What if Russia stopped selling uranium to the US tomorrow | Scott Nolan, General Matter

During college I was an intern at Boeing on what seemed like the coolest project you could do, which was a government military project at a top secret location. I actually got feedback from coworkers who said, you're doing too much, you're going too fast. They literally said, we have to make this project last. On day one of joining SpaceX, we had an employee handbook. And I remember the first page of the employee handbook in big bold letters said, this is not a science experiment. This is an engineering problem. And it was like night and day. Congress passed a law banning the import of Russian uranium starting on January 1, 2028. We need to be online as quickly as we can so that there's viable source to really offset that loss of US supply. Our goal is not just bring back US capability, it's also make it more scalable so we can power all the things the US wants to do.

0:00

Today I have the pleasure of sitting down with Scott Nolan. He is the co founder and CEO of General Matter. Scott, I think today Russia supplies roughly 25% of all the enriched uranium to the US and I want to know what would happen if Russia decided to just stop supplying the US Tomorrow.

0:43

If Russia stopped shipments of uranium to the US Tomorrow, utilities would start eating into their inventories. We'd continue importing from Europe, but soon we would have to figure out how to replace that. We would probably talk to the European suppliers about purchasing more. Maybe there would be a lot of pressure to start even thinking about purchasing from China, who's one of the major producers. And that would be the solution. We'd be buying more probably from Europe, and then maybe Europe's buying more from China and then maybe China's buying more from Russia. So we'd be kind of bringing in through that channel. And at some point maybe it would be too hard to purchase or prices would go up a lot and utility prices would go up a lot, and maybe there'd be occasional brownout. So I think it would be a scenario where you truly have a shortage of about 20, 25% of uranium supply onto the grid. And we would be scrambling to figure out where that's going to come from. And probably for some period of time, rates would go up a lot.

1:00

And uranium isn't really like an elastic supply. Right?

2:00

There's a lag time on it, which is why people carry inventory. But the whole supply chain of enriched uranium, if you go end to end, is really, you start with mining. And so mining ramping up, mining has major lag time on it. Then you convert it into a gas. Those facilities exist, they're not that flexible. They're basically at capacity right now. And then enrichment is this other capacity piece where you need a large factory running enrichment equipment, enriching and depleting uranium at the same time. And then you're turning that into fuel. So there's a lot of fixed capacity in the supply chain. It's not something that's easily just ramped up overnight. So I think it would be a race against time to actually bring new capacity online, whether that's inside the U.S. in European allies with a fallback plan of finding someone who can make it. And at some point that becomes China. And maybe I should add, the reason I think you're asking this question is because in 2028, that becomes reality under current law.

2:04

What do you mean by that?

3:06

Congress passed a law banning the import of Russian uranium starting on January 1, 2028. And currently we're under a waiver process where utilities can say, I don't have a good source of this uranium, this enriched uranium, or this fuel other than Russia. And so we do need to import this to make sure that we can keep the lights on. And so that's the waiver process that's going on now. On January 1, 2028, that path, that waiver process actually expires. And so as written in the law today, that is the deadline.

3:08

Would you expect the US to kind of change their stance ahead of that, to basically make sure that that doesn't just immediately get turned off overnight?

3:41

I mean, this is why we're doing what we're doing. So the Russia uranium ban was a bipartisan act, extremely, extremely supported in Congress, very consensus. I don't think it's something that the US wants to back down on. And so the reason we started as a company was to address the supply chain issues that the US Is facing from low enriched uranium that powers today's grid, which is 18.6 to 19% nuclear. The advanced fuel that all advanced reactors need called Hailu high assay, low enriched uranium that we can't get at all today. So every advanced reactor company that you've heard of is, is looking for this fuel and trying to find a long term supply. And so it's this, this inability to expand the grid, to expand nuclear because we don't have the, the missing piece of enrichment is why we're doing what we're doing. And whether Russia's in the market or not, this is a major issue that we're trying to address.

3:48

I'd love to talk about the history of uranium enrichment. I think back in the early 2000s, space and actually getting astronauts to orbit was one of the last things that Russia was only able to do and the US didn't really. We were shutting down our shuttle program and then SpaceX came in and now that's kind of flipped. How did the world kind of unfold where Russia is actually supplying us with this very critical material?

