Instigators of Change

Star In a Bottle: Can Fusion Solve Our Energy Crisis?

Khosla Ventures Season 1 Episode 17

Imagine if we could harness the power of a star, right here on Earth. Our energy challenges, as we know them, would disappear. That’s the promise of fusion technology, a way of generating electricity that scientists have been working on for decades, but have never quite cracked. Are we any closer in 2022? Commonwealth Fusion Technology, a venture-backed company spun out of MIT, is working on a prototype reactor that they believe could begin replacing conventional power plants as early as 2030. Commonwealth’s CEO Bob Mumgaard recently sat down with Vinod Khosla at KV's CEO Summit in the Bay Area to talk about how their technology works, what challenges have yet to be overcome, and what fusion would mean for the planet.

Kara Miller:

Welcome to Instigators of Change, a Khosla Ventures podcast where we take a look at innovative ideas, the people who come up with them, and those who invest in them. I'm Kara Miller, and this week, what if everything you thought you knew about energy and how bad it is to use it is wrong? Today, Bob Mumgaard, the co-founder and CEO of Commonwealth Fusion Systems, talks about an idea that grew out of MIT and was once considered a pipe dream.

Bob Mumgaard:

It's a way to boil water the same way a coal power plant or a gas plant is, but it's using a fundamentally different reaction.

Kara Miller:

But if fusion technology works, it could pull us away from fossil fuels and make energy incredibly abundant. In this special episode of Instigators of Change, we will hear Mumgaard in conversation with Khosla Ventures founder, Vinod Khosla, at KV's annual CEO summit.

Vinod Khosla:

People have been skeptical of fusion. The running joke is it's 30 years away and it will always be 30 years away. But tell me what got you thinking differently about fusion. Maybe start with what you're actually building.

Bob Mumgaard:

Why do you do fusion? Why does fusion end up in science fiction and things like that? I think that's been... The thing that's really been clarifying is the climate crisis. If you've got the climate crisis, that's fundamentally... We have too much carbon in the air. How did it get there? Well, we dug up a bunch of it and we put it through a reaction that made CO2 and that's the most prominent reaction in the world and that's how we powered our entire society. 50 gigatons a year. That's got all these big downsides. But fusion has been out there as this thing that is a completely different reaction. It's a reaction that's 200 million times more energy per reaction than coal.

And so it's really an interesting... We used to think the stars were 5000 years old because that's how long they would last if they were made out of coal, but once we found that they were made out of fusion, it meant that they were 10 billion years old. And so if you could do that on Earth, it'd be this big deal and you'd go from a world where you were consuming resources to a world where you just had energy from a machine. People have been searching for that for like 50 years. Trying to building machines that can create the same reactions, the same conditions inside the stars, but on Earth. That we can control. If that were to work-

Vinod Khosla:

[inaudible] reactor.

Bob Mumgaard:

Yeah. So, they built-

Vinod Khosla:

I still vividly remember where I stepped on this [inaudible] reactor. Bob said, "You know, a few minutes ago, this was 600 million degrees below you."

Bob Mumgaard:

You have to build a star in a bottle. That's not an analogy. That's literally what it is. You have to build this machine that can create the conditions that the stars do. We've been doing that scientifically in very large nationals labs and universities for a long time. We've actually been making really good progress to make that reaction happen. It's faster than Moore's law in almost parameter you could look at scientifically. We're sitting there within a factor a few of making it useful. Of actually making the reaction so vigorous that it makes more power than it takes to actually start the reaction.

The world was on this trajectory to get that done in 2050, which is a little bit too late for climate. It's actually quite a bit too late. And so what had happened at MIT is we stumbled across a technology that was completely unrelated to fusion that, if it could be developed quickly, could be applied to fusion and change that trajectory. Basically, bend it into near term. It was a technology that someone mentioned earlier about superconductors. It was a new material. This is something that came out of material science labs. Not out of physics like where we came from. If you could do that, you could take all the science that we'd done until that time and all the demonstration that we've done and machines like at MIT and France and the UK and you could put it into something that was like a product. Something that was economical. Something that you could demonstrate soon.

