Episode 75: Feeding the Matrix: Realta Fusion

Molly Wood Voice-Over: Welcome to Everybody in the Pool, the podcast where we dive deep into the innovative solutions and the brilliant minds who are tackling the climate crisis head-on. I'm Molly Wood.

This week I’m continuing my series Feeding the Matrix about AI and energy use

You’ve heard from the chip and software giant Nvidia the cloud computing giant AWS otherwise known as Amazon

Both of those companies have sustainability commitments and customer demands and economic realities that have them pursuing energy-efficient technologies and techniques in their big huge computing facilities known as data centers

But you’d be forgiven for thinking that maybe considering how fast companies are trying to build new data centers we might also want to start working on some like HIGH IMPACT solutions

So for the second half of this series let’s talk innovation like the holy grail of clean energy generation fusion virtually limitless carbon free energy but without all that pesky trying to connect to the grid business and designed to deployed RIGHT alongside a data center or something else that needs a bunch of clean energy.

TAKE TWO: 

Kieran Furlong 

my name is Kieran Furlong and I'm the CEO and one of the co-founders of Realtor Fusion. We're a company that spun out of a large experiment at the University of Wisconsin at Madison and we're focused on developing and deploying zero carbon, always on fusion energy.

Molly Wood 

Wonderful. Where did the name come from?

Kieran Furlong 

the name, it's good you ask. Because this is my opportunity to try and tell all of your listeners how to pronounce it as well. It's realta, it's realta, that's how we're saying it. Any listeners in Ireland who speak Gaelic may quibble because there's different pronunciations there too, but it comes from the Gaelic word. It's the Gaelic word for star.

Kieran Furlong 

So being from Ireland originally, I thought, there's a unique name, but that's what we could use. And I like it because we're making fusion real, which is the clue on how to pronounce it as well.

Molly Wood 

That's so great, thank goodness I asked that. Gay look for star. You stinker, that's outstanding.

Molly Wood 

Let's start with an overlay of the Fusion energy landscape. We have done one other interview about Fusion on this show, but I think it's good to point out that there are lots of companies, but also lots of technologies in this space. Tell us where we sit, and then we'll dig into your version of Fusion Tech.

Kieran Furlong Ha

Kieran Furlong Sure.

Kieran Furlong Sure, sure. think first of all, it's good for everyone to be aware that fusion is the core energy, if you like, right? It's what powers the universe. It's what powers our sun. And of course, all of the energy that we consume on Earth is solar energy in one way, shape or form. It's either kind of biomass, like charcoal on your grill today, or ancient biomass in fossil fuels, solar power, or wind that has, you know, that's the currents that have been driven by the atmospheric

heating and cooling and so on as well. So fusion first of all is this ore energy, right? And really it is the fusion of very light elements. So ultimately this is how all of the elements in the periodic table were originally formed. But essentially we're going after something called, well, we're using two isotopes of hydrogen, deuterium and tritium, essentially heavy hydrogen and even heavier hydrogen.

fuse those together and they produce helium. So essentially we're bringing the nuclei of atoms together and we're basically pushing them together under extremely high energies causing them to fuse. When they fuse they release a small bit of the mass that they started with as a massive amount of energy. So essentially that's what we're trying to tap into. It's just very difficult to do, especially very difficult to do in a controlled fashion. So that's what we're working on.

Molly Wood Right.

Kieran Furlong Within fusion itself, then there's different flavors of it. So, kind of the two big branches that folks may have heard of are inertial confined fusion and magnetic confined fusion. And so our technology sits under magnetic confined fusion. We're essentially, we're trying to get to a plasma, which is, you know, the fourth state of matter. It's where you basically keep heating up or energizing a gas to get to a point where you begin to strip the electrons off the ions and you've got charged particles just bouncing around.

To do that you need get very high energies, essentially very hot temperatures to put it in, kind of everyday speak.

we will be dealing with temperatures that are hotter than the sun. So, kind of hard to find a material that you can contain that in. There's not many materials you can make a bottle out of that would hang onto mass at those temperatures. So instead we create a magnetic bottle. So essentially using magnetic felines to confine a plasma because again it's a charged, it's a mixture of charged particles that can therefore all be confined

Molly Wood Hmm.

