Nuclear fusion seems to have been “twenty years away” forever, but recent advances could mean fusion is finally on its way to becoming part of our energy mix. Azeem Azhar speaks to Nick Hawker, co-founder and CEO of First Light Fusion, a UK-based company that uses an approach it calls “projectile fusion” to generate energy.
That method uses an electromagnetic launcher to fire material towards fuel at high speeds. Nick explains why it’s one of the most promising approaches to a problem that has puzzled physicists for decades. He breaks down the challenges, both extraordinary and mundane, that fusion experiments involve.
They also discuss:
- Why First Light Fusion’s system could generate power more efficiently than the sun.
- When the first fusion reactor could plug into an electrical grid.
- How many lasers it takes to turn on a light bulb.
@Azeem
@FLF_Nick
Further resources
Investing in Deep Tech for an Abundant Future (Exponential View Podcast, 2021)
AZEEM AZHAR: Welcome to the Exponential View podcast. I’m your host Azeem Azhar. This week’s guest works on one of the world’s most exciting and fiendishly difficult. Technological problems: nuclear fusion. The promise fusion is well understood; harness the power of the stars to provide endless clean energy. So channeled, fusion power could support our growing energy needs while maintaining and helping us achieve our sustainability goals. Electricity to demand may triple in the next 30 years, an overall energy demand may follow a similar trajectory and all that while we’ll need to decarbonize the energy system. Fusion could be one part of the answer, but it’s also proven to be a really hard problem, seemingly 20 to 30 years away from fruition. Everything about fusion seems like science fiction, depending on the approach chosen a fusion engineer might need to control temperatures above up to hundreds of millions of degrees, accelerate objects from a standing start to 30,000 miles an hour, all in the width of a fingernail. Fuse exotic ions together, maintain liquid metal walls and containment chambers, build magnets hundreds of thousands of times, more powerful than the Earth’s of own magnetic field, or even synchronize hundreds of lasers pumping millions of joules of energy at a point in space in just a few nanoseconds, whichever path you take, and there are a few, it seems rather complicated. Now there does seem to be some kind of engineering progress. In recent years, dozens of scientists have founded start up tackling this opportunity. These teams are racing alongside longstanding public sector projects like America’s National Ignition Facility and the international collaboration [inaudible 00:01:54]. The entrepreneurs seem to be gathering some momentum; between 2010 and 2018, $250 million of private capital. A comparatively small amount, went into fusion startups. In the first 11 months of 2021 alone, more than $2.5 Billion went to fusion companies. So are we on a path to this milestone? What obstacles remain, and what don’t we know? Nicholas Hawker is one man who can throw some light on those questions. He is the founder CEO of First Light Fusion, an Oxford based fusion startup entering its 10th year, and he is at the forefront of addressing this challenge. Nick, welcome to Exponential View.
NICK HAWKER: Thank you very much. My pleasure to be here.
AZEEM AZHAR: What first prompted you to build a fusion reactor?
NICK HAWKER: Well, I’m very motivated by what fusion can mean for society, for climate change and for human development. But if I’m being honest with you and honest with myself, what motivated me at the start was that actually what we are doing is a new approach to fusion, which had never been explored before. And actually what attracted me was being off the edge of the map of human knowledge, and exploring and finding something new. The challenge of the problem, the difficulty of the problem, and the fact that this was an opportunity to do something and learn things known on the planet had ever learned before.
AZEEM AZHAR: When you frame it that way, it’s a sort of a challenge wrapped and a challenge, because of course we haven’t built successful fusion reactor yet, and we have sort of some theoretical approaches, and some experiments. And one could argue that even using the established set of experiments was pretty challenging. You’ve got on even further off that map.
NICK HAWKER: We, all of us stand on top of that research, which has gone before. It took all of that government funded work to get us into the position where we are so close and where the private sector can come in and kind of build on top. I think the thing which fascinates me about fusion is yes, it’s such a hard set of problems, but it’s also, the solution space is so enormous. There are so many different ways of doing it, and so many different roots through all of the difficulties, that actually you have radically different technologies being pursued by different startups, different government organizations. And it’s such a malleable thing, even though it’s an incredibly hard problem. And that’s one of the things that just gives me confidence that it is going to be solved. There are lots of ways to do this, many of which are right at that moment of being on the cusp.
AZEEM AZHAR: So let’s explore from the basics, perhaps what fusion really means and how it operates. I mean, I think we get this idea; it’s a process that happens in a star, but could you perhaps show a little bit more detail on what it actually entails?
