The road to commercial nuclear fusion

View Full Transcript

Eklavya Gupte:

Hello and welcome to Energy Evolution, the S&P Global Commodity Insights podcast dedicated to covering the fascinating and ever-changing world of the energy transition.

This week, we're checking into nuclear fusion, which is the technology that can transform how we power our world. This episode will examine how big tech companies are racing to secure their stake in what many see as the holy grail of clean energy. And I'm lucky to be joined by our correspondent, Camilla Naschert. Welcome Camilla.

Camilla Naschert:

Hi, Eklavya.

Eklavya Gupte:

So nuclear fusion is basically the process where two light atomic nuclei come together to form a heavier one, and in the process they release a ton of energy. Now, research on fusion kicked off in the early 20th century with scientists like Hans Bethe doing a lot of the foundational work back in the 1930s, figuring out how stars produce energy through this process. And the push for harnessing fusion for energy really took off in the 1950s, especially during the Cold War. And countries were eager to find new energy sources.

And the first big breakthrough came in 1952 with hydrogen bombs. But it wasn't until the invention of devices like tokamaks and inertial confinement systems that serious research began. But here's what's exciting. We have hit some major milestones recently. In December, 2022, the National Ignition Facility in the US achieved something scientists had been working towards for decades. They got more energy out of a fusion reaction than they put in, and that's what we call ignition. Yet despite this breakthrough, we're still facing significant hurdles before fusion becomes a commercial reality. Camilla, what are you seeing in terms of the current landscape?

Camilla Naschert:

Yeah, that's right. Aside from the technical challenges, there are material limitations as well when it comes to fusion energy. The materials used in fusion reactors must be able to withstand extreme heat and radiation over a long period of time, and critics also point to high costs for the research required and argue that this money could be better spent on renewables. And as with other forms of nuclear energy, there are safety concerns as well, which play into public perception and support for the technology. So in a nutshell, fusion research has come a long way since it started with some exciting developments as you said, but there is a long journey ahead before it becomes a reality for powering our homes and our businesses.

The enthusiasm for the technology is growing though, and we are seeing increased willingness from investors to back nuclear fusion startups, for example. And as we'll hear in this episode, the tech industry is actually among the front-runners in this game. That's given their expected power demand growth from AI data center computing power. Now to hear more about all of this, let's bring in our first guest.

Francesco Sciortino is the co-founder and CEO of Proxima Fusion, a German company working on fusion power plants. Francesco, welcome to Energy Evolution. Maybe you can start out with introducing yourself. I'd be curious to hear a little bit more about your background and how you ended up founding Proxima Fusion.

Francesco Sciortino:

Yeah, happy to. Thanks for having me. I am a physicist by background. I am originally from Italy, but I lived in the UK for five years. Then one year in Switzerland, five years in the US, did my PhD at MIT on Tokamaks. Maybe we'll talk more about what those are. Basically the most conventional way of doing magnetic confinement fusion. Then I joined the Max Planck Society, which is an organization of national labs in Germany, still on the tokamaks side. And then in 2022 I got more and more interested in stellarators, another kind of magnetic confinement fusion device. And early 2023 we spun out from the Max Planck Society and created Proxima Fusion. And today we are about 90 people between Munich, Zurich and Oxford, all across Europe basically. And we are working on stellarators for fusion power plants.

Camilla Naschert:

Great. I can imagine that for many listeners, stellarators is not a household term. So let's dive into the technology. Can you explain what stellarators are, how they differ from other ways of producing nuclear fusion energy and how you at Proxima are working on this?

Francesco Sciortino:

Yeah, so effectively we are trying to build a star in a laboratory and we're trying to do that by taking this hot ionized matter that is called the plasma by confining it into a magnetic bottle. And that magnetic bottle for us looks like a donut, and that donut has magnetic field lines which hold this plasma, this hot ionized matter away from material surfaces. There is no material surface that can hold the temperatures of this plasma. We're looking at a hundred, 150 million degrees, so more than 10 times the center of the sun. And so the magnetic fields have the function of creating that confinement, that bottle effect.

Now the difference between a tokamak and a stellarator is relatively subtle. I would say that a stellarator is more intuitive in the sense that this donut shape, we don't just want magnetic field lines to go in the long direction of the donut. We actually want them to twist around to create a helical pattern, let's say. That gives us the best way of confining the plasma. So a stellarator is more intuitive in the sense that you want magnetic field lines to go around in a helical pattern so you can try to make magnets to create this magnetic field that also twists around to create something like a helical pattern. In practice, you can make those magnets not really go around in a snake-like form all around the doughnut. You can break them into modular magnets, and that's the kind of stellarator that we are looking at as Proxima Fusion.