Kieran Furlong by these magnetic field lines. So you essentially, yeah, you build this magnetic cage to hang onto it.

And that keeps us separate from touching any of the, you know, kind of vessel materials that we would build the fusion energy system out of. So that's the magnetic confines. There's a variety of different approaches in magnetic confines. Some of your listeners will have noticed major news that was announced by the US Department of Energy at the end of 2022. And that was the first break even fusion reaction. And that was actually at a facility known as the National Ignition Facility, Lawrence Livermore National Lab in California.

Initial confined fusion basically is where you try to compress a fusion target.

So again, these heavier hydrogen isotopes in one space, and you compress them all so much that it heats up. The inertia keeps it all confined, right? It heats up before it can expand, if you think about how hot air expands and so on. So that's inertia-confined fusion. That was the first break-even reaction that was announced by DOE at the end of 2022. Big deal.

Molly Wood Right. And that was the first time, and this is why, you know, to reiterate, this is why this is hard. That's the first time that had ever happened.

Kieran Furlong It's really, this is first time it's ever happened in a controlled manner, right? it's really, it's really tricky to do, to hang on to. One of my colleagues and co-founders, Professor Kerry Forrest, always refers to this as, it's like trying to hang on to jello with a bunch of rubber bands. So that's really what you're kind of trying to do with a plasma.

Got it. And then where is Realta on that kind of continuum between breakeven and power plant?

Kieran Furlong Yeah.

Yes, exactly. and this is it. So, breakeven is something that really, different technologies will approach breakeven in different ways. That's not what we're currently focused on. What we're currently focused on is the jello and the rubber bands kind of step is can we actually stabilize our plasma? We know we'll be able to get to breakeven if we can just hang on to it for long enough and pour enough energy into it, right? So, where we are in realtor is we spun the company

out of a large RPE-funded experiment at the University of Wisconsin-Madison. It's what we call HWAM, W-H-A-M. It's an acronym. We can get into that later. Wisconsin HTS Axisymmetric Mirror. So there you go. There's nested acronyms in there. That's what we spun out of. So that experiment is now up and running. And REALTA is sponsoring research with the University of Wisconsin. We're sponsoring that project. We've got REALTA personnel working alongside the University of Wisconsin personnel on that.

as well. we are one of the small handful of fusion companies that, you know, we've an operating experiment. We've got fusion plasmas. We're, you know, kind of twiddling the knobs and, you know, making use of different actuators on the experiment to try and stabilize that plasma. So we're kind of moving quite far along the path in terms of it's no longer just, you know, kind of a nice CAD drawing and a physics simulation. We're moving into actual experimentation.

and plasma operation.

Molly Wood Right. There's a building we should, you for listener purposes, I had an earlier call with 

Kieran and your background is like basically what looks like a big rocket holding a bunch of plasma. Yeah.

Kieran Furlong Yes. Yeah, it's pretty cool. It's pretty steampunk. Yes, we've got this nice big shiny steel vessels, the vacuum vessels, that's where we'll have the plasma. And then we've got a variety of different ways of heating that plasma as well. And yes, I definitely encourage you, Monty, to come and visit it, should you be in the Madison area. We'd be happy to let you push the button and fire a plasma shot.

Molly Wood I mean, great, I'm booking my flight now. That's, I can't think of a better field trip offer, honestly.

Molly Wood OK, so there's, so when we talk about the kind of eventual reality of fusion, there's the science, there are multiple approaches, there's the ability to harvest energy, then there's deployment and commercialization. And you guys are taking kind of an interesting approach to that that I would love for you to tell me about.

Kieran Furlong Yes.

Kieran Furlong Yep.

Kieran Furlong Yes.

Kieran Furlong Yeah, so I mean in terms of there's a variety of our kind of different concepts out there right now. Myself, I've been through kind of the last clean tech boom and subsequent bust and been very involved in climate technologies over the course of my career.

it really is important to figure out that first plant, right? So I'm focused, absolutely focused on plant number one, right? I don't really worry too much about getting to the hundredth plant or thousandth plant or anything beyond that. Getting that first one built is super important. And that is one of the challenges for fusion, as we've talked about before, you know, a lot of the kind of experiments that are out there, they're classic big science, right? Massive, massive investment, huge amounts of capital, takes very long time to build these things and then to operate them and so on.