NICK HAWKER: So in many senses, fusion is the opposite process to nuclear power. With nuclear, you’re splitting big, heavy elements like uranium and plutonium, and when you split those elements into parts, that releases energy, there’s a slight mass difference between what you start with and what you end with, and the difference comes out as energy. Whereas fusion is the opposite, because it’s joining together rather than splitting apart. And it’s the very lightest elements, the types of hydrogen, rather than very heavy elements. And it’s slightly strange that you split things and get energy and then end of the spectrum, you join things and then you get energy, but that’s how the physics that works. And so, when you fuse hydrogen, you make helium and yeah, you release energy in that process. That’s the process which sun uses, but the sun is actually not a very good fusion reactor. Actually, the heat output per unit volume of the sun is about the same as heat output of the compost heap. It’s just really massive.
AZEEM AZHAR: I just need to make sense of that. So, I have a compost heap in my back garden and in summer it does get kind of quite warm. It’s warmed by the sun and there’s all these biological processes coming out of it. You couldn’t even cook an egg on it, but so just help me understand that. So the sun is out there, this huge thing, and it’s kind of inefficient?
NICK HAWKER: The comparison I’m talking about, just so you know, is per unit volume. So if you had a piece of the sun that was the same size as your compost, he then you’d get the same amount of energy out from it. But obviously, the sun is rather larger than compost heap. The reason for this is to do with the fuel that’s being fused. So the sun fuses normal hydrogen, which is actually astoundingly difficult to fuse. All of the leading approaches to fusion on Earth are using different forms of hydrogen, different isotopes of hydrogen called deuterium tritium. Hydrogen just by itself is a proton and a mono electron, and deuterium, the nucleus is a proton, and a neutron and tritium is a proton and two neutrons.
AZEEM AZHAR: And deuterium, I understand, is sort of naturally occurring on Earth and there’s sort of famous race during the Second World War, to get access to deuterium for early experiments. I remember for some reason there was lots of deuterium in the water in Norway, if I remember correctly, but tritium is not naturally occurring.
NICK HAWKER: So deuterium is present in all water on the planet. So the water of the seas, the water in your body, actually, if you had a pint of beer, there’s enough deuterium and a point of beer to power an average home in the UK for a year is just present everywhere. Tritium though is not stable, it has a half life of 12 years. So any process which naturally produces it, it decays. So you can’t get it from nature. One of the challenges of fusion is to produce your own tritium within the plant. So you consume and tritium and you have to produce your own fuel supply. And how you do that, turns out to be quite involved in a number of ways, which we can talk about.
AZEEM AZHAR: Yeah, we’ll do absolutely do that. What are the actual benefits though, of the fusion process? Does it give us much more energy for the cost or for the volume, or is it just a sort of fungible replacement? What’s our mood, should we build some nuclear fission and should we build some fusion?
NICK HAWKER: I anticipate that it is going to be very different in terms of costs, but the reason for that is because it’s very different in terms of safety, and it’s to do with the nature of the process. So, if you interrupt a fusion power plant, the process stops straight away. You have to put so much effort to get these conditions that you need for fusion with such precision control, anything which disrupts that, it just turns off immediately. So a meltdown is just impossible with fusion. It’s not how the physics works. Also, the only direct product that you have is helium. So there’s none of the kind of spent fuel leftover radioactive waste that you have from nuclear. And then the last thing which does have a big bearing on cost is there’s no weapons grade material in a fusion power plant. You can break into a fusion power plant, steal some special metals and then make something really unpleasant. But all of these things, ultimately, in nuclear translate to cost, because it’s triple security fences around the outside perimeter, and it’s redundancy in every system within the power plant. It’s not good enough to have a backup diesel generator in case the power gets off. You’ve got to have a backup diesel generator for the backup diesel generator, right? Sort of depth of safety engineering won’t be necessary for fusion.
AZEEM AZHAR: So we’ve got this fusing of isotopes of hydrogen, and you’ve got the longer term safety benefits of it. But obviously from that pencil sketch, it becomes quite challenging, because the nuclei are both positively charged, they obviously repel each other, we have to overcome that force in order to get them to fuse. That is the hard part of the physics, right? Or one of the hard parts, right? How do you overcome that force? So how do we think about that?
NICK HAWKER: It’s a bit like if you imagine playing crazy golf and you had like a hole that was on top of the hill, you’re trying to putt the ball into the hole on top of the hill, but you got to hit it hard enough to the hill, otherwise you don’t have a chance to do it, right? So this is where the temperature requirement comes from, which is to get to 100 million degrees kind of temperature, or 10 to 100 million degrees temperature. That’s where you need to be in order to get enough of them to fuse that you get positive energy back up.