A tokamak, the more conventional approach that I used to research, is similar but has at the center of the doughnut, maybe it's better to think of it as an apple rather than a doughnut, in the apple, the core of the apple is what is called the transformer, a component, a piece of powerful electronics that is creating a large current in the plasma. So it's a extra piece of technology that is adding a large current to basically simplify the overall design and that overall design becomes easier when the magnets become flat.

So a tokamak, when you compare a picture of a tokamak and a stellarator, the first thing you notice is a tokamak has all the magnets flat. It looks simpler, it's simpler to design, simpler to manufacture, but the challenge with a tokamak is what you don't see and it's really hard to control In practice. It has instabilities that we have been trying to control and to partly live with for many, many years, many decades. In a stellarator, the challenge instead is in making the magnets that look twisty and that twistiness is a challenge in both design and that eventually in manufacturing. And Proxima was founded on the observation that the modeling and the scientific basis for stellarators is now advanced to the point where we know that we can design these devices.

We also have a fantastic prototype of what a stellarator power plant could look like. That is represented by a stellarator in the north of Germany called W7-X. That's the most advanced stellarator in the world. It's a facility that costed the German government and the European Union about $1.5 billion. And that's the facility that Proxima effectively span out from, and we think that scales much better to a power plant.

So both tokamak accelerators are magnetic confinement devices. At the end of the day, they're donuts with magnets. There are many other kinds of fusion devices with lasers, with pistons, with supercapacitors. Each of these has their own parts and you won't be surprised to hear that there are many opinions in the field on what works the best. Right now there are 60 fusion companies, give or take. Stellarators are fairly well represented in that category, and Proxima is now the best funded stellarator company in the world.

Camilla Naschert:

And you just mentioned already that there's still a host of challenges that you and the wider industry are grappling with, but what would you say would be the next milestones that you are working on and your team are working on?

Francesco Sciortino:

It's related to the magnets. So the intensity of the magnetic field that you produce in these donut shaped devices is very strongly correlated. The relation between magnetic field intensity and fusion power that you can produce is very, very strong. If you don't mind me using a formula on the podcast, the fusion power scales with the magnetic field to the power of four. That means if you can double the magnetic field intensity, you get two to the four equals 16 times more power out. We are talking about quadrupling the field intensity at the center of the plasma. That means we are talking about new magnet technology that we are developing that can increase the fusion power by more than two orders of magnitude. The key milestone ahead for us is in mid-2027 to show a full scale, full power demonstration of one such magnet of the right kind of shape. We call this the stellarator model coil. And once we demonstrate that magnet at full power, we think we've knocked out the most important technology milestone that then leaves a clear way, a clear path to making a net energy stellarator.

Camilla Naschert:

And where does that path end? What kind of timeline can we imagine for this to move out of a pilot to real power plant?

Francesco Sciortino:

So the magnet, the stellarator model coil as we call it, is in mid-2027. Then we have four years of assembly of a device that we call Alpha. That's a net energy producing device. In other words, it's a demonstration. It's not quite a power plant. In other words, you give me 50 megawatts, I give you back 50 megawatts. Your 50 megawatts might be from coal, gas or whatever. I give you back 50 megawatts from fusion. That's an important milestone. It's the demonstration that you are at a scale, at a capacity that is very interesting, especially if your device is able to naturally go into steady state. So continuous operation and if it's intrinsically stable. And that's the key focus of Proxima, the stellarator opposed to a tokamak, especially a category of stellarators called a QI stellarator is a device that is naturally able to go into continuous operation and that is always stable.

So if we knock that out in the roadmap, we want to have first operation of the Alpha Stellarator in 2031 and then full power net energy demonstration as soon as possible, hopefully 2032. And then we will proceed to a first-of-a-kind fusion power plant. Right now we say as a goal for the company to do a first-of-a-kind power plant within the 2030s, and that means as soon as we can, as close to 2035 as possible. But generally within the second half of the 2030s. If we get that as expected, then the 2040s should be the time when we scale up as fast as we can. We hope to see a great exponential, the likes of which we've seen, for example in France when there was a deployment of fission power plants in the '70s, '80s. We know as humanity how to deploy hardware very large scale for energy systems. We need to have the right kind of motivation, the right kind of drive, let's say, to make a new energy transformation, the likes of which fusion can bring about.

Camilla Naschert:

And clearly you've been able to convince some investors of this roadmap as well. Proxima raised 130 million euros of series A funding quite recently. Can you explain why now was the time that you were able to raise this funding and how does that fit into the funding requirements for the longer roadmap?