So that's a challenge for Fusion. We looked at that challenge and said, well, hey, we think we've got a better approach. We have got a technology that can operate, first of all, at a smaller scale than some of those.

It's a lot more robust. We feel we can tinker around with it a lot more so that we can get to what's the optimal operation of that plant. We believe when we have it operating, it's going to be easier to keep online so it'll have higher uptime, easier to maintain, the kind of things that industry cares about as opposed to trying to do kind of a trophy shot in a scientific experiment or something. So we're very focused on that. And part of that was, OK, what's the go to market?

Molly Wood Okay.

Kieran Furlong to a large regulated utility and saying, let's build a power plant with a first of a kind technology does not fill me with joy. Right. I just look at that going, well, hey, these regulated utilities, you know, many of them are conservative. Their business models are kind of, you know, they're baked in already. It's difficult for them to do something risky like that. When it comes to zero carbon electricity, they already have many other options out there as well. So we're looking at no, actually, let's leverage the

fact that we can do something at smaller scale. Now, it's still large, but we're looking at the kind of thing it might be a cogen facility for a large industrial site or campus or something like that. So ideally what we're looking at is a go-to-market strategy where we identify a large energy user and say, hey, let's build one of our plants right on your site, right next to you, and deliver heat and power.

are the combination of those that you're looking for. So I'm a chemical engineer by background and when I first kind of got involved with the scientific team, part of it was, okay, so we do the fusion reaction, the neutrons are captured in this thermal blanket that wraps around the fusion core, so we have heat. And then you take that heat, you put it through a conventional steam cycle to generate electricity, but you're losing a significant portion of the energy you started with there. How about we just look at who needs heat?

heat as a product, specifically zero carbon heat. So, you know, I mentioned chemical engineer. My very first posting in the chemical industry was actually down in Argentina on a large cracker. A cracker is a plant, basically a petrochemical complex. It's the front end where you will have...

Molly Wood Hmm. Mm-hmm.

Kieran Furlong feedstock based on crude oil or natural gas coming into the cracker and at the back you've got ethylene, the main product, which is one of the basic building blocks for all of our polymers and plastics and many of the petrochemicals that are in everyday products. That's a huge big furnace essentially, a big hotbox on the front end. uses a lot of energy that's as process heat. So I started looking at well, what are the use cases we could go after where you've got a need for heat.

but in industries where it's hard to decarbonize, right? And so can we provide an option there? So that was kind of one of the initial approaches that we are looking at. And really, you've got some very hard to decarbonize sectors, but they've got companies there who've made commitments to decarbonize and have zero carbon weather by 2030 or 2050. And trying to address that was one of the first things we've looked at. Also, these are companies who are used to large capex projects.

sophisticated engineers and project managers, an ideal partner for a start-up like us to build a first-of-a-kind plan to bring it to market. That's one use case.

Molly Wood Right.

Well, and I should say that, so keep going.

Kieran Furlong What's over the past year, we've also seen a huge increase in interest for data centers. So again, back to this concept of we can build a little smaller, right? We can deliver, we can basically say, hey, who's a large energy user at one site location? And it really helps if there's a user there who's looking for zero carbon energy. And that's a top priority for all of the hyperscalers. I'm from Ireland originally. Ireland is the country that sends the highest proportion of its

Molly Wood You read my mind.

Kieran Furlong electricity to data centers. It's currently over 20%. Singapore is the next closest at around 7%. Now, some people will look at that and worry about it and think it's a bad thing. In Ireland, it's a golden opportunity. You know, it's got ample wind energy capacity. It's really a way to turn wind into cash is one way of thinking about it. But the wind's not always going to blow, so you're definitely going to need to have firm, always-on, base-load energy to go alongside that.

Molly Wood Hmm. Mm-hmm.

Kieran Furlong and ideally that should be zero carbon as well. That's definitely what the hyperscalers are looking to do and I think they're being very prescient there. They're looking ahead and knowing, listen, if people are really going to allow us to continue to use more more energy for AI, for queries that folks are doing online and all of the other massive simulations and so on, kind that even we run, they know they need to earn their social license or maintain that social license with zero carbon.

energy and I think that's where it's very complementary to other forms of zero carbon electricity like wind and solar.