4:46

It's an interesting history. So the US was first to do enrichment. We did it during the Manhattan Project. We did it for nuclear weapons to win World War II. And that process, that was done using a process called gaseous diffusion, it was a first generation technology. The US did it at scale, did it across a few sites. Oak Ridge, Tennessee, Paducah, Kentucky, Portsmouth, Ohio, three places that we did enrichment for decades and that enrichment was later. It was used for weapons originally, but very quickly then used for nuclear power. The US was a leader in enrichment. We were doing it 85 to 90% of worldwide enrichment was the US using this gaseous diffusion process. That was something we thought was really important during the Cold War to have our own supply chain. Later, after the US got this capacity online, Europe came up with their own enrichment. Russia came up with their own enrichment capacity. These were with a second generation technology, the gas centrifuge. And so these two paths kind of diverged. The US kept doing gaseous diffusion. Europe and Russia did centrifuge. The US kept it going until the fall of the Berlin Wall. And at that point, the US said, hey, we're clearly now in a free trade world. We don't need to do this gaseous diffusion process anymore. The process that Europe and Russia went with is more efficient. We can trade for this. This is an economic decision, it's not a national security decision as much. We have a huge stockpile, we have plenty of weapons. We do not need more of that. Let's focus on nuclear energy. And for nuclear energy, we don't really need to have our own enrichment capacity. This old process is a lot more expensive, it's more energy intensive. Let's work with our foreign allies to trade. And so over the next few decades, post the Berlin Wall, the US progressively shut down its gaseous diffusion capacity to the point of in 2013, it shut down the last plant that was owned and operated by the US in Paducah, Kentucky. And we said, we'll keep trading. And so over the last decade, it's become more obvious that that probably wasn't the right approach.

5:11

Is there a reason that the US Kind of made that decision in the first place. It seems like a very strategic thing to have here in our borders. How did that happen?

7:28

I think it's like a lot of industries where we said we're in a global free trade world and we're, we're allies with Europe, we're on good terms with Russia, we're trade partners with China, freely trading with them for the last decade. You think of all the way back to the Clinton administration, this seemed like the consensus. China's a growing economy, we can move manufacturing there, we can do other things here and then free trade. So I think we just undervalued China. The importance of having our own capacity, whether it's manufacturing or critical minerals or things like the nuclear supply chain. On one level, it's purely a domestic US thing, like how do we make sure we have our capability. But the second tier of it is really global influence. So today, because we can't produce fuel for reactors unilaterally within the U.S. russia and China are really driving the vast majority of reactors that are deployed elsewhere. So north of 70% of reactors internationally, I think are Russian and Chinese designs because they can come with a fuel supply.

7:36

One of the things that really surprised me when I was researching for this interview was apparently the US actually has the most reactors out of any country right now. I think it's like 94 and China is at something like 59. But China is turning on like 28 to 30 right now, like they're under construction, whereas the US maybe has two. And so they are just like solar, just going to absolutely destroy us. If the current trends persist. How do you think we kind of stop and buck those trends and get back to building nuclear?

8:52

Yeah. Even at a bigger level, if you look at China's grid in 2010, we were tied completely neck and neck, roughly 4,000 terawatt hours. Since then, they've doubled their grid, almost tripled it, while we've stayed flat. And so this issue of capacity and power production, it's upstream of everything. And it's like, yeah, all the things that you hear about, probably half the companies that you'll talk to are doing something that has a major energy input, especially AI. So AI, data centers, manufacturing driven companies, defense, all of these things require a solid grid and good baseload. And, and China now has 2x what we have. This is something we have to fix as quickly as possible. So this is why we're working on this. It's really upstream of everything. The ability to enrich is upstream of making nuclear fuel, making Nuclear fuel is the required input for every reactor. And it seems like the only clean, it seems like we're going to go with. For the grid to expand, it needs to be clean baseload. And I think the only form of clean baseload that we have today that's economical is nuclear. It's always been the cleanest, it's always been the safest baseload. So we see that as being the, the way that the grid expands through existing reactors, advanced reactors, but without the fuel, you can't do that.