Francesco Sciortino:

Right. So the total raised amount for the company, including this latest series A is about $200 million between public and private funding. What I think investors both on a European and US level have appreciated is the speed of quality of execution. We're a team that is based all across Europe. Many of us have actually spent quite a bit of our lives in the United States. We have seen the opportunity to translating best engineering quality from German national labs, bringing together a community from across Europe and also the US into a engineering project that now we have shown can actually scale much faster than most people expect. So over the first two years we started out saying we can produce a reactor design over 24 months. In practice, we took only 12 months. We then decided to publish that and that design is called the Stellaris design and you can find it online. It's a published peer review paper that we decided to share with the world.

We then got into R&D work in the lab, especially focused on superconducting magnets. Now it's been some time. We are transforming again as a company into a manufacturing company. And this rapid transformation is what venture capital really wants to see. Not an ambition of knocking out one milestone after the other one in a very predictable way, but rather surprising everyone in terms of overcoming previous barriers. So for us, the why now moment was really 2022, something got unlocked on the design side. We showed in 2022 as a community of research in stellarator that it was now possible to optimize numerically stellarators to come to new trade-offs between the physics and engineering constraints. And that really changed the whole game.

Also, we started as a company from W7-X and the ability to transfer the know-how from the world's leading Stellarator in the world, W7-X is really incredible in what it has achieved. The ability to translate that into a commercial path that we have shown over the last couple of years to be transformational. The investors fundamentally are not in the game to support a moonshot project for the sake of humanity, but really to make money. So part of the enabling element of the story is that we are creating intellectual property that will be at the basis of an entire new industry. Fusion is not something to tiptoe into. It's understood by our investors that this could be entirely transformational. The market size is enormous. Maybe energy is the single largest market that anyone can actually try to tap into. We're looking at $10 trillion more or less. The question is how do you access it? And it's right now in fusion a question of feasibility rather than how cheap is it?

Eventually we have to make a device that is commercially attractive. So it does boil down to levelized cost of electricity. We need to reach less than a hundred dollars per megawatt hour, and we need to do that with a clear path to decreasing costs over multiple generations of power plants. So iteration speed is very important and I think over the first two years of Proxima's life, the investors have seen that quality and iteration speed is the one thing we're selling.

Camilla Naschert:

Now, if we look at funding rounds more globally, we've noticed a trend of tech companies putting quite a lot of money into funding fusion ventures recently, partly driven by this concept that we've talked about on the show before of exponentially growing data center power demand or projections of that at least. Is that where you see a lot of the potential, so there is additional growth in power demand or which use case settings do you envisage for your stellarators in the future?

Francesco Sciortino:

Yeah, there is a lot of demand and a lot of excitement around the market represented by data centers. I would say that the data centers are really at the forefront of people's imagination because AI is enabling so much of new applications in society. But actually we didn't need the data centers to realize that human civilization scales with energy effectively. So the access to energy is directly proportional pretty much to our ability to manufacture, to give quality of life. There is still much of the world population that does not have access to energy, and here we're talking about clean energy, safe energy, and that has a completely different kind of appeal compared to gas, coal and whatnot. So data centers are definitely raising the attention.

Maybe worth mentioning here, the power purchase agreement that Commonwealth Fusion Systems, a fusion company based in Boston, as announced just a couple of days ago with Google, that is a major milestone for the whole field. The Commonwealth Fusion Systems colleagues are doing great at building a tokamak. So the kind of device that I mentioned earlier I used to work at, and regardless of whether one thinks that the tokamak is the best path forward or not, it's incredibly good that now we are at the point where the magnetic-confined fusion technology has reached maturity and there is enough momentum to see the first customers effectively.

So Google needs an outrageous amount of energy to power the data centers that directly determine the success of Google itself. So now Google is looking at multiple paths to its energy procurement. In the short term, they have made public that they have been looking at solar and wind and batteries. In the middier term, they're looking at geothermal and small modular reactors using fission. Fusion is in the long term, but it's now time... Right now in the 2020s, we are creating intellectual property that is at the basis of this new energy generation and massive scales. And it's great that Google is recognizing that and not only Google.

Camilla Naschert:

And that kind of brings me to my final point. You mentioned this PPA in Boston, whether there is a global race to be first, a race for fusion and to what extent Europe in a good position there maybe relative to the US, relative to China.

