — these tend to be the kinds of places associated with volcanic activity, like Iceland. In other places, you’d need to drill down several miles before it becomes hot enough to fry an egg.
Source: SMU Geothermal Laboratory Temperature Maps, 2011. Permeability: In order to extract heat from deep within the earth, you need to capture it in some sort of working fluid — water, for example — which can transport that heat up to the surface. (Because that’s where your power generation or industrial process equipment resides.) This means that you need to find enough surface area, deep underground, in order to efficiently exchange heat between the earth and your working fluid. But just like high temperature heat, the earth is not equally permeable all over the place. Some geology is porous enough for hot water to flow freely, and pick up heat over a large area, while some is not at all.
Water: In some places, you can find enough water underground to use as a working fluid. But in some places, you’ll need to bring your own water, or some other fluid. This is probably the least critical of the three variables, since water is not so hard to find closer to the surface, and we already consume a lot of water in other forms of thermal power generation (which we’ll presumably be substituting with geothermal).
All of the geothermal power plants in the world today have been built where these three variables have come together serendipitously, entirely of their own accord. In fact, many of these prime locations were discovered by accident, as well drillers looking for drinking water ended up discovering hot water or steam by mistake. Hence, the challenge for geothermal technology developers is to find a way to bring these three variables together deliberately across a much wider geologic range.
If you truly want to unlock geothermal anywhere, then the first variable is the linchpin: we need to be able to confidently, reliably, and affordably drill much deeper than we can today.
One of the reasons that this is so daunting is that we humans are already very good at drilling holes in the ground. In the United States alone, the oil & gas industry has drilled more than 4 million wells since its inception in the late 1800’s. Each year, around the world, we drill tens of thousands of new boreholes in order to produce more fossil fuel.
From a civilizational standpoint, we’ve recently made especially big strides in drilling here in the US, thanks to the “fracking revolution” of the past twenty years. The fracking boom produced a bonanza of drilling innovation oriented around hydraulic fracturing (the revolution’s namesake) and directional drilling. New technology has also enabled the industry to seek out oil & gas resources at lower depths. This has led the average well depth to more than double from a little over a mile, to about two and a half miles, since fracking kicked off in earnest around 2010.
But unfortunately, that’s not deep enough to support affordable geothermal energy beyond a very limited geographic range, as the maps above make clear.
Moreover, the relationship between drilling depth and drilling cost is not a linear equation. Drilling costs tend to increase exponentially as you head deeper into the earth. As a result, the US Department of Energy has estimated that drilling just three miles below the surface would yield an average geothermal power cost of roughly $100 per megawatt-hour; while reaching six miles down would increase that cost to over $300 per megawatt-hour. (Recall that solar power in Saudi Arabia already costs about $10 per megawatt-hour.) And these deep geothermal estimates should be considered pretty rough, because we simply don’t drill many holes more than a few miles deep. In fact, the deepest hole humans have ever drilled — the Kola “Superdeep Borehole” in Russia — reached just 7.6 miles below the surface.
There are multiple hurdles to drilling a super deep borehole. For starters, it’s wicked hot down there. That’s the point, of course, but it raises all kinds of problems for the drilling system. We’re not talking about a fusion-level materials challenge, but it’s no easy feat to prevent drill bits from rapidly melting or disintegrating. Additionally, deep holes mean high pressure, so the hole is constantly at risk of collapsing in on itself. Even steel casing can be crushed under thousands of pounds per square inch. Then there’s the question of what to do with all the rock you’re grinding up. That stuff needs to be continuously extracted from the well, even once it’s miles from the surface.
And all of this needs to be controlled with whatever sensory feedback you can glean from the tip of a drill at the bottom of a boiling hot hole in the ground.
There’s no question that we can do this. Unlike fusion, drilling five to six miles below the surface is entirely possible, and it’s not too much of a stretch to imagine we could reliably reach seven or eight miles down. It’s just prohibitively expensive. And in a way, that makes the challenge of achieving affordable deep drilling even more daunting than achieving fusion, period — because we have a lot more understanding of what we’re up against in the realm of super deep boreholes.