Molly Wood Voice-over: Time for a quick break. When we come back, we’ll talk a little more about the potential for creating data centers that are islands of power not bothering anyone around them and how long before we get there.

Molly Wood Voice-over: Welcome back to Everybody in the Pool. We’re talking with Kieran Furlong of Realta Fusion for our series on AI and energy use Feeding the Matrix.

Molly Wood That's one of the things I think is so compelling here. All of a sudden you have OpenAI and whatever it is that Elon Musk is doing, and they're talking about needing gigawatt scale energy. And you're saying we would like to build, we will be able to construct a plant next to one of these new data centers, decentralized, unrelated to grid energy, saying,

Kieran Furlong Yeah, yeah, I mean, you go back.

Molly Wood You're just going to operate as your own little island of energy generation and LLM output and everybody's happy and nobody dies as a result of you all needing your own special gigawatt.

Kieran Furlong Right, right. I think that's exactly it. So they're going to need to huge, you know, consume large quantities of energy. You go back five years, 50 megawatts was a big data center. Now data centers have been constructed around 200 megawatts. And I'm just was out in Singapore recently learned of a new one going in in southern Malaysia. That's 450 megawatts. And people are talking, beginning to talk now about gigawatt scale data centers. So yes, massive amounts of energy, but again, all concentrated in one location. And people will talk about increases in energy efficiency.

you know, better chip design that's designed for energy efficiency and so on. Well, yes, but that just will really open up more space for more data processing. And this is a key change that I think people need to make is if we can just, if we get fusion to work and if we can have fusion where it's reliable, it's cost effective, we'll be moving from a scarcity mindset for energy into an abundant mindset, right? So back to that example of Ireland, right? The amount of energy going to data centers, it's like, it almost shouldn't matter.

99 % just keep building more capacity because that's a great business for the country to have. So that's a critical thing is can we just unlock that abundant energy and we'll find all kinds of things to do with it. You know who'd have thought 10 years ago we'd be talking about energy for mining cryptocurrencies or now for you know the massive amounts of energy for AI for data centers. We'll come up with humanity I mean in some ways it's kind of like the endlessly

elastic market, we will find many many new things to do with energy if we have enough of it and it's cost effective. Yeah, I think that's just like, you know, a real good thing to work on. think it's something that gets us, it's all very motivating.

Molly Wood Yeah.

Molly Wood Yeah, that's...

Well, it's a really interesting, I wanna sort of just talk more about this go-to market because I think it is, it's an innovative way to think about the abundance mindset, right? The kind of obvious direction is to say, okay, well, if we're going to be producing energy, we should put it on the grid. It's a public good, like energy just go, you know, and so this, I wonder what led you to think, what if we just didn't do that?

What if we just buy, because I am fascinated by a direction toward bring your own energy, which I actually bring your own power, I tend to call it, which I think is gonna probably become increasingly common. But it takes a like, it takes a mindset to say, what if we did not do this the way that we've only ever thought about doing it? Kieran Furlong Yeah.

Kieran Furlong at you know.

Kieran Furlong Right.

Yeah, look, some of this comes from my own personal history. One of your prior guests was talking about biogas generation. And I spent some time in that world as well. And in fact, I worked on a project where we were aiming to, within a start-up, much earlier in my career, we were aiming to gather together all of the food waste in the Bay Area, you know, where the population there are pretty well trained in terms of separation of waste and everything at this point. We gather together all of the food waste, bring that to the large wastewater

treatment plant that's just across the Bay Bridge from San Francisco in Oakland and look at using all of that food waste instead of having it rotting in a landfill generating methane which is an even more potent greenhouse gas instead you'd have that in a controlled environment capture the methane the biogas burn that generate electricity and plug it into the grid

plugging into the grid was going to be a 10-year chunk of the Gantt chart, right? So you look at getting that project to market and that's just, well, some of it might be because of the bureaucracy and the arcane nature of regulated utilities, but I won't pretend.

Molly Wood Right. Plus California, will stipulate.