9:22

This is kind of a hypothetical, but let's say you were given like $25 billion today and just said go turn on as many reactors as possible. How would you actually do that?

10:44

The administration is doing a version of this right now with AP1000s. That's the reactor that we have that really moves the needle on the grid.

10:54

And what is that?

11:02

So AP1000 is roughly 1 gigawatt reactor, 1 gigawatt of electricity production. And so you look at a lot of these data centers and they used to be a few hundred megawatts, now they're gigawatt scale. So. So in order to really move the needle quickly on approved reactor designs, you're going to have to do something with a 1 gigawatt class reactor. And that's what you see the administration doing with deals around their deployment. In the next few years, we expect there to be a bunch of approved next generation reactor designs. Pretty much all the startups you've heard of are doing this type of design. A lot of people call them SMRs, small modular reactors. There's also micro reactors in there. And so the big question people have or the big thought behind this, the big idea is how can we take something that was historically a huge construction project and turn it into a factory build that can be done at scale. And so almost all of these companies have the premise of we're going to take a civil works construction project and we're going to turn into a factory build that's modular, small modular reactors. We're going to take those modules and we'll take them to the site and then we'll connect everything and get it online without the same 10 year construction effort that's required historically in the US when you do that, you end up needing to make your reactor smaller. And if you make your reactor smaller, your core smaller, and if your core is smaller to get good nuclear reactions, to get good burn up, good efficiency, you end up wanting to run more enriched fuel.

11:04

So that's like the difference between like 3 or 5% and like 20%.

12:40

That's right, yeah. Yeah. It's typically 3 to 5% or 19.75% is where most reactors are going.

12:42

And what is the difference in the process for, like, enriching uranium if you're just doing like 3 or 5% versus, you know, 19.75 or 20?

12:48

I would think of it like a refining process or a distillery. You know, even a distillation process. It's just like refining another time. So most enrichment methods are a series of steps to just further and further and further refined material. And so all you do is further refine it some more. So when uranium comes out of the ground, it's about 0.711% U235. And so that's the fissile material that you want. And that is elevated to 3 to 5% for the reactors on the grid today. So it's refined until you get to that threshold for advanced reactors, that's taken up to 19.75. But the process is not fundamentally different to do that.

12:55

And is this something where, like, tomorrow, if there was a small, you know, SMR that could be turned on, does the fuel even exist right now to. To make that, like, work? Or is this something that has to be kind of scaled in parallel?

13:37

Yeah, not in the West. So the. Yeah. If we go back to the whole origins of general matter, it was, it was around this question. To answer your question directly, of if. If there were $25 billion, what would I do? I would probably, you know, if you want to get started today, you would deploy some large reactors, you'd deploy some AP1000s. There's a huge wave of advanced reactors coming through. Many startups, some incumbents. These are very promising. It is likely the path to bring down the cost of nuclear and to scale it rapidly. And so I would save some of that money for when they are licensed and they are ready to deploy and deploy some of those. You might do a mix of some microreactors for some environments, for some remote applications, or if they can get their costs low enough, you can tile them together and then other SMRs for applications like data centers and behind the meter.

13:50

When we were talking before we started recording, you mentioned that basically with Founders Fund, the way that you guys decide or you decided to start this company is almost under duress. Like, you ideally are trying to go find someone else that's doing it and go fund them. But if something is super critical and no one else is doing it, then maybe you're the Right person to do it and you go do that. How did you kind of make that decision?