AZEEM AZHAR: Yeah. But of course you have to get these ions to smash into each other as well, because if you’ve only got two ions, even if they’re moving very quickly, the likelihood that they’ll collide is quite low. So I guess there’s also a density question. The denser, your ions are, the less time you need for them to have a chance of colliding with each other. And if it’s really not very dense at all, you have to hold that temperature for much longer to make sure you get enough collisions.
NICK HAWKER: Yeah. What you are alluding to is something very important for fusion, which is the fusion triple product. So we’ve talked about the first bit of that, which is the temperature, and then you need density, and you need time. So, the amount of fusion you get goes simply with density squared. With temperature, it’s a very nonlinear relationship and more complicated. And actually you can have very, very low density in very, very long time, and there’s schemes out there like that. People talk about magnetic confinement fusion, magnetic fusion. They have very low density, those schemes, and very long confinement times. And then what, we’re working on is inertial fusion, where with the complete opposite end of the spectrum; the density is off the charts high, compressing the fuel to more than the density of lead. And the time is extremely short, but actually both approaches can have the same triple product.
AZEEM AZHAR: So there are two general approaches; this magnetic confinement, on the one hand, I think some of us probably have seen pictures of the most expensive donuts in the world, which are these sort of [inaudible 00:11:40] holding with powerful magnets. And then, the other approach is inertial confinement, which is the approach that you’ve taken. Can you just briefly give us a flavor of what the density and the times look like? You’ve given us a sense of the temperature; you said it’s sort of 100 million, maybe a little bit less. How do we think about the density and how do we think about the time?
NICK HAWKER: So, in magnetic fusion, the density is actually lower than the kind of atmospheric density of the air in the room. Actually, it’s a vacuum, it’s a vacuum state, but the confinement time can be very, very long. Whereas in inertial fusion, we’re talking about tends that might be 10 or 100 times higher than even the dense metal, but the time scales are very short. So you’re talking about kind of one billionth per second.
AZEEM AZHAR: You’ve chosen to go down the inertial confinement approach. What are the advantages of it?
NICK HAWKER: One of the big advantages that we have is that we have a process that we can simulate. The physics of the process is amenable to simulation is attractable thing to simulate. The kind of advent of the computing power that we now have, and the simulation tools and methods that we now have make a huge, huge difference to the rapidity with which we can explore the parameter space. We are doing a campaign right now, an experimental campaign, it’s a 20 shot campaign, and we’ve done over 100,000 simulations going into the design of that campaign. So all of that learning that we’re able to do in Silico [00:13:17] allows us to make much more rapid progress than we would be able to make otherwise.
AZEEM AZHAR: And when you say, “Shot” you mean a single experiment, is that right?
NICK HAWKER: Yeah, that’s right. Each experiment, we call a shot. So we fire one projectile at one target, make a set of measurements and that’s an experiment.
AZEEM AZHAR: So with other approaches, the parameter space is really complex, it’s really hard to run these experiments in silico, as you describe it. And that means that all the optimizations and all the learning have to come when you crank up these devices, which is time consuming and expensive. And so, the learning rate is lower. Whereas because of the parameters of your particular approach, you do a lot of the learning before you ever press the sort of big green button to run the experiment?
NICK HAWKER: That’s right. That’s my hope anyways, that’s the thing. The biggest challenge we face is, are the simulations actually bright? That’s the question. So it’s all very well and good deploying best in class machine learning, for example, with the simulation tool, but what you are learning with that is simulated reality. It doesn’t tell you anything. You can be as fancy as you like there, it doesn’t tell you anything about whether a simulated reality is the same as reality or not. So what we’re able to do is, we’re able to break down our process into the different steps and we have different ways of making different kinds of measurements to provide that crucial validation data for the simulations.
AZEEM AZHAR: Is this approach only possible recently, given how computers improved in the last several years?
NICK HAWKER: Yes. And it’s the computers, but it’s also the methods for the simulation as well.
AZEEM AZHAR: Can you share with us what some of those sort of methods are?
NICK HAWKER: Sure. I hesitate to get too technical, but so, our main code uses something called the finite volume approach, where basically, you split up your region you want to simulate into lots of little boxes and we’re modeling loads of processes, all of which ultimately shuffle mass and energy around the system between all these different boxes. This gets very challenging though, when you have very different materials involved. So if you have hydrogen on one side, you have, say a metal on the other, you have something which in physical reality, you can kind of zoom into that join between the two materials. So this is an area where we have something we call front tracking, which is a particular method for tackling this challenge and allows you to simulate that accurately at a box size, which is trackable.
