Unleashing Isotope Abundance with Atomic Alchemy’s Thomas Eiden

What’s the cause of these persistent and chronic shortages of medical isotopes?

The United States consumes about 50% of the world’s nuclear medicine every year, but we effectively make a tiny fraction of the supply of radioactive material that goes into that nuclear medicine. It’s primarily sourced from about 6 reactors around the world, but all those reactors are getting old. I spent the early part of my career at Idaho National Laboratory working on reactor core design with the Advanced Test Reactor, which was about as old as the reactors making medical isotopes. I knew from first-hand experience that when a custom part breaks in these very complex scientific machines it could shut it down for months for repairs. So I knew that these machines could do amazing things, but at their advanced age, there’s no way they should be underpinning the entire supply chain for nuclear medicine. 

The first problem is with the access to manufacturing capacity because the reactors that are able to make these materials are in short supply. The other issue is that the supply chain is really fragmented. When you take your material out of the reactor it’s screaming hot, radioactively speaking. So you have to put them in some very special containers that are not cheap to license and then you have to ship them to a chemical processing facility where you’re extracting all that radioactive material from the feedstock at a high purity, especially for nuclear medicine applications. Then once you have your isolated and purified material you ship it off to a drug manufacturer to make the drug. At that point, it goes to the hospital. So you’re shipping this radioactive material all over the world and in nuclear medicine many of these materials have really short half-lives so producers sometimes have to produce up to 30% more than is actually ordered to make sure that the hospitals get the quantity that they need. It’s a race against time. 

We’re going to fix that problem because that’s a lot of risk in the supply chain. Our facility will vertically integrate as much of that supply chain as possible so that you don’t have to ship things all over the place and we have more control over the discrete steps. You won’t have all these middlemen brokering things between like five different commercial companies. We can be a one-stop shop that streamlines the whole process and we can make these materials less expensive and more efficient than the current status quo.

How does your reactor differ from a reactor that is used for power generation? ‍

Generally speaking, the way radioisotopes are made is you take some boring, non-radioactive material, place it into a little aluminum capsule, weld the top on it, and plunk it into a reactor. Then you bombard that boring material in the reactor with neutrons that are created from the nuclear fission chain reaction that keeps that reactor going. You’re making a certain number of excess neutrons that hopefully get absorbed by the boring material that you place in the reactor and turn it into a not-boring radioactive material with a very valuable use case. 

But you need to build a special type of reactor to make these materials. I like to say we’re the SpaceX of nuclear medicine because our reactors are a lot like their rockets in the sense that a Falcon 9 is not used in place of a Starship and vice versa. Rockets are designed for certain payloads, orbits, and so on. Reactors are very similar. You can put these types of capsules into a power reactor or a research reactor because you’ll still get those excess neutrons, but that doesn’t mean you can profitably make isotopes. The Advanced Test Reactor at Idaho National Laboratory, for example, takes too long to shut down and depressurize so it can be opened up to take out materials for most nuclear medicine use cases. And the density of neutrons whizzing around in a power reactor is much lower because the core is so much larger so you can’t cook these materials as efficiently. So the design plays a very important role in how cost-effective it is to do this. 

What we’ve done at Atomic Alchemy is design a reactor that can make these materials as cost-effectively as possible. We wanted it to be very simple and very low OPEX compared to a large research reactor. But it also had to be very power-dense. A power reactor may be putting out several hundred times more megawatts of thermal power but it, and therefore those neutrons, are spread out over a couple hundred fuel assemblies in a huge reactor core, but our entire reactor, which you can almost put your arms around, puts out 15MW in an extremely small volume. So the density of those neutrons is very large and we can cook this stuff very effectively and make a much more pure end product.  

Did you design your reactor to be isotope agnostic or do you have to adjust the design for different types of isotope production? 

The reactor is called the Versatile Isotope Production Reactor because, well, it’s pretty versatile. We can make 40+ materials in the reactor at the same time. 