14:42

Yeah, so the origin story was really at Founders Fund, I was looking at many different engineering companies, everything from transportation to space to nuclear, and so met a number of nuclear companies, other energy companies over the years. By the late early 2000s, there were a bunch of interesting advanced reactor companies, and Founders Fund invested in one of them. And I was talking to them about the path forward, and the prior on this was that the hardest thing would be getting licensed, because for a long time no one had gotten a license on a new reactor design. And they said, yeah, there's something there that's not trivial, but there's a known process and we're going to do that process. It's just that no one's really tried to do that or no one's done it well, and we're going to do it well. That was what the reactor company told me. And then they said the much bigger challenge besides getting a license is actually getting the fuel. And so I started looking into this and trying to understand, what do you mean, why the fuel is so hard to get? Clearly, we have reactors on the grid today that are running on fuel. Is it that much harder to make Hailu, the more advanced fuel that you need, than to make LEU, the 3 to 5% enriched fuel that all reactors run on? And I said, why don't the companies in the US that make that fuel make Hailu? And they said, there aren't. What do you mean, there is no company? Yeah, there are no companies that make it in the US we completely import it from Europe and from Russia. And so I said, well, okay, where are you going to get Haylou? And they said, there's one place, it's Russia. We have to import it. And so combined with the Russian, this was already challenging for them. But then combined with the Russia ban, it was extremely difficult.

15:07

When you initially get that information, what's going through your mind on this? This clearly doesn't work. Like there's this massive roadblock that's going to happen that you can see is going to happen. And for some reason, people are just kind of like saying, you know, we'll figure it out.

16:51

Well, I think it wasn't something people were worried about or it wasn't relevant for a long time. So if you're a utility operator, you have a bunch of nuclear reactors, you know, you're purchasing fuel from Europeans or enrichment from European firms, and you know, you're importing some from Russia because it's been historically very low cost and it hasn't really been a problem. The grid hasn't been expanding. You keep your contracts going. The utility fuel buying teams purchase where they can get it from. And there hasn't really been the impetus to create new enrichment capacity in the US as reactor developers, like all the startups working on reactors, I think they have been the first to really encounter that the supply chain in the US is not where it needs to be. I think utilities may have been aware of this, but. But it's not within their scope and so they'll buy from where they need to purchase from. New reactor companies kind of discover this as they go and realize that, okay, it's time to start thinking about our supply chain now that we have a design. And then you realize this really weak part of the supply chain. So the Founders Fund story was they were telling me about this issue and then I spent basically all of 2023 trying to find a company that was solving this and working on this. And after an extremely thorough search, understanding all elements of the supply chain from mining to conversion to enrichment to fuel fabrication, they were right, it was really an enrichment issue. And so spent months trying to find the right company that was doing enrichment, realized it didn't exist, realized that like in the past, like with Palantir and Anduril, a new company needed to be created. And so started pulling, thinking about how do we pull together that company in 2023 and then officially incorporated, officially launched at the start of 2024 with a core team. And so to your question of under duress, yeah, it's not something that Founders Fund does programmatically or even desires to do, but it's something that when there's an opportunity that no one is solving, it's been the case that people on the team find it, find it worthwhile to put everything else aside and go start something. So we started this with the desire to bring back domestic enrichment to make it much more scalable, much more reliable, it's already extremely safe. And then to bring down the cost that we can power. Everything the US Is trying to do that's going to require a massive increase in the grid or behind the meter power production.

17:04

I was reading this doc that you gave me earlier this morning and you basically said most of the time when you're talking to founders, you're almost trying to convince people to not start companies because they're not necessarily doing it for the right reasons. For you yourself, this is like a very long journey. How did you kind of come to the conclusion that you wanted to spend the next maybe decades of your life working on it.

19:27

Yeah, I think, yeah. To address that point of like trying to convince people not to start companies. Yeah. I'm certainly not the first person to say this. I think there's been allusions to this by Peter at Founders Fund has long said the average, like the hundredth person at Google made way more money, to put it in financial terms, than your average founder in Silicon Valley. So if people are trying to start companies for financial reward, it's often much better to just join a great company. And I think a year or two ago Jensen said this also someone asked him, if you were going to do it all over again, what would you do differently? He said, don't do it. Because yeah, I think starting companies is hard. And I kind of knew what I was getting into after 12 years of working with different founders at founders fund and three, four years working at SpaceX, very early on, it's something you should do only if you have to do it. So my criteria, a lot of people would ask me during my decade plus at Founders Fund, would you ever go start a company? I had a pretty standard answer which was, yeah, I would do it under three conditions. If there's a really important problem that's not getting solved and somehow I have a good shot of helping solve it, then I'll do something with this one. Through talking to reactor companies, it became obvious this important problem of actually making the fuel that they needed was not getting solved. So that checked number one and two. And then when I really dug into it during 2023, it became obvious you would need a few different things that had to come together for this to work. And I felt like my background unexpectedly set me up to do that.