Kieran Furlong Plus California, yes, okay, plus California. But I won't pretend that it's all just dysfunction. It's actually, operating a grid is pretty complex, right? All of the grid balancing and so on that needs to go on. So there's reasons why interconnect takes time, takes a lot of engineering, a lot of study before you can plug something in. So we just did not want to have that grid interconnect and the publicly regulated utility that may have its hands tied around what they can and can't.

do as our first customer. That just seemed to be adding more more risk on top of the technology and execution risks that we're already addressing within our company. So instead we're looking at, how can we just set that whole grid connection problem to one side? We'll still like to do that. At some point in the future, we'd love to be doing grid scale electricity generation plants as well with our technology, but not as the initial go to market. And that's what matters for me and it matters for a start-up.

gone through this before, it's like getting plant number one financed and built is going to be probably the highest mountain we have to climb. And that's higher than all of the, you know, kind of technology mountains that we're climbing already.

Molly Wood I wonder, so then that does raise the kind of question of financing. I suppose that if your goal is connect to the grid, provide a public good, you know, that people pay for and there's a profit motive, you are still saying, hyperscalers, we are trusting that you have committed to your sustainability goals enough, or that public pressure will demand that you do this in a clean way that you will also add on many millions of dollars.

Kieran Furlong Yes.

Kieran Furlong Yeah.

Kieran Furlong Right.

Molly Wood to your CAPEX budget for your new data center to co-locate a fusion energy plant next to it.

Kieran Furlong Right.

Yeah, but look, I've been through this before in many other start-ups, right? You you can't rely on a green premium, right? It might be enough to get something started, but you have to be able to show a line of sight to cost competitive energy. And we think we have that. So we've done, you know, pretty robust techno-economic analysis ourselves. Obviously, we're basing that on lots of assumptions, but we believe they're reasonable engineering and scientific assumptions at this point, that we can point to cost competitive electricity, cost competitive

Molly Wood Mm-hmm.

Molly Wood Right.

Kieran Furlong heat on future plant bills. And the first ones are...

Molly Wood Okay, so you're saying to them you'll get zero carbon energy, you will also get a payback period. That's reasonable. Yeah.

Kieran Furlong Right, at some point in the future, exactly. And it might not be for that first plant, but the first plant needs to be built at URISTA Technology. And that's where you look at, okay.

Where do we get the financing to defray that risk on the first plant? Is it direct public financing, know, non-dilutive contribution from government agencies? Well, that's a possibility. Is it actually no? There's a of a willingness to invest from the hyperscalers, let's say, because they want to basically more indirectly earn their social license to continue doing what they're doing. There's a variety of different ways that you can see that the initial investment

that needs us to get us down the curve to where we can be cost competitive will happen. this is a...

bone of contention when you get into discussions with folks with different sides of the, when we're looking at energy as more of the conventional fossil fuel side and folks who are pushing the alternative energy agenda, I really think we have to just look at, no, we need to get to cost competitive energy, ideally zero carbon energy, how do we do that and how do we actually all together collectively look at what's the investment needed to get us there? And then we can figure out which channels does that

investment come through.

Molly Wood Right. Well, and this is the abundance seems to be a powerful part of that argument here too, right? If you're saying like, look, fusion energy is a really good bang for the buck, right? It's abundant, zero carbon energy that will pay itself back can be deployed anywhere. Like you already see these hyperscalers making, by the way, when we're saying hyperscalers, we're talking about like Microsoft and.

Kieran Furlong Right. Yep.

can be deployed. Yep.

Kieran Furlong Right.

Molly Wood right, Amazon and we're trying not to name names, but companies who have an interest.

Kieran Furlong But every day names the Silicon Valley, well, I guess Washington, giants who are out there building massive data centers or utilizing massive data centers being built and operated by others. Yeah.

Molly Wood Yeah.

Molly Wood And what they know, and you see the really forward thinking ones making energy investments. think OpenAI and Sam Altman, Sam Altman has made a bunch of bets in fusion and also nuclear fission and solar. They're all, in a way it feels like the AI boom might be a really great thing that happened just in time for your business.

Kieran Furlong Yep. Yep.