Any other reactor that’s making neutrons can do more than one material at the same time too, but there are potential issues with using other types of reactors. In the case of power reactors, you’ve got this issue where its main mission is electrical production so you have to be careful of what you put in there because if your material starts soaking up all the neutrons that’s going to reduce your fuel cycle and limit how much electricity you can make before you have to refuel. The issue with government-funded research reactors that supply radioisotopes today is that they’ve got some very specific missions that are being budgeted for and it’s not their main mission to make isotopes. They’re not a commercial enterprise so if they take a little bit longer than you’d like to make your isotopes you’ve just got to be patient and hope you get your materials when you need them. 

What makes your design different from existing reactors used to produce isotopes?

We designed it to be the most cost-effective thing that we could build as quickly as possible. The reason build speed is important here is that the operating licenses for five of the six research reactors abroad that make the bulk of the materials for nuclear medicine expire by 2030. So we wanted to be able to build these things in the next couple of years. But we also wanted to make sure that we had very low technology risk. 95% of the components that make up the reactor are commercially available off-the-shelf in the sense that they’re used in some other reactor system already, which means we already know how they perform. We know their upper set points, what their temperature limits are, how much abuse they can take because radiation is a very punishing environment, and so on. It also means the regulator is very familiar with the technology and we’re just arranging it in a more optimal way for isotope production. So we eliminate a lot of the regulatory risk as well. 

What we hope to do then is make it almost a magical experience for anybody who wants to buy this material. Whereas today you have to have a reactor expert on your team to interface with the reactor folks to try to figure out how to get this stuff into their reactor, our process will be as simple as an order form where you send us the specifications of the material. Our business is to make this stuff and so we want customers to be able to procure it as quickly as possible. As somebody who used to do 100-page analysis reports every 60 days to put stuff into a very fancy research reactor I want to eliminate that experience. 

A shortage of medical isotopes is a huge problem that can cost lives. Why aren’t we building more of these types of reactors domestically? 

That is a very multifaceted question. It’s so obvious to me what needs to be done and yet no one is doing it. If people aren’t familiar with the space, they’ll just assume that if it’s such a big problem and no one is doing anything about it then the government will solve it for you. I can tell you the government is not coming to save us on this issue. There’s infrastructure that we need that won’t exist unless someone decides to build it. The good news is right now is probably the best time to design and build reactors. Part of the reason for that is that the cost of compute has dropped so much over the past 10 to 15 years. When we do reactor core design, we use neutronic modeling codes that we parallelize the crap out of and we can soak up a thousand cores of computing power to run a reactor simulation for a month. If I wanted to do what we were doing 15 years ago, I probably would have had to partner up with a national lab that has a supercomputer to do things as quickly as we are. So now we can iterate at the speed of ourselves because we don’t have that constraint on computing power. 

Another reason no one is doing this right now is that the people who are most impacted by it are downstream. There are folks that are doctors of nuclear medicine who have tried to build reactors, but they didn’t know enough about the fundamental reactor technology to make these things. That meant they had to contract a lot of things out and their budgets blew up. It’s very expensive to do the things that we’re doing in-house if you’re using contractors. At Atomic Alchemy we understand the nuclear medicine problem even if it’s not at the same level as a doctor, but we do know how to make radioactive material so we can optimize the infrastructure and process side to do it as cost-effectively as possible. 

The last big reason I think this hasn’t been done yet is the perception that nuclear is very hard, very expensive, and extremely regulated. And those are all kind of true. But one advantage that we have as a company is that we sat down and combed through the regulations to understand what we really need to do to put together a license application to build this stuff. We’ve had plenty of people tell us how it’s always been done in the past, but that doesn’t lend itself to innovation. So we were able to craft a very modest 700-page construction permit application because we followed the letter of the law rather than guidance from other folks who did it 20 years ago for large power reactors that are so much different than ours. So we’ve been able to get cost estimates to build these things that for most people are unbelievable because of this persistent perception that building a reactor like this must cost a lot. 