AZEEM AZHAR: So that’s fascinating. So simulation, which is one of the advantages of your approach has become more useful as a computing power, has got cheaper. What are the other advantages of your approach?
NICK HAWKER: So what we’re doing is, it’s a new method for inertial fusion, which we call projectile fusion. What we do is, we fire a high velocity projectile that flies for a short distance and it hits the target and the fuel is inside the target. So this is not dissimilar to how the National Ignition Facility works. So there, they’re using a laser instead of the projectile and the laser light shines onto the target, and what happens is, it causes the fuel capsule inside the target to implode and to get to very high density in the temperature. Our approach is exactly the same, but we have very different target designs, which work with that projectile input instead of the laser. So, the advantages of doing it this way, firstly, the machine, which more the projectile is a lot cheaper than the equivalent laser. It’s much more energy efficient. We launch our projectile with electromagnetic forces, so it’s like a rail [inaudible 00:17:04], and what you get is kinetic energy in the projectile. And if we look at the efficiency of that, that can be 10%, whereas if you look at the efficiency of National Ignition Facility, it’s 0.1%.
AZEEM AZHAR: The National Ignition Facility is using this sort of spherical array of 192 lasers. It looks really quite incredible, but its energy efficiency is actually much lower than the kinetic approach that you take?
NICK HAWKER: Yes, that’s right. That’s right. And that sort of leads me on to another advantage of projectile fusion. So if you have actually 192 lasers, then your reaction chamber and your plant is really quite geometrically complicated, right? It’s a very complicated thing to build in a very harsh environment. So with projectile fusion, we only have one entrance hole into the reaction chamber, which is just from the top. We simply drop the target in from above, it falls down under gravity and we fire the projectile straight down on top and it hits the target. The whole point of inertial fusion, how it works is, you form this incredibly hot, dense state of matter and it’s trapped by the inertia of its surroundings. It wants to expand. It wants to relax to that high pressure state that it has. Nature doesn’t want that to happen, right? It wants to cool down, but the surroundings can’t just disappear out the ways instantaneously. So it pushes on the surroundings, the surroundings push back on it, and it’s actually the inertia of the system itself. Once you’ve formed that state, it just takes a certain amount of time for it to dissipate again. And that’s where the kind of billionth of a second comes from. Nothing is holding the plasma together. Its own inertia holds it together for that very short time.
AZEEM AZHAR: But it’s also a much smaller type of device when we think about out the common picture of what a fusion system looks like; it looks like it’s a really big physical machine with lots of volume, but with inertial confined fusion, the densities are much higher. The volume is much smaller as well, in some sense.
NICK HAWKER: That’s true and sort of not true at the same time. The plasma at its final condition might only be 50 or hundred my icon size. So it really is tiny, but to absorb the energy that it releases, you need a big system. So there’s quite a large surroundings and to get the projectile to the necessary velocity, the necessary energy, you need quite a lot of capacity, so that’s quite a big system. But an advantage of inertial fusion is that driver system, be it projectile or laser, it’s very sort of physically separated from the place where the energy is released. [Tokamaks 00:19:41] are very integrated devices with a lot of quite complicated components all in one place. Inertial fusion is quite a lot more decoupled in terms of the engineering of the plant.
AZEEM AZHAR: So if I can summarize a list of advantages, as I’ve understood; you’ve got something that is easier to simulate, which means that you can hopefully go through the learnings and more quickly and more cheaply by doing a lot of this in the computer. You have an engineering approach, which makes this a simpler and cheaper than either using my magnetic confinement or using the laser-based inertial confinement that’s used for example, in the National Ignition Facility. And that additional simplicity has benefits around longevity because you identified that within the national ignition facility, you have sort of neutron spraying everywhere around these expensive lasers and so on. Whereas in your simpler system, the energy is being dissipated by neutrons, but there isn’t the same Ruth Goldberg apparatus of sort of expensive kit that’s being hit by it. So that sort of roughly seems like there of three distinct advantages. Is there anything more?