The Path
Given that deep geothermal is so clearly on par with fusion as a civilizational energy resource, it’s been surprising how few startup companies there are working on solutions with true breakthrough potential in this domain.
Just a handful of companies are targeting sufficient depths to reach “supercritical” temperature and pressure conditions anywhere on earth. Quaise is pioneering millimeter wave drilling. GA Drilling is was (?) working on plasma drilling. (The company has also developed a novel approach to conventional drilling, which is now the exclusive focus of its website.) GeoX is adapting drilling methods from offshore oil wells — typically the deepest in the oil & gas industry — with a range of innovations in drilling and well design.
I wish these companies the best, and frankly I wish that even more effort was being spent on these kinds of geothermal moonshots. (“Earth’s-core-shots” ?) In my view, society ought to be spending at least as much on super deep geothermal bets as we are on fusion bets.
However, I don’t believe that this is the most likely path forward for geothermal energy.
The most likely path forward is probably going to be a lot more incremental, and frankly a lot humbler. I believe this path has a good chance of making geothermal finally matter as a global energy resource, but is unlikely to make geothermal a fusion-level bet.
This is the path of “Enhanced Geothermal Systems”, or EGS. But frankly, writing about this path as though it refers to a big category of startup companies is disingenuous, because EGS is effectively a category of one. Fervo Energy is not just the flagship company in this space. It’s the standard bearer, the cockswain, the gunner, and the scullery maid, all in one.
The basic idea of EGS, which Fervo has almost single-handedly piloted from concept to commercial reality, is to “frack” a geothermal well into being. In many parts of the world, including the “Mountain West” here in America, there is enough heat just a few miles below ground for geothermal power, but developers have historically required a lot of luck to encounter enough natural permeability. To solve this problem, Fervo has adapted proven techniques from the oil & gas industry in order to drill more affordably and reliably into hot rock, and then crucially, frack that rock into a permeable reservoir. (Another interesting geothermal startup, Zanskar, is taking a potentially complementary approach — using a treasure trove of data and AI in order to identify the ideal sites where heat is closest to the surface, and some natural permeability probably exists.)
So far, Fervo has demonstrated that EGS can work at a 3.5 megawatt site in Nevada. Now the company is focused on its first full-scale project: a 100 megawatt site in Utah. I think it’s fair to say that this project, “Cape Station”, is not just a crucial milestone for Fervo; it has become a vessel for the hopes & dreams of practically everyone working in geothermal… and most of the extended “Climate Tech” community to boot. So, while there are certainly other geothermal startups with innovative approaches that are worth following — e.g. Sage Geosystems and Eavor — all eyes are on Fervo, and rightly so.
I’m already beginning to see a wave of “fast followers” waiting in the wings, with plans to leverage the EGS knowledge which is now beginning to diffuse outward from Fervo, through its oil & gas industry partners.
That brings me back to The Path for geothermal. Here’s how I envision the most likely scenario:
If Cape Station is successful and the next phase of EGS goes well, I expect we’ll see investment begin pouring into the sector. Given the existing scale and expertise of the oil & gas industry, we could fairly quickly see sufficient resources mustered to make rapid progress down the EGS cost curve. From there, incremental progress on deeper drilling techniques and greater fracking precision could gradually expand geothermal’s range of geographic viability. In America, we could see geothermal begin to head Eastward towards the next most fertile ground (around Texarkana and Appalachia).
In this scenario, we probably wouldn’t end up with “geothermal anywhere”, but I can imagine ending up, several decades hence, with hundreds of times more geothermal energy than we have today.
And here’s Part III in this series: Hydrocarbon-like energy carriers, without carbon emissions.
If you like what you’ve been reading, please hit that SUBSCRIBE button!
Global final energy consumption, not primary energy.
Even with geologic hydrogen being cost-competitive, how do you envision the transportation of hydrogen to it's end use? Isn't that part of the challenge as well given the physics you mentioned?