And it was at the right scale. You could think about, well, you might be able to actually do this commercially. You might be able to do this as a company. So, instead of doing as an international collaboration of scientists, could you do it as a company? It was really a confluence of a bunch of events. That material, the science from the fusion side, the climate poll, the position that MIT had found itself in terms of having a bunch of talent altogether, the funding market as it was coming back from Cleantech 1.0. All at the same in 2015. 2016.

Vinod Khosla:

So, the ITER project is the multi-nation... How much is it now? 20 or 40 billion or something?

Bob Mumgaard:

It's like 50 billion now.

Vinod Khosla:

50 billion dollar project. Bob decided it was possible to compete with that project that'd be done by 2050. Tell me what made that possible. Why you thought you could take on something with 1/50th the resources. Back then, you didn't even have a billion dollars. What made it possible?

Bob Mumgaard:

This goes to... You're trying to do something new. You have to... The fundamental principles. In these fusion machines, the fundamental principle is they are basically magnetic bottles. They hold the star stuff with magnets. They need to insulate that very, very well. You have a hundred million degrees in the center. It's the hottest thing in the solar system. Our machine at MIT is literally the hottest thing in the solar system. And then, a meter away, you have room temperature. So, very well insulated. That's because of the magnetic field. It turns out that the magnetic field... The power of a fusion machine is the magnetic field to the fourth power. It's this huge lever. We'd shown that over and over again. That if you just change the magnetic field a little bit, you could change the size of the machine by quite a lot.

Everyone had always been up at the edge of the magnetic field they could build with technology. And so what we did is we took this new material that was a new type of superconductor that eliminated a limitation on the magnetic field that had existed before and we said we think we can actually get enough of this superconductor together and we can actually build a magnet that's twice the magnetic field than anything existed before. A complete step change in the 300-year history of magnets. At twice the magnetic field, you could then take a machine that was like a large multinational machine like ITER, which is the largest construction project in Europe, and you could put it... Instead of on a stadium, you could put in a garage. That's the type of big shift that you look for in a new technology. Not an incremental thing.

We went to the scientific community. Very transparent. We published. We collaborated with everybody. And said, "Hey, what do you think of this idea?" The conclusion was, yeah, if you had that magnet, that machine would work and it would be that small. And we don't see any show stoppers except for that magnet. And so that's laid out a very, very concrete plan to do risk retirement in this field that people have worked on a long time of, first, build the magnet, then apply it to that machine, demonstrate for the first time ever that you make more power out than in from a fusion reaction on Earth, and then that is a product and a power plant that we know how to deliver. It's a way to boil water the same way a coal power plant or gas plant is, but it's using a fundamentally different reaction.

Vinod Khosla:

How long do you think it took ITER to build their magnet?

Bob Mumgaard:

We actually just did a study on this. We're running at about 10 times the speed of ITER. Whether that's in how fast it takes us to build... building... or fast it takes us to develop a magnet. It took ITER about 25 years to develop their magnet technology.

Vinod Khosla:

It's a very old project.

Bob Mumgaard:

Yeah. Yeah. This is the scale like the large [inaudible], right? It took them about 25 years. We were able to go through all the same stage gates in just over three. It was one of these things that... One of our physicists was just giving a talk at the main plasma physics conference. Middle of the talk, standing ovation. Spontaneously. When they talked about this magnet.

Vinod Khosla:

That's awesome. That's why the world needs entrepreneurs. And only entrepreneurs can change the world. But our first task was then making the magnet. We had to raise money. We used a very entrepreneurial pitch. Said, if fusion doesn't work, what would happen to the 100 million dollars you had to raise to build the magnet?

Bob Mumgaard:

Well, it turns out that the magnet was useful for other applications. It's useful for things like wind turbines or things like MRI machines or things like generators. And so we were able to line up a bunch of different applications and said, hey, this isn't so binary. Yeah, it's new, it's novel, but... It has this big upside of fusion, but if it doesn't quite go right, we've still really advanced the state of the art in these other areas. Actually, we're just shipping a magnet to the University of Wisconsin that's unrelated out of the same program.

Vinod Khosla:

And so how long did it take you to build your magnet?