NICK HAWKER: What I would say is, we’re trying to achieve a power plant design that can be realized with existing engineering, with known technology from nuclear and other areas that you can just port straight into our device, maybe with some small tweaks and things. We think we have that. We have something which can be built from existing nuclear steels, that you can… It’s really boring stuff, but you can look up the steel we want to use on the internet and you can find a detailed guide of exactly how to weld it so that you don’t ruin the heat treatment. If you don’t have that existing material and you don’t know how to weld it you’ve got to have to figure that out. It just takes ages to go through that boring stuff. It’s a compromise, because you never get anything for free. The compromise is, in our core process, this projects, our fusion process, and I’ll be very plain about it, we have more to prove in the physics of that than many others out there. The level of evidence for the plasma physics of Tokamak is higher than the level of evidence we have for our scheme. And that’s why I kind of go to the simulations as, when you’re asked about, “Why are you doing it this way?” I think that there is a huge, huge amount of design space available in that target design that we have. And in a sort of stupid sense, I think of it like Minecraft; I’ve got like a 20 mil, by 20 mil, by 20 mil cube and I can put whatever material in there wherever I like to get maximum performance out from this projectile. We have gone past what I thought was the limits of the possible simple, several times already. Our best target design has improved by a factor of 10,000 over the course of 2021. I’m setting down the challenge to the team by summer, finally, a factor of 100. And we’re like, “Well, how can we possibly find that? We’ve maxed out our current designs.” And they’ve gone off and tried some crazy stuff we never thought of. And now it looks completely tractable that we’ll find another factor of 100.
AZEEM AZHAR: How far away are we from the [inaudible 00:22:49] most matters in the minds of the ordinary listener, which is when can we expect to get out more electricity from a system like this than we put in?
NICK HAWKER: I’m very confident that that challenge is going to be solved this decade. Very confident. I speak for the fusion community to say that there are magnetic few projects, which have very good chance to demonstrate that this decade there is more than one inertial fusion project that has that possibility. Actually, the National Ignition Facility had a really great result this year, a real breakthrough. They got what people call an energy gain. So en energy gain being the fusion energy out, divided by the energy input that you’ve used, so the energy of the laser in that case. They got to 0.7, And that actually equaled the magnetic fusion record which has stood at 0.7 since the 1990s. I thoroughly expect them to show energy gain next year. Now, will that be a commercially viable system? Not necessarily straight away, right? Because that laser is very inefficient and it’s very expensive, but there’s other technologies coming through; there’s higher efficiency lasers, and then we have our approach, which is the same physics, but with a different driver. Yeah, I’m very, very confident that the physics challenge be solved for this decade.
AZEEM AZHAR: So Nick, could you just clarify what we mean by the gain factor or Q, and what does 0.7 mean in this context? And what does one mean in this context?
NICK HAWKER: So in inertial fusion world, people tend to talk about gain and in magnetic fusion world, people tend to talk about Q, but they mean basically the same thing. So it’s the ratio of the fusion energy out, to the input energy going in. So if that’s less than one, it means you’ve consumed energy, which obviously it’s not what you need to do. And if it’s more than one, it means you produce energy, which obviously is the goal. For inertial fusion. You need to be at gains much higher in order to get it to work. The physics starts to change as you get a large amount of fusion energy release. And if you can show that you can understand that, and model that, and the experiment works at that low level of gain, then that’s de-risking the physics improving that it can be done.
AZEEM AZHAR: I wondered about that NIF results when I read about it in the summer, because 0.7 is, as you say, where magnetic confinement has been hovering since I think Tony Blair was Prime Minister of the UK back in 1997. So it’s a sort of a threshold that has been there for a couple of decades. And then, you’ve just alluded to the fact NIF might get the gain above one next year. And of course we want that gain to be well, well above one for these things to be viable, but no progress for sort of 20 years and then 0.7, then jumping to one in the next year suggests that the pattern of progress in this field is not linear; it’s more punctuated with sort of significant steps, or is it that we’ve reached a moment of compounding progress that we’re going to start to see a smoother steeper gradient of progress?
NICK HAWKER: Actually, for the National Ignition Facility, it’s been a real journey. They’ve been operating and doing these experiments for 10 years, there’s been a succession of improvements coming through. Perhaps I sort of think of it like overlapping S curves; get on a new idea, progress is slow to start with because you don’t know what it is, and you don’t understand, and you really get the hang of it and the optimization. And then, right at the very end, it’s just the tail end; there’s not much left to optimize. And there’s this sort of history and you can see it in the literature from NIF and what they talk about, of different changes being made and different designs coming through. The previous sort of design point for them had a gain of about0.05 that had this jump to 0.7, since they’ve done that, they’ve actually repeated it, and they didn’t get the same thing they got before. There was a slight change in the target, and that had a big impact on performance. So they’re right in that bottom of the S curve where they’re still exploring and understanding and very, very confident that they’re going to make that progress and get up further.
AZEEM AZHAR: If we think about your journey within First Light Fusion, what are the key challenges that you have to overcome?