Kieran Furlong It might, exactly. The two things can kind of leverage each other, right? So to get it up and running because I think that's the other thing is if people are thinking either are, right? Well, no, we should just be building gas peaker plants and using natural gas or no, we need to be just focusing in on nuclear fission. Energy is a classic example of where we need all of the above, right? And energy transitions take time. Daniel Jurgen, kind of very well known author and, you know, kind of knowledgeable person in this space lays those out or I'm trying to think the other guy, Vaclav Smil, who talks about the history of energy, energy and civilization. They point out that these things take a long time. mean, people think, yeah, we've moved away from the era of firewood and that was, you know, 17th century and so on. It's like, not in the largest parts of the world and we're still burning a lot of biomass for energy as well. So you end up with trying to do a lot of these things. I mean, it might be firewood for a stove in a home in Africa or it could be

SAF, you know, jet fuel from biomass in, you know, Boeing 737. So there's, we're still burning biomass in a variety of different ways as well. So it will take time to transition, but we absolutely need new sources of energy. It's not a matter of, well, let's close the door on this one over here, let's open this one. We need it all for a while. We've got 10 billion people are going to be with us by 2050. Already we're at what, 8 billion?

We've maybe two billion who are living the quality of life that we enjoy here in this country.

some consuming a little bit less energy for that quality of life, some consuming more. energy is what underpins, you know, the food we eat, the clean water that we can drink, the materials that make our clothing and our homes and everything else, right? So energy is just critically important and we need more of it. We need an awful lot more of it. And if the AI boom is one of the things that can help, you know, scale new forms of energy, like fusion, fantastic, right?

Kieran Furlong gets it up and running and helps us get down those cost curves to where we'll be cost competitive and become the choice. Just as we've seen with solar and wind. In many cases they now become the default for new power generation, electrical power generation because they are the most cost effective way to do it in certain parts of the world.

Molly Wood Yeah. What does your timeline, taking out interconnection, what does your timeline look like?

Kieran Furlong Yeah.

Yeah, so obviously as I mentioned we've got a fusion experiment up and running right now that started this past summer in July and that will be continuing to operate that.

to demonstrate how we can get to that plasma stability, how we can structure the rubber bands just right to hang on to the ball of Jello. And then we are currently designing the next step up facility from that, which essentially will be a larger version of the experiment at the University of Wisconsin. We call that ANVIL. We aim to have that online by the end of this decade, so 2029 or so. That actually will be the last scale up step that we need for the critical components, things

like the HTS, the high temperature superconducting magnets, the plasma heating systems and so on. Once we demonstrate that we get to the right plasma conditions in that vessel, or that device that we're calling Anvil, then we will go and build a tandem version of that. So two of those together, connected by a longer central vessel. That is what we'll be calling Hammer, and that's our first net energy producing plant.

commercial plant and that will be in the mid 2030s.

Kieran Furlong So that's the timeline for us. It's going to take some time, right? There's a lot of work that needs to be done. But that's not out of the question. Ten years probably sounds like a long time for folks in the software industry, but if you're talking to people who are doing, you know, drilling for new oil wells off the Arctic shelf or constructing a new cracker or whatever else, ten years is a reasonable project timeline. We have an awful lot to do in that time to make that, but that's the goal that we're setting.

Molly Wood Mm-hmm. Right.

Kieran Furlong if we push things off further and we start saying well you know let's just be safe let's talk 2040 or beyond people aren't going to pay attention it's going to be too late for many folks so we need to have this sense of urgency around every day coming in like we need to solve this problem as soon as we can

Molly Wood Yep, amazing. Karen, thank you so much. Keep an eye on the space. It's fascinating. Appreciate the time.

Kieran Furlong Well, thank you very much. Thank you. I really appreciate having the opportunity to speak to you. Thanks very much, Molly.

Molly Wood Voice-Over:

That's it for this episode of Everybody in the Pool. Thank you so much for listening.

There’s one more official episode to go in the Feeding the Matrix series although you can imagine this is a topic that will keep coming up

Next week we’re going to imagine not just how you power a data center  but how you rethink these facilities COMLETELY  to be buildings that blend in to their surroundings that don’t just hog resources they actually give back to the earth and the communities around them

A project called regenerative data center design that actually originated with Microsoft.

Email me your thoughts and suggestions to in at everybody in the pool dot com and find all the latest episodes and more at everybody in the pool dot com, the website. The newsletter is BACK find that at Everybody in the Pool dot com or read all the back issues at mollywood dot co.

Thanks everyone for your support and for listening it might not always feel like it but together we CAN get this done. See you next week.