Bob Mumgaard:

First, we broke a lot of things. We built stuff and you're at the cutting edge. And so just build it, test it, test it, test it, break it, build the next one, and build that cycle as fast as possible. I remember we were going to a big national test stand in Switzerland. We called them up and said, "Hey, we want to show up and test every week." They said, "No, no. People test every year." We said, "No, no. Every week, we're going to show up with a new sample." And these are big, big samples. They didn't take us seriously, so we said, "What's your operating budget? We'll pay for the entire year. And if we don't use it, you can have the money." They said, "Okay, fine." And so we showed up every week and just threw things through their test stand, pushed them to failure, took them out, smoldering thing, looked at it, took it apart, said, okay, called back at MIT and CFS, build the next one. About 20 iterations of that subscale until we got confident enough that we had the technology and could go to full scale.

And then we went straight to full scale. Not intermediate. This thing has got to work at this scale. It's got to be 10 tons. It's got to carry 10 million amps of current. It's got to have the most superconductor in it that's ever been put in this by a factor of 100. Let's just try it and do it. And we built that magnet. It took us two years to build the full scale one, which is still... I mean, why did it take so long? It takes long to cut metal. It's like, damn it. Why can't we cut metal faster? All the AM companies. We eventually, pretty much on budget, on schedule, got the full scale magnet. And it worked straight out of the box. We live-streamed it.

Vinod Khosla:

It took about three years.

Bob Mumgaard:

Yeah. Yeah, the subscale R&D was about a year and then the big build was about two and a half.

Vinod Khosla:

So, last Labor Day weekend was exciting as we watched the magnet come up to 20 Teslas slowly. A long time.

Bob Mumgaard:

It's like a day long party of everyone that had built it watching on pins and needles... and we live-streamed it to all the stakeholders and people that had Blink... of Tesla by Tesla by Tesla. It was a 20 Tesla magnet, so 20 Tesla. An MRI machine is like a Tesla, a Tesla and a half, so this is really, really high field.

Vinod Khosla:

So, you eliminated the key risk, which was can we build a magnet? What risks remain and where are you? Tell me. When we we know this actually works?

Bob Mumgaard:

One of the techniques that we really thought about was you have this new-

Vinod Khosla:

I just have to set the stage. ITER started 25 years ago. They expect to have proved that they can generate more energy than they consume by 2035. Roughly.

Bob Mumgaard:

Yeah. They're going to be late, but yeah.

Vinod Khosla:

You started three years ago. When will you prove this?

Bob Mumgaard:

'25. Right now, we are in the act of building a machine. It's called SPARC. It's a fusion machine easily fit in half of this room. And that machine is part of a building. It's all first of a kind type stuff. It's a project an hour outside of Boston. It's got about 1000 people that work on it. It's about an 800 million dollar project. But it will be a machine that, in 2025, you push a button and, for the first time on Earth, you will make more power out than in from a fusion plasma that's about 200 million degrees. Cooling towers will have a bunch of steam go out of them. You let your finger off the button and it will stop. You push the button again and it will go. That's a new paradigm for an energy source. That's like lighting light bulbs with fission the first time. Or fire. Things like that. And so that's something that the world has been waiting for since we basically understood how stars work.

Vinod Khosla:

It is very exciting. Now, you'll prove this, but what happens next?

Bob Mumgaard:

Right. That's still a demonstration. I think this is one of those things that... These really big problems... I think about big problems we all work on. I work in energy. One out of every 13 dollars that's spent is spent on energy. It's 10 million people that work in that field. These are huge problems. So just having a proof of principle isn't enough. You've got to be able to make it into something that is infinitely replicable. That has deployment models that can be scaled around the world. It can't just be this one-off scientific jewel. So, everything that we've done is to build it so that we can build 10000 of them. When you're in a meeting at CFS about how we should design the magnet, someone will say, "How do you make 10000 of them?" Because if you make 10000 fusion power plants, that's enough to be about half of the global energy use. There's about 60000 power plants in the world today. And that's what you have to do for climate. By 2050, you have to do that. Otherwise, the world burns. It doesn't matter if it's fusion power plants or solar power plants or whatever. There's that many of them. You have to go out and be able to build them as if they're Walmarts or build them as if they're 737s.

So, you start with that in mind. You say, okay, that means we got to make this thing into a product that is as simple as it can be. That can be built by as many people as can build it. We have to go as fast as possible to show it works and that it can be made. And so we'll do that in SPARC. Even in SPARC, even though it's a demonstration, it's built by a regional constructor. Someone who's used to building Amazon warehouses. Not by a very specialized, fancy firm. For construction. For pouring concrete. It's built to scale.