NICK HAWKER: Yeah, so we are working towards a milestone at the moment, which is to actually show fusion with that process in the real world for the first time. The next phase for us is going to be to build a new experiment. So a bigger pulse power machine, a bigger electromagnetic launcher to get to high velocity projectiles, and to more energy out than so it’s gain experiment. That’s the next step for us. And then alongside the gain experiment we’ll be developing what we call the basic design of the plant, which is then more engineering detail for all of the systems that you need. So you don’t just know that you need the heat exchanger, which takes this and much energy from here to there at this temperature and this flow rate. You actually have made some design choices and done some work. So, that’s the kind of forward milestone, and that’s putting us in a position to actually then straight away start the first of a kind plant build at that point with the physics de-risk. And with that basic design in hand. Actually the problems that we face on a day to day level are so mundane at the other end of that spectrum. So, we had something the other day where we found we had some voids in a particular component in the target, which we didn’t previously know were there. And that the kind of difficult thing that challenge is you don’t know how many of those issues you have until you solve them all, and you’ve got the result.
AZEEM AZHAR: The target is a small piece of perspex that is stuffed with the deuterium and the tritium, is that correct?
NICK HAWKER: Yeah, it’s a precision engineered, precision manufactured object, which it’s very hard to model, hard to understand, hard to make, hard to figure out how to make it.
AZEEM AZHAR: So, the heart of this of course, is this thing that you have to design, first of all, and then manufacture within precise tolerances and a sort of mundane example of the kind of challenge that you face is that, actually the steps were mucking up the realization of the original design and you are dealing with presumably dozens of these things, some of which you know today, and some of which are yet to be discovered.
NICK HAWKER: Yeah, that’s right. So we’re constantly going back and forth between the simulations and the experiment. And whenever we’re looking at the experimental results, we take the simulation results as gospel. And we say, “The simulations are right. There must be an experimental problem. Where is it? Let’s find it.” And then completely other way around, we, when we’re looking at the simulations, we’re assuming that the experiment has been executed perfectly. “Where is our physics modeling incorrect? Where could there be some additional process or some different thing happening?” We keep going with that, and then if we do that, we’re going to get there.
AZEEM AZHAR: You painted a picture of what the next several years look like. And you’ve also talked about some of the types of issues that you have to contend with. The way you solve these obviously will be with people. And so, I’m curious about what that experience has been for you as a founder and a CEO, and starting to think about how you build the kind of team that needs to tackle these problems. Because I will confess, if you said to me, day one, you’re starting up a fusion company, who’s your first hire? I would not know where to go.
NICK HAWKER: Yeah. I think the skill in managing a very complex program, like this is getting the best out of all the skill sets of people that you have and giving proper ownership over, so we break all our work down into milestones, which are kind of six to 15 month kind of pieces of work. And they each have a milestone PI, milestone leader who is responsible for delivering all of that work. And so that’s principal investigators. So we have milestone PIs, and then the milestone’s broke, down to projects, then we have project PIs. And once you are in charge of project and you have been given a shot budget, that’s it. It’s yours. That’s really important to give that kind of level of empowerment to people. And I think that’s how we make the progress, right? In many senses, it doesn’t matter if you do A, then B and C, or B, then A, then C. If the team wants to do it one way, or someone in has a real great idea or thinks they see a path forward in a particular direction, then we let them go that way, and we trust that the whole thing will be getting need to.
AZEEM AZHAR: I want to share what some reflections on that. One is that the team is complicated. You know, you have a machine learning and data science team, that’s building simulation models, you have high performance computing team. You’ve got physicists, you’ve got literally a workshop that makes physical widgets, some of which are really small. And then there’s like a 25 meter long steel gun barrel. There were bags of gun powder caged away. I mean, it’s a complicated setup, it’s not like building a dating app at all, it’s as far away from it as possible. And then, the cycle time that you’ve talked about, it’s really interesting. It’s quite daunting. Six to 15 months, which is short in academic terms, but in traditional startup terms, even traditional hardware startup terms is quite long. And, I’m quite curious because there isn’t something that you can just pick out of academia, or pick out of a traditional software based startup and use as a playbook, but do you think that there is sort of a kind of a commonality of approaches that you and your team are starting to understand that work, that are distinct to other types of organizations?
NICK HAWKER: There are things that we, I think, understand, and as a team understand and know are important. I don’t know how unique it is compared to other organizations. We very much have an agile management philosophy, and if I can say something slightly rude-
AZEEM AZHAR: Of course.
