This is the first European fusion company we've spoken to, a spin-out from Max Planck. For someone meeting the company fresh, what is Proxima building and why does it matter right now?
We spun out of Max Planck in early 2023. I came from the world of tokamaks, was at MIT when CFS spun out, and then joined Max Planck, where in 2022 I was one of the European coordinators for tokamak research. Around that time I lost faith that tokamaks offered a path to commercial power, and I was fortunate to be in the one organization that also hosts the most advanced stellarator on the planet. Proxima's vision came largely from seeing what W7-X is like: it was designed to do certain things and hit exactly those targets in 2022, which was spectacular. That same year a theoretical study showed the last big physics drawback of stellarators had been overcome, which told us the QI stellarator path was the right one.
QI stellarators are a type of fusion that has already been scientifically established. These devices are steady state, so they operate continuously, and they are intrinsically stable, meaning they cannot go unstable, which is very different from tokamaks, laser fusion, or magnetized target fusion. The open question is the magnets: you have to build more powerful superconducting magnets, as you heard from CFS. Proxima was founded on the belief that if you get that technology jump right, the concept is robust and scientifically clear. So we can build an engineering company focused on those magnets, on simplification, and on commercial viability, grounded in what Max Planck has done rather than wondering about scientific principles.
The CFS speaker was keen on fusion having a ChatGPT moment. Given you say the science is de-risked and the magnet is around the corner, do you think we are getting close to such a moment?
I agree with your framing. A ChatGPT moment is a moment of shock, an unpredictable one. Demis Hassabis might say there was nothing unpredictable about it, that we knew it was coming, but for the vast majority of people it was a surprise. The same was true of the National Ignition Facility crossing Q greater than one in late 2022: it was exactly on the roadmap and was what the facility was designed to do, yet most people experienced it as a great surprise. I prefer to think of it as a milestone rather than a breakthrough, and the Q greater than one of SPARC, which is not far away, will hopefully be a great milestone for the whole field, not just one lab.
These things build on each other. When MIT and CFS demonstrated their large-scale high-temperature superconducting magnet with the toroidal field model coil in 2021, and scaled up the HTS supply chain, Proxima would not be doing what it does without that proof.
Humanity raises the bar milestone by milestone until it is high enough to fit a stellarator power plant underneath.
I do not expect to be shocked over the next few years, but I am watching very carefully, and I think the world will find Q greater than one inside a commercial entity to be a genuine shock that wakes people up.
Your 2027 model coil is the big milestone. Concretely, what does it have to show for you to call HTS risk retired?
It has to generalize the result of the toroidal field model coil to a non-planar magnet. The reason is geometric: QI stellarators cannot be designed, as far as we know, with flat magnets. A stellarator has no central solenoid, which is the extra electrical circuit that in a tokamak drives a large current and forces pulsed operation. Removing it lets the device run continuously if you design and manufacture it correctly, and the QI property is what makes it intrinsically stable and free of the disruptions that plague tokamaks. To get all of that, you have to build twisty, non-planar magnets, which is a hard challenge.
The company that can design an entire stellarator meeting both the physics and the engineering feasibility metrics, manufacture those magnets, and do it fast enough to suggest power plants can be deployed competitively, is a transformational company. For us, that 2027 magnet is the next level, just as the model coil unblocked CFS to build their tokamak. I do not doubt a tokamak can reach Q of one, and I do not doubt a QI stellarator can exceed Q greater than one, but you have to show me the magnet. I believe it is feasible because I have seen the design, the simulations, and the prototypes, but we are not done building the full-scale, full-power prototype, so that is what we are doing now.
Fusion timelines famously slip. How long do you think this will take?
The magnet milestone is around the corner; we are manufacturing it, so you can count on it. Our Alpha device, the net-energy-gain stellarator, is already being designed in partnership with Max Planck and is reaching its conceptual design review, the line past which you say this is clearly feasible and commit to full detail and funding. In late February the Bavarian government stepped in with 400 million euros of non-dilutive capital, RWE is supporting, Max Planck is fully behind it, and we are now advocating for the German federal government to add significant support, so a real ecosystem is coming together.
Alpha targets the early 2030s. That is a broad statement, not a promise about an exact date, because we are building a reputation for not promising what is not feasible, and we would like to push it as close to 2030 as possible. After that, with the same philosophy, a first-of-a-kind within the 2030s, meaning as close to 2035 as we can, somewhere between 2035 and 2040. We have more detailed internal plans, but there is no value in giving Washington Post readers a precise number that will surely change, hopefully for the better.
Net gain, engineering break-even, and competitive cost of electricity are three different things. How many years and how much capital separate Alpha's net gain from arriving at utilities with a competitive cost?
The first point, maybe obvious, is that a first-of-a-kind of any technology is not a compelling product. Nobody sensible thinks a first-of-a-kind flies off the shelf. We need multiple companies, multiple projects, perhaps different concepts, going out and proving this works connected to the grid, and some scheme has to be found so that someone pays for the R&D that brought each company there. Then you need a second iteration, and a third.