19:48

And why was it that when you went to search for the company that was already doing this, nothing came up? Like, what about this problem in particular didn't seem sexy to people?

21:32

I think it was just a completely unknown problem. It's the sort of thing where people

21:42

just do people just think about the reactor and then they don't actually think about the supply chain behind it.

21:46

Or you assume the supply chain is fine because existence proof we have nuclear power in the country today. Obviously the supply chain exists, obviously there can't be an issue. But then when you actually realize, okay, we need to do something slightly different and we don't have the capability to do that, you start pulling the string and realizing you do the research and you realize the current status. So I think it was off. People's radar. I think when you think about an enrichment business, it sounds like a big industrial chemical business, a processing business, which it is, and that doesn't sound that sexy. And then I think there was also an element of wondering, are we even allowed to do this? And so when we started, our first conversations were with the doe. Hey, this is what we're hearing from reactor vendors. This is what the problems that we're seeing. Is this really an issue? Is this something that a private company could help solve? And the answer was yes. But prior to that you would say, okay, let's look at who's done this before. And it's all nation, state or state owned entities, spin outs of national labs, companies that used to be part of the Department of Energy. Clearly very, very strong government DNA and a history within the government. And so is this even something a private company is allowed to do? Through a lot of conversations we were encouraged to do this. We learned that, yes, you can do this, but we need to work really closely together. I think it's a combination of, to your point, it's not the world's sexiest business. It's also not something that's been generally on people's radar because there was no need for new expansion of enrichment in the US at least, so people thought. And then there was a question of is this even allowed? And what would it take to get the government support. And so now through a few years of work, really starting in 2023, we're working very closely with the DOE and the NRC on bringing this capability back.

21:50

You worked at SpaceX from like 2003 to I think like 2007? Ish, right. What was that like? And then what were you able to kind of like take from that to starting this company?

23:56

Yeah, so SpaceX was not my first job, even though it's my first job out of college. The place I worked right before SpaceX was Boeing. And so during college I was an intern at Boeing on what seemed like the coolest project you could do, which was a government military project at a top secret location working on a military program. And so I had been a mechanical engineer, aerospace engineer, undergrad and master's. And so this seemed like a great place to go. Much more interesting than working on commercial airliners, I thought, so let's work on a military program and let's try and push the limits on what's possible in the aerospace world. And it was a multi thousand person project out in the desert and it was relatively slow. I tried to work as fast as I could do as much as I could. And I actually got feedback from coworkers who said, hey, you're doing too much, you're going too fast, you might as well, you should slow down a little bit. We have to make this. They literally said, we have to make this project last.

24:06

Is it because of the incentives with how they were getting funded?

25:11

I think it was partly culture, it was partly the people there. But ultimately you could say it came back to the incentive structure of running a cost plus program where whatever you spend, you get that times 1 point X and that's the money that you make. And so, yeah, there's a very clear incentive to maintain the contract, do the best job that you can, take very little risk, and on the margins maybe go a little bit slower to do a better job and take less risk. And that's going to increase the budget of the project. Now you only want to do that up until the point that the project keeps going. You don't want to push it so far that it gets questioned, but to a limit. That was the incentive. And so I had worked on a very large aerospace program in college, and then after college graduated and well before graduating was an intern at SpaceX in 2003 and then joined full time in 2004. And when I got there, it was, you know, 30 something people as an intern, 50 something as a full time employee, and it was like night and day. And even after my internship, I knew SpaceX was going to win in the launch market just by seeing the two cultures.

25:15

This was like years before. Didn't they land their first rocket in 2008?