NICK HAWKER: A lot of the agile management books and things out there are written by people who have built dating apps, as you allude to, and ultimately, it’s just not that hard to build a dating app and it doesn’t really matter how you manage it, because it’s just not that hard. So, it really does matter how you do it here. I think I saw a tweet from Elon Musk, which said, “The rate of progress is the amount that you learn per iteration, how long it takes you to do an iteration.” And I thought, yeah. And I thought about it a bit longer and the problem with what that implies is it implies that you do an iteration, you learn, and then you design the next one, and then you do the next one, and then you learn, and then you design the next one, you do the next one, you learn. If you’re doing it like that, you’re going to have a maximum velocity that you can get to. The complicated bit: how do you overlap the iterations and still make fantastic progress? One way that you do that is you run ride along experiments, right? You have experiments which have a primary purpose, but then have secondary or tertiary purpose that don’t put any risk on the primary part of the experiment, but advance you somewhere else, in some other area. I’ll give you a really simple example. In a recent shot, it was a perfusion shot where we don’t have any other diagnostics, other than the neutrons of [inaudible 00:34:17] to know if we’ve got fusion or not. We were testing whether the target was moving during the shop, by shutting a laser off the back, completely decoupled from the actual real experiment we were doing, but it allowed us to transfer that. But then, that’s not always possible. So, sometimes you’re in that tightly coupled loop, the tightly coupled problem. How do you make progress there? I wish I had a concise answer for that question of how you overlap the iteration cycle, it’s easy articulation, but it’s just really complicated.
AZEEM AZHAR: Let’s zoom out a little bit and talk about commercialization. Is commercialization something that’s on your mind at this point?
NICK HAWKER: Yeah, absolutely. I mean, no one wants to spend their life working on something which isn’t going to be used, right? Or I don’t. I want to see it and the real world.
AZEEM AZHAR: So I would like to just flip to what fusion reactors would look like when they’re connected to, to the grid, right? What scale will they be? Will we have hundreds of them, or will we have thousands of them or tens of thousands of them?
NICK HAWKER: I think fusion will be a smaller scale than nuclear, although it does depend on which technology approach ultimately succeed, or which of the set of possibles will succeed. Some of the more mainstream approaches do go towards that very large kind of one gigawatt electrical size, all of the startups, we are all pursuing designs, which aim at smaller plant sizes. So for First Lights, we are working towards something similar to 150 megawatts electrical, that sort of size, but it’s quite hard to kind of grasp what that means, I find. So fusion is big. The fundamental physics of it has this time scale in it, as we’ve talked about; you can translate timescale into a length scale. So there is always an advantage in the size in terms of the plasma physics. So fusion for me is actually not a disruptive technology; the disruptive technology in energy, actually more than any other is solar power, because you can buy solar panels and you can put them on your house, but that’s not going to power us. That’s not going to meet the electricity demand which we’re going to have, we’re not going to get enough energy doing it that way. So yeah, fusion is still centralized generation with a thing that you would call a power plant in a specific location, which is disputing power out onto the grid. It’s a revolutionary technology, but it’s a continuing technology. It’s still the same business model for all the power plant builders out there, and you and I, when we interact with energy, it’ll still be, you turn on the light and the light turns on and you get it from the National Grid. So yeah, fusion fits in with that existing paradigm.
AZEEM AZHAR: 150 megawatts would be a sort of medium size coal or gas power plant capacity to power 150,000 homes. It’s still significant. And is your view that, there are obviously, to your point, advantages in being able to have technologies that can operate at different scales. So solar can operate on a single rooftop at couple of kilowatts and it can operate in these utility scale farms, much, much higher capacities. And one of the things that does is it provides cost advantages, because you get sort of scale effects. It also provides market advantages because you have many, many different types of operators who can compete to provide power that’s solar generated. And you are very frank, I think about looking at where you expect fusion to end up, which is actually quite big, which means that only certain types of companies will be able to afford to build it and then to connect it to the grid. Because it’s one thing to connect my solar panels in my garden to my house battery, it’s another thing to plonk 250 megawatts of generating capacity into the grid. It reduces the number of players. So that for me is quite a fascinating way of looking at the challenge. So does that mean that your sense would be that the people who will run fusion plants, the firms would look quite similar to the types of firms that currently run nuclear or fossil fuel based plants?