Francesco Sciortino:

Yeah, the race is real. Right now there are these 50, 60 fusion companies all racing to be first. In reality, I believe this is going to be also a quite geopolitical game. It's quite likely that now we have about 10 companies that have raised more than a hundred million, which is enough to go and demonstrate key technology enablers. Let's say for us that's a full scale superconducting, non-planar, so weirdly shaped magnet. There are other companies that are trying to prove themselves through their technology enablers. There is going to be a sub selection of those that actually manage to get to that. Maybe six companies are going to go out and have enough funding, $1 billion scale funding to go and build a net energy demonstrator. And then there will be a few players that go and build first of a kind fusion power plants. These are not likely to be all in China, all in Europe or all in the US. It's quite likely that there will be one or two US champions, one or two European champions and one, two Asian champions.

In the US and Asia, I think in some types of fusion R&D, some champions have been emerging. In Europe now with the latest round, Proxima takes quite a stance, but really the competition in fusion is still quite open. There is still great diversity of thinking about what will work best. So players in the game still need to prove fundamentally themselves eventually. It's not just about doing the energy, it is about delivering energy at a cost that makes a compelling case for the energy markets. It's also a game where the public is going to play a significant role.

For Proxima now in Germany, the alignment with the German government and with various state governments is crucial. I personally believe that deep tech and climate tech, large scale technologies that try to tackle big problems, they require an alignment between the public and the private, which is not entirely common. And right now we have a big problem related to the energy transition is incredibly large problem. I think most people quite really struggle to get their head around the fact that we are not trying to build one gigawatt. We need thousands of gigawatt, we need thousands of power plants. That means we need the public and the private to come together to realize how do we scale up new technology as fast as we can. So in the competition, how governments will play will also make a big difference.

Eklavya Gupte:

Thank you, Camilla. As you mentioned, much of the investment behind fusion companies has come from the big tech sector and data from S&P Global market intelligence 451 research shows that recent global funding exceeds $7.1 billion across 50 startups, many of whom are backed by big tech.

Camilla Naschert:

Yes, that's exactly right. And to get into some of the details on how tech companies are thinking of fusion in their energy procurement strategy, let's hear from Steve Piper, director of energy research here at Commodity Insights. Steve, thank you so much for joining us on the show. Now you and your team have recently written about this growth in funding for fusion ventures that we've seen recently, especially from the tech sector. So I'm curious to hear a little bit more about that research and how you see that trend evolving. Do you want to give us an overview of some notable recent funding rounds and maybe some numbers and some names and how that's changed in the last couple of years?

Steve Piper:

Yes, Camilla, thank you for having me on. Fusion has rather famously or notoriously always been 10 years away as a source of commercial energy, commercial electricity. But it has of course intriguing properties that align with the goal of tech sector investment hyperscale loads that support cloud computing and artificial intelligence infrastructure with zero carbon electricity. And more recently, a number of companies have made sufficient progress and developed pathways of research that hint at potential commercialization of electricity from fusion. There are a lot of private sector, call them startups, firms that have been working on different approaches to fusion for quite a bit of time, but are now independently pursuing this technology.

There are probably in the United States three companies that probably have been the most successful in raising funds for their approaches. They're actually more than three companies trying different approaches. But for an overview of this space, we have Commonwealth Fusion Systems has raised about $2 billion in pursuit of a more conventional tokamak fusion reactor. This is a team that was spun out of MIT in Boston. They made a big splash just early in July with the planned reactor that they have in Chesterfield, Virginia, roughly 400 megawatts is their goal to have that in service by the mid-twenty-thirties. So yes, still 10 years off, but with a greater degree of confidence in ultimate success, Google has signed a power purchase agreement for 200 megawatts of that capacity. So we're seeing these tech companies have enough of an expression of confidence in the progress, what these fusion companies are doing to put ink to paper and at least to outline a power purchase agreement. So Commonwealth Fusion Systems about $2 billion raised, 200 megawatt PPA for their first reactor if they're able to bring it to service.

TAE Technologies has raised about $1.2 billion. Their approach has established a stable plasma at 70 million degrees. The key to any of these approaches is to establish a stable plasma in a strong magnetic field of somewhere between 150 million and 200 million degrees centigrade. For context, they've got a stable plasma at 70 million degrees. They'll try to gradually step that up to a level that can produce commercial quantities of electricity.

Third is Helion. They've raised about a billion dollars. They made a big splash a couple of years ago announcing a 50 megawatt power purchase agreement with Microsoft and a target date of 2028. Now that's just around the corner. So if they were able to meet that target, of course that would be big news. Their next prototype, which aims for a stable plasma at around a hundred million degrees would be the next threshold for them to achieve. They had hoped to achieve something like that by the end of 2024. So they're a bit behind schedule. That might ultimately affect their continued funding, but that's the next step that Helion is aiming at. So a lot of activity in this space. The tech sector is I think, aware that these investments are to some extent moonshots, but it's a lot of interest in pursuing these independent minds of fusion research.