And then we'll move to ARC, which is a little bit bigger. About twice the size. That will be about a 200 to 400 megawatt plant. That's the scale that is a lot of coal and gas plants. And so that means you could have a model where you show up where there's a coal or a gas plan and you build this machine and you hook it to that grid connection. Now, you're replacing something. You're not having to reformulate the entire regulatory regime or the entire model in which we finance power plants or any of that. You're co-opting all that infrastructure, but you're just changing the reaction inside the machine. And so that machine we hope to start... We will start it in parallel to SPARC. We're not quite sure yet. But turn it on, 2030.

Vinod Khosla:

Explain this concept of fusion boiler. This is a fusion power plant. We've talked a lot about it.

Bob Mumgaard:

It sounds really crazy, right? You have fusion and it's in Doc Ock comic book stuff. But at the end of the day, ti's just heat. You are making heat the same way the Sun makes heat. You have to capture that heat and you have to turn that to work. It turns out we've been making heat to work since watts. It does that the same way. You make, literally, a fusion boiler. Except instead of being heated by the combustion of carbon into CO2, it's heated by the fusion of hydrogen into helium. Out of that comes a pipe and out of that pipe comes 650 degree heat in the form of moving fluid. In our case, it's molten salt, but it could be steam. It could be other hydrogen helium. And that goes into a heat exchanger and eventually makes steam and eventually turns a turbine. The good news is that we know how to do that. 90% of all the energy we use today, that's exactly how it's done. All that infrastructure is already there. It's already sitting there. It's been refined by 100 years. And so the fusion power plant is the whole thing, but really, the innovation is in the boiler part. It's in this machine that makes a bunch of heat using this reaction.

Vinod Khosla:

Power plants have very long lives. We've got quite a few in the US. Probably 5000 coal plants or natural gas plants. Those have 50-year lives. These are businesses. They hate the idea of having to shut them down. Obsolete them. How do you help them? How do you get them on your side instead of fighting you in Washington?

Bob Mumgaard:

The first thing is to realize, in almost everything a crowd like this does, there are a lot of stakeholders. If you want to change the world, it turns out your stakeholders are the entire globe. And so you've got to factor that in from the very beginning. In the case of a power plant, the stakeholders are the people that own it. They're the people that live near it. The people that work at it. They are the people that supply it with goods and services, whether that's fuel or parts. And so they have to be a part of this. Otherwise, you're fighting a very, very big uphill battle. You can't show up and build a power plant in a community that doesn't want it. It's just not going to work.

And so that means it's not just the technology. It's actually all the other pieces around the technology that you have to get in front of. If you're out there building an entirely new industry, you literally have to write the rules of that industry. We had to, for instance, go to the federal government and write what we think the regulatory rules for this technology would be. And pass legislation to get those rules as the rules. And then we had to go find communities that wanted this. And so well before we ever decided where to build SPARC, we were out in many communities around the United States, asking, hey, what do you think about this? And trying to find places that were very excited about it. We ended up finding one in Devens, Massachusetts that understands technology, very concerned about climate, high tech jobs. They were like, yes, let's do this. Let's put this right next to our water shed for our river and show that this can be done in an ecological, safe way.

And then also on the people that own power plants. We looked and said who's going to lose in the fossil fuel transition? Well, it turns out there's a lot of money in big oil companies. What's their plan? And we asked them. Like, hey, what's your plan? Well, we have a gap. We're not quite sure what this looks like. And we were like, well, come learn about fusion. Through a couple years of talking to them, a couple of them, probably the most progressive ones, they got really excited. Like, we could actually see a future here where our expertise of how we build pipes and how we build projects and how we run platforms... That expertise could actually be used to scale this technology. They became some of our biggest advocates. We're the largest investment ever for ENI and Equinor, the Italian and Norwegian oil companies. That also means that they're in Washington right alongside us, saying, this is what the future of energy should be. This will be zero carbon. We want to be a part of this. Let's go forward.

The power plant owners, likewise. They don't want to write off those assets. They're going to be forced to stop running those fossil fuel plants because they can't omit. So, what do you do with that land? What do you do with that infrastructure? Well, if you had a technology like this, that would be a really big deal for them. That means that that community still keeps jobs. That means that community still gets taxes for making energy. That means that the site and the infrastructure continues to be reused and it can stay on the books as a profitable thing for that business. And so by pulling all those in way earlier than what you would normally think... Don't wait on a go to market strategy until after you have the product. Pull them in really, really early. And then build a coalition. And we've been able to do that with pretty good success.