NICK HAWKER: Yes and also no. So let’s separate out the people who run the plants from the people who actually build the plants. So Westinghouse there’s PDF there’s the Chinese designs, there’s not that many. As you get further out from that sort of core point, there are many more engineering companies that are involved. So you end up with groups who can be called architect engineers, who assemble all this technology to produce a plant and put it on the grid. And they hand that over to the operators of the power plant. And there are many, many operators of power plants around the world. Megawatts, that sort of scale is a lot of the balance of plant is literally catalog items, at that level. As you get up to kind of one gigawatt level, it starts to become a research project how you build the turbine for that. So you’re are incurring engineering risk. Whereas plant there’s a turbine from Siemens where there’s literally a brochure. I was Googling it. I literally found one on eBay for $2.5 billion. I don’t know what an Earth it was doing there, but it was an old one that was like 30 years old. It’s just not that unusual. The other part of it is total financing cost as well. So we’re aiming to try and keep a total Cap Ex bill, less than a billion for that power plant. That becomes something which is then much, much easier to finance, just like project management, project financing risk. And then, the kind of key mistake all of us in fusion need to avoid is, we’re going to build the firsthand, and it’ll work, and it’ll be kind of a bit rubbish in a lot of areas. And there’ll be loads of improvements that you could make. Don’t. Don’t bother. You’ve just figured out how to make one, just deploy them as quickly as you possibly can, because we need as much electricity as you can possibly get to solve this dual problem we have; climate change and human development, right? We need more energy. Just build them.
AZEEM AZHAR: Nick, you’ve got a very pragmatic approach to this. There are a couple of dozen other startups tackling this problem, so we might have some progress. What would you say is a likelihood that by 2070, we’ll have fusion powered electricity in our grids?
NICK HAWKER: Very high. I think in 2070, we are going to be pulling an awful lot of carbon back out of the atmosphere and it’s going to need a gigantic energy source to do that, and I think fusion is the prime candidate.
AZEEM AZHAR: Okay. So what about 2050? 20 years earlier? What’s the likelihood that we’ll have fusion powered electricity in our grids?
NICK HAWKER: Yeah. I still think very high, but I am a realist. And to think that fusion is going to be a dominant energy source at that time, I don’t think is realistic. If you look at how quickly other energy sources have been built out, I think fusion can be a part of the mix at that time, but it’s very important from the point of view of net zero and climate change, we don’t get distracted from the things we have to do now. We have to be building out solar and wind the maximum rate we can.
AZEEM AZHAR: And what about even closer, in 2030? Do you think there’s any chance that there might be an economically viable fusion reactor connected to a grid somewhere?
NICK HAWKER: I think there’s a very, very high chance that there will be a fusion reactor connected to the grid somewhere. If it’s one of ours, I think it would be economically viable in that first plant.
AZEEM AZHAR: Now, fusion seems like a really collaborative endeavor between different groups, public and private, and there seems to be some friendly competition. Whereas in other deep tech carriers like quantum computing or artificial intelligence, there seems to be a much, much more competitive environment. Is that how you feel it? And does that change the way that this burgeoning industry is operating?
NICK HAWKER: I don’t feel it as a hugely competitive environment, actually. There was a time when the private fusion enterprises were sort of looked on with an element of distrust from the public efforts. But I think that has changed. I think there is a much broader church now in terms of community. Literally yesterday, I and four other startups were presenting to Europe Fusion Consortium, presenting at one of their internal workshops. So, I think that the kind of barriers between the different areas are coming down, and I think that’s really important, because there is a lot of cross-learning that can happen. We don’t know which technology is going to succeed yet. We should be pursuing all of them. And actually, everything is going to be limited by deployment rate. Like solar and wind are limited by deployment rate, I think fusion will be limited by deployment rate. Any extra source of clean energy we can get, we desperately need, so I don’t care if there’s five approaches for fusion, there’s a massive market for all of them, and we need as much as we can get.
AZEEM AZHAR: Nick Hawker, thank you so much for taking the time with me today.
NICK HAWKER: My pleasure. Thank you very much.
AZEEM AZHAR: If you enjoyed this discussion, do check out the podcast feed. You can listen to previous discussions about the future of our energy mix, and how tough tech like nuclear fusion can get us there. That’s something I’ve spoken about with technologists and investor Ramez Naam, decarbonization economist, and [inaudible 00:43:34] year and early stage climate tech investor, Dawn Lippert. They all speak brilliantly on what it will take to make green grids a reality. Now to become a premium subscriber of my weekly newsletter, where I cover this and other topics you can go to www.exponentialview.co/listener, where you’ll get a 20% discount to stay in touch. You can follow me on Twitter. In the US, I’m @Azeem, A-Z-E-E-M, and elsewhere, @Azeem, A-Z-E-E-M. This podcast was produced by Mischa Frankl-Duval, Fred Casella, and Marija Gavrilov. Bojan Sabioncello is our sound editor.