The key objective for every fusion company, in my view, should not be a demonstrator or even a second-of-a-kind, but reaching infrastructure financing, the point where you no longer need equity and loans can finance a project. Then you are in a different phase, and questions arise about whether you are a design-build-operate model or an OEM. That is noise at this stage. We need to see a great deal happen over the next decade, but you are looking at the 2040s for that scale-up. The journey we like to invoke is EDF after the 1970s oil embargo: with a clear government statement, a company that could execute, and a population that wanted it, France decarbonized its grid in about fifteen years, possibly the most incredible hardware deployment humanity has ever done. That is what we have to do in the 2040s, which is why we believe we must be a vertically integrated company, modeled more on aerospace than on EDF, since EDF is not exactly a model of deployment speed.
What stopped fission's growth in France was EU and German policy. Is fusion exposed to such risks once it is developed?
It is not, for clear physical reasons. There is no chain reaction in fusion, so the process cannot run away. I will spare you the big statements about it being super safe and then even safer; none of that is necessary. These are big machines with a lot of energy, so you must be careful, and there is radioactivity, quite a lot of it, but no long-lived radioactivity if you do it right.
That safety profile has been recognized in the UK, the US, Japan, and Germany, where it is becoming law, and hopefully across Europe. The effect is to reset fusion regulation toward radiation protection rather than atomic law, as the Germans put it, which removes much of the policy fragility that affected fission.
You chose to build on former fission ground, in a country that spent twenty years exiting nuclear, now reframed around European strategic autonomy. Are you getting from European and German governments the sense that this is a strategic, autonomy-driven project?
Proxima is not just about doing fusion; it is about being in the right place at the right time, and not only because of an energy crisis, a neighbor invading from the east, a weak European economy, or the fact that we have not built great tech companies in a while. It is the combination, plus the fact that we are not pumping gas out of the ground, that means we really need fusion to succeed, and anyone in the German government will agree and do what it takes. The last time a European leader said "whatever it takes," Mario Draghi, and stuck to it, it built lasting confidence, and that reliability of institutions over extended periods is one of the real values Europe can now offer fusion, in deliberate contrast to other geopolitical contexts.
So when Chancellor Merz says Germany will build the first commercial fusion power plant, full stop, and then repeats it, and his minister repeats it, you conclude they intend to signal something and want to stick to it. As in the EDF story, you need a government that truly wants it, a company that can execute, and a public that feels it is necessary. That said, I would tell any politician who called Proxima the single basket for all their eggs that they were being irresponsible, because no one company changes a continent's energy. Europe's real opportunity this century is to jump into deep tech with both feet, because its manufacturing ecosystem for high-precision mechanical systems is far ahead of the United States, so Proxima must play its role: this is as much about inspiring a continent as about building a power plant.
You have raised around 200 million while Alpha alone is reportedly a roughly two-billion project, and CFS has raised far more. Do you see European companies in a battle for capital at the US level, and how dependent is the model on the state?
Macroscopically, Europe has a capital deployment problem, but in Proxima's case that is no excuse not to succeed. We have raised more than all the stellarator companies worldwide put together. We are the youngest stellarator company, the largest, and the best funded, which says something. Yes, we have not raised as much as CFS, and all I can say there is wait and see.
If we are good at what we do, international investors want to make money, and the Singaporeans, Americans, and Japanese all love making money. If we offer the best team and the clearest roadmap with the lowest scientific risk and the greatest clarity in separating dreams from what is buildable, why would they not invest? Europeans do need to stop putting money only into real estate or Palo Alto funds, and the number of European family offices I have met in Palo Alto rather than at home says a lot, but that is changing. You will not convince people to invest in Europe by guilting them into it; you convince them by showing they can make more money, and Proxima needs to help drive that shift rather than just be subjected to it.
You were at MIT in 2022 and have lived both systems. Is the gap in pace and execution between the US and Europe actually real, and is there something the Americans have that Europe lacks or vice-versa?
They have founders. The rest does not really count. There is a clear statistical difference, what Patrick Collison calls the ABC, arrogance, bureaucracy, complacency, and it affects Europe more than the US. That said, anyone who has lived in the United States knows the average person there is not an MIT fusion engineer, and there is a great deal to be proud of and defend in Europe: public health, public education, public transport. But that does not come from the sky; we have to build great companies to pay for it.
Does Europe need to change its culture? Yes, and culture is the thing we have to change, which can happen relatively quickly, because building great companies does not require majority votes in parliament. Scale-ups and startups becoming larger change culture first in small bubbles, then through their supply chains, then possibly as lighthouse projects. San Francisco is spectacular, Boston's biotech is fantastic, London has a lot, Paris quite a bit, and Munich in deep tech is becoming something. Ten years from now I think people will see that deep-tech investing in San Francisco in 2026 had odd things going on, valuations that did not seem right and products hyper-funded before validation, whereas Munich has a scarcity mindset that pushes founders to do exceptional things, build real teams of people who believe in equity rather than hopping between ever-higher salaries. Different ecosystems, different pros and cons, and if Europe does this right, in ten years the game will have changed.