Vinod Khosla:

I thought this was brilliant strategy. Instead of all the pressure on every coal power plant to shut down and write off their assets and basically be out of business, to say we'll just build you a new boiler, which is replaced regularly in coal plants anyway. It's the part that needs the most maintenance. You cut your time to build a power plant substantially. No regulatory. No permitting or minimal permitting. I thought that was brilliant strategy.

Bob Mumgaard:

And the local community doesn't have particulate emissions anymore. There's fewer trains delivering coal. There's a bunch of benefits. Actually, the one that we haven't figured out how to get them on board is the train companies. The train companies end up not winning. So, if anyone has ideas, see me.

Vinod Khosla:

There was a surprising piece of research I did about 15 years ago and wrote a paper on it. On this energy transition. Most people don't realize that, around coal plants, there's more radiation because of the ash that goes up in the air and more radiation deaths than around any nuclear power plant. Period. Coal power plants are radiation hazards and you get to avoid the problem. But let's be optimistic. You might say pragmatic. You build a power plant in the next seven or eight years. In 2030, we have our first one up and running. How do you replace all 5000?

Bob Mumgaard:

This is a general problem with anything that's fighting climate. Most of what we do in entrepreneurship... Your timeline is set by maybe the cost of capital or maybe your competitors. There's some social dynamic to it. In climate, the timeline is set by a molecule that's in the air. It doesn't care what we do. The world is going to continue to heat. Unless we stop putting more of those molecules up, it will get really, really bad. You have to stop by 2050. And so no matter what you do in climate, if you're going to displace fossil fuels, you have to be able to build that at a speed that is not like we've built anything like that before. It's a speed that's closer to the way we've rolled out cellphones than it is the way we built even wind or solar. It has to drastically accelerate. But the good news is that we know that's doable. We have instances in history where we've done that. Whether it's building the Liberty ships for Arsenal for Democracy in World War II-

Vinod Khosla:

Talk about the Liberty ships example.

Bob Mumgaard:

You think about how much work it is to build something really big. It's usually very slow. But it turns out, once people got organized and said it needs to be done, they figured out how to build ships, thousands of ships in Los Angeles, an assembly line of ships, to fight World War II. And they took Ford from making automobiles to making way more complicated Boeing bombers. That was because Boeing adapted how they would make those bombers so that Ford could make them. There's this huge industrial base that exists that can be co-opted to do these types of problems. And it's been done several times in history. Even in energy. France actually decarbonized their grid. France is completely decarbonized. They decarbonized their grid in 10 years, in the '50s, and they said we're just going to build nuclear power plants and we're going to stamp them out as if they are products. As if they are not one-off stick build things. As if they're all the same and here's the foundries that make the parts and these are the people that go install them. And just go from site to site to site to site. They've been able to do that. And they supply electricity to almost all... A lot of the clean energy in Europe comes from that fleet.

Vinod Khosla:

So, here's the numbers. Before the Second World War, the US had built 10 Liberty warships in the decade before the war started. Not far from here, the San Ramon Iron Works was started. In the next five years... And Liberty warships are complex beasts. In the next five years, during the Second World War, the US built 4600 Liberty warships. I think you have a shot at building 4600 fusion boilers, which are simpler than Liberty warships. It's the most optimistic story I've heard on climate. I definitely wish you great luck because, if we're going to solve the climate problem... and most people give up and talk about how it's not possible... I think there's a ray of hope here. Frankly, more than a ray. It's real optimism because I think it can be done. Thanks very much and we'll break here.

Bob Mumgaard:

Thank you.

Kara Miller:

I hope you enjoyed this special episode of Instigators of Change featuring Khosla Ventures founder, Vinod Khosla, and Commonwealth Fusion Systems CEO, Bob Mumgaard. If you want to hear more conversations about seemingly impossible solutions to some of our most challenging problems, remember to subscribe to the show wherever you get your podcasts. I'm Kara Miller. Instigators of Change is produced by Matt Purdy. I'll talk to you next week.



People on this episode