To give you a sense, last year the University of Missouri announced it was going to update its research reactor, which first came online in the early 60s and they understand there’s this really lucrative medical isotope market out there so it probably wouldn’t hurt to update their reactor. When they unveiled the plan at a conference last year, someone aptly asked how much it was going to cost. The gentleman at the podium said it would probably cost about $1 billion. Everyone around me—a bunch of incredibly smart folks who’ve worked at national labs all their lives—said that sounds about right. I was sitting there wondering if I was missing something because our cost estimate for building our reactor is about a tenth of that for four reactors. I’m really excited for when we build our facility and I can show the rest of the industry that you don’t have to spend a billion dollars to replace a 10MW research reactor.  

In general, I think people in this space defer too much to authority. They assume that because all these really smart people have been doing things a certain way for 30 years, who am I to say they’re wrong? Well, the other side of that is who has built a reactor in the last 30 years? We have the two Vogtle reactors that just went online, but those got started over 10 years ago. Not a lot of people in the industry today can even say they’ve built reactors. I think people just need to do their own thing and it may sound surprising but it’s the best time to be doing this right now.

You just announced a collaboration with Oklo. How will your two companies be working together?

As a fast reactor company, Oklo is also working on reprocessing used nuclear fuel—basically recycling nuclear waste. And what’s really fascinating is that there are a lot of radioisotopes in the waste left over from the recycling process that are actually very valuable to what we’re doing at Atomic Alchemy. A lot of the initial collaboration was to look at the unit economics of pulling certain materials out of their waste stream, but the end goal is to be able to collaborate on a facility where I can take their waste and pull out the stuff that I can use for many of my commercial segments so that what’s left over is a true nuclear waste with absolutely no commercial value. There’s also just much less material left over and it’s thermally cooler because of the materials we’ve pulled out so it’s easier to store. 

In the long term, our reactors are in the thermal spectrum, meaning we use slow neutrons, which have similar physics to a lot of light water reactors used in nuclear power plants. The physics of Oklo’s reactors are very different. They use fast neutrons without slowing them down because the type of fuel they’re using lends to just a different way of doing the fission. But sometimes there are some isotopes that are very interesting, especially for nuclear medicine space, where you have a simplified production pathway if you have access to those fast neutrons. But there aren’t many fast reactors in the world and the ones that do exist are government research reactors. Most of them are in Russia and we won’t work with them anyway. So what this collaboration opens up is the potential for a small subset of isotope production that is uniquely tailored around Oklo’s fast neutrons that most other folks don’t have access to. That’s really exciting because we can make some isotopes even more efficiently with access to fast spectrum neutrons. 

You’re looking to break ground on your first facility soon. What does the timeline look like?

We are looking to try to be up and running between mid-2027 and early 2028. A lot of that’s driven by the construction firm’s timeline and you can expect there to be some regulatory slop. But in any case, we should be able to start breaking ground in about 18 months with pre-construction activities. In terms of legal nuclear construction, our 700-page application will need to be approved first, but a lot of this can be done in parallel. Once we get legal permission to construct the actual facility that’s going to take about 21 months. Then we’ll have a short commissioning time frame to test the reactors and bring them up to production capacity. We’re running a tight ship and I expect those dates to hold firm. 

How does the world look different in the future if Atomic Alchemy is successful?

The byword is “abundance” right now. I’m a member of the Society of Nuclear Medicine and I get notifications when there are supply shortages. It’s 2024. Why is the word “shortage” in our lexicon? That’s ridiculous. This isn’t the 1950s when there were just a small handful of reactors controlled by some military or government entity. We want to make the word shortage disappear forever. Once we begin providing access to these materials we can satiate the demand for well-established use cases, but there are also interesting new use cases we can explore more. For example, space batteries. Right now, NASA can maybe launch a deep space mission maybe once or twice a decade because of the shortage of these materials. If NASA can launch all of the deep space missions its budget permits, what does that mean for commercial space in the sense that now they have access to these same materials? Maybe that allows for a world where we have Starlink around Mars without needing hyper-efficient solar panels.