Camilla Naschert:

And I mean your research obviously focused on the tech sector among the investor base behind fusion ventures. Are they the most significant backers and why is that?

Steve Piper:

I think it's fair to say that tech companies investing in sort of hyperscale data centers have become significant and growing backers of a number of different clean energy technologies. They have signed agreements for the next generation of small modular reactors on the nuclear side across a number of potential developers there. They have assigned power purchase agreements with enhanced geothermal systems. There's a growing awareness that the suite of mature emissions-free technologies, think wind power, solar photovoltaic carrying that at times with batteries doesn't deliver 24/7 electricity with those zero emissions characteristics. We're still looking for that kind of ideal power source that can deliver baseload electricity. And in pursuit of that, data centers have placed a number of funding bets across a variety of technologies. Fusion in terms of its contribution, it's a little bit further out and it's at a little bit lower scale, but there is a recognition that if that bet were to pay off, then data centers could commercialize emissions-free, potentially inexpensive and reliable electricity and help manage the implications and consequences of all this load growth that we anticipate coming out of the cloud computing and the AI infrastructure.

Camilla Naschert:

And all of this load growth is fairly real and fairly tangible. I mean, plans are for the next years here and maybe the next decade. And as you say, fusion is a little bit further out maybe in the. 2040s, 2050s. So how are they so motivated to invest in something so early stage given the potential delivery risks of this?

Steve Piper:

It's a really interesting question. I think part one to that answer is that there's a growing realization that fusion could be available somewhat sooner on the one hand. And that the research pathways for fusion are becoming maybe a little bit shorter. Lawrence Livermore National Labs we're able to achieve positive energy on a fusion reaction. The system energy was still negative, but that provided a path forward for additional research because you can look at that whole process chain and say, okay, can we make the lasers that trigger the reaction more efficient? We're dealing with 30 or 40-year-old infrastructure, can we look at that process, end to end and leverage off of the positive energy achieved by the national laboratory? The longest plasma that's been sustained at this point is about 22 minutes, certainly well short of commerciality, but they were seeing fusion occur in that reaction, still at negative net energy, but the research paths are unfolding that sort of dance around that possibility of positive net energy.

So they're seeing a research path forward. The tech sectors are funding independent lines of research, and all of the sources of power that data centers are pursuing are going to take time. The next generation of nuclear plants, they're looking at early-twenty-thirties to see those prototypes come in. Even natural gas generation, the turbine orders are backing up and to put new generation on the ground is a three to five-year timeline. So part of the answer or the second part of the answer is that it's going to take time to put large amounts of generation on the ground regardless of what you're investing in. Fusion, given the shortening timelines and research pathways and the lengthening timelines to add other forms of generation starts to look like it can fit into the mix.

Camilla Naschert:

Something else that I was wondering is what exactly Google or Facebook are buying as in, what are they getting in exchange for their investment? Is it just the knowledge that they're supporting a technology that will become relevant for them in the future? Or are they securing some kind of future output already or some kind of intellectual property that they have access to?

Steve Piper:

Well, I think it's a combination of those things. I mean, first and foremost, access to reliable electricity to power data centers to allow them to scale up artificial intelligence infrastructure, the training and inference, all of these firms are vying to be first to market with the most AI tools they can support and see a need for securing long-term access to the electricity necessary to sustain those operations.

And then secondarily, and maybe not secondarily, I think it's top of mind for all of them that their customers, their users want emissions-free electricity. They want it to be as economical as possible and they want it to be sustainable. And so they've pursued these mandates for net-zero generation from their operations. That's been part of their culture going back 10, 15 years or more. And so these approaches are all meant to bring together reliable baseload power that is also emissions-free.

Eklavya Gupte:

All right, listeners, thank you for tuning into this episode of Energy Evolution. Hope you enjoyed it. I'd like to take a moment to recognize the Energy Evolution podcast team, including my co-host, Taylor Kuykendall, Karen Willenbrecht, Dan Testa, and Camilla Naschert, as well as our agency partner, the 199 and the Commodity Insights digital content team.

And please don't forget to subscribe to Energy Evolution on your preferred platform to stay up to date with our future episodes. And if you have any ideas for podcastings or guests, please send us an email at energyevolution@spglobal.com. Thank you for tuning in and we'll see you next time.