26:26

First successful launch or first successful orbital launch was 2008. And so, yeah, to your question, what was it like? It was a, it was a team of, let's say, 50 people working on the basic infrastructure required to achieve the mission of SpaceX, which was, let's make spaceflight much more affordable so we get much more of it and let's make it reliable and safe and bring down the costs and let's actually help humanity make the next step towards Mars. So there were 50 people who were a mix of new grads, people from the industry, all true believers in the mission, working in this small company that was relatively unknown at the time, not taking the standard aerospace job, but believing that a small group of people could do something to move an industry forward that had completely stagnated for decades. And so when I was there, it was 2003, we were starting on building test stands, there was initial engine designs, initial engine tests, and we Just scaled that up into engine development program on Merlin, on other engines, get those working, get performance up higher and higher, integrate those into Falcon 1 and the original Falcon 1 launches. So that was like a three year arc to see that progression. And it was just a lot of people working as hard as it took to make things succeed and working constantly, 60 hour weeks or more, but you know, not 100 hour weeks every week. It was, this is something where we're going to, we know where we're headed, we want to get there as fast as possible at the lowest total cost. And yeah, let's do our best to make this succeed.

26:30

One thing I've noticed over the past probably 20 years is Elon's way of building companies has drastically shifted. Like even 2008, 2009, SpaceX and Tesla were near bankruptcy. Same thing with like 2018 production. Hell, that brought Tesla close to the brink. I think for a very long time. Elon didn't want to run multiple companies. Tallulah Riley at one point said he was like waking up in the middle of the night, throwing up and stuff, trying to basically survive. What was kind of the evolution of the way that he operated over that time frame?

28:10

I would say yeah, when I, when I started there in let's full time 2004, the pace was fast, everyone was working hard. But I think the lesson that's been learned since then by Everybody at SpaceX is just the feeling of urgency of how long these projects can take if you allow them to and how close survival might be to the line and how you just don't know where the line is. And instead of letting things get to the brink with like you said, two companies that were on the verge at the same time trying to survive that nearly impossible. Instead of getting to that point, let's work harder upfront, let's go faster upfront, let's have higher expectations of ourselves to avoid ever letting it get that close ever again. I think that is the lesson I would draw from that is people were probably too comfortable. We all were very optimistic, we all believed it was going to work, but it did turn out to be close. And so we probably should have been stepping on the gas even harder on day one. And so maybe that's one lesson I think, yeah, there's probably many, many lessons learned over the years about how to maintain a really strict meritocracy, how to reward the best people, how to attract the best people and how to just have extremely high standards around all these things to actually achieve the long term mission. Because if you allow any drift, you're very unlikely to actually make a huge impact. And so you've just got to be extremely strict, extremely focused on what you're trying to achieve. And then today, I guess we're about 24 years into SpaceX. And so you realize these things can take a very long time. And so we should try and go as fast as we can so we can really get it done.

28:42

What was the feeling internally when the first three rocket launches failed?

30:28

So I was there for the first two. The first one was, I think people were pretty discouraged. That was, we had obviously hoped it would get to orbit and we did get off the pad and got some altitude, but then there was a fuel leak in the engine area, which caused a fire, which led to the rocket not making it to second stage ignition or separation or even max Q. And so that was pretty discouraging. But we knew that the first one, you know, it's almost like in the Matrix, no one ever makes their first jump. Very unlikely for our first rocket to actually make it. And so we knew, okay, we should expect that this was a coin flip on the first one. And so that's why we have enough funding for 3. And the second one really was very close. 2 and 3 were very, very close. The second one had a fuel slosh instability that led to spiraling and then eventually air and gas ingestion, which made the engine.

30:33

And they're basically just not in the fuel tank. There wasn't these stoppers to stop the sloshing of the fuel.

31:31

Yeah, there weren't sufficient baffling. There wasn't sufficient baffling or there weren't sufficient baffles inside the second stage tank to prevent that oscillation from having a positive feedback loop. And so, yeah, eventually the engine ingested gas and then, you know, that was it. On the third one, it was a really fluke thing where one component changed and then that created a time delay and some gas release. And then the two stages bumped just as the second stage was trying to fire off, which caused an explosion there. But after the first launch, it was really close. There was just these little things that came up two more times. But I left thinking that after the second launch, okay, we've absolutely got this in the bag for the third one. Um, we're in good shape. Let's, let's go try and do other things. But obviously There was a 20 year history after that of doing way, way bigger, more amazing things with land, like reusability, starship. And so that was really just the Beginning. But as someone straight out of school, you don't always realize that coming out

31:37

of SpaceX, how'd you kind of decide to figure out what you wanted to do next?

32:41

That is the important question that is often very hard to answer, what to do next. I left thinking that the next step was to really to learn to gain experience in the business world. And so I did just under two years of consulting at a management consulting firm, thinking I would learn a lot from many existing businesses and how they operate, how to make them better, that this could be really useful in the future for working at another company, whether it was my company or someone else's company. And then you learned that actually many of these big companies are just doing things the way they've always done them. There may have been a moment in time where they came up with these processes, where they thought from first principles, but now they've been around for decades and the way they're operating might be out of date. And so let's go rethink these things. But the moral of that story was just that you can just think things through and figure it out. And so basic tools like the ability to just analyze systems, whether it's a business system or an engineering system and sort of an engineering or math background was really useful for thinking this stuff through. And often the people on the ground didn't have any really unique special insight into why they were doing things the way they were doing. And there was often a better way to do it. And so that was probably the main lesson drawn out of that, but knew I didn't want to stay in that industry forever and was thinking about what was next and ended up thinking, okay, maybe I will go to another tech company or even could go back to SpaceX at some point. And then actually met Peter from Founders Fund, who convinced me, hey, don't join Startup xyz, join our startup in the venture capital space. And that was 2011.

32:47

What did Peter say to you when he was trying to convince you to join?

34:36

I don't think it was any one thing. We just had a series of conversations about what I had worked on, what I was thinking about doing next, what was interesting. And at that time, that was early years of Y Combinator and there was clearly this shift in how companies could be started, how they should be started and funded founders funds. So Y Combinator was taking the thesis that companies are much lower cost to start now and a version of software will eat the world. And there's all these opportunities and teams of two living in an apartment can get something started and improve real traction. Their thesis was very focused on that, which I do think was separate from what came later of MVP lean startup sort of philosophy. Many of those companies had big missions and just were just trying to go as fast as they could on low budget founders funds. Philosophy at that point in 2011 was very focused on really two things. One, in the name Founders Fund, we will back founders unconditionally no matter what. And this was not the case at the time. This was coming out of a 2000s era History of thinking of the company as a horse and a jockey. And the horse was maybe the market or the technology and the jockey was the founders. And maybe you had a great horse, but you could just get a different jockey. That had been the conventional wisdom at the time. Founders Fund took a very different stance on this and said basically the founder is almost the company. If you think that you're creating a great company and you can just swap the CEO and they're just a fungible resource, then you're probably just building a very normal, very standard commoditized company in some way. The idea was that the founder who really noticed the problem, came up with the idea, had really the moral authority to lead and only they really understood the inner workings of the company and could drive it properly. That maybe was the one view of it. The more pessimistic view would be something like, if a company is struggling so much, this would maybe be the VC view. If the company was struggling to the point where you need to replace the CEO, maybe you should spend your time on finding different companies to invest in, not trying to fix companies. There's a positive version which I think was really the real version, which was great. Companies are run by their founders and this was statistically true. If you look at public companies, then you look at two baskets. Founder run, non founder run, the founder run have done much better historically. And yeah, then there was the VC version of let's focus on the great companies and put all our energy and time there and not be too obsessed with helping a struggling company barely make it. And so that's played out as well and was known for throughout the 2010s even today as the power law where the best companies are worth much more than all the others put together.

34:39

Yeah, the way I think about it is basically the founder is kind of the keeper of the soul of the company and if you pull the soul out, you're not going to have the same like the mission disappears. Honestly, if you think about like Uber Pre, while Travis was running it. It was a completely different company than the one that exists today. I also find it a little bit interesting when you have that situation where Travis does leave or he has to leave, where there's almost like a loss of story, almost.

37:43

Yeah, you could say there's like, some people refer to NPCs, but you could have like an NPC company where it's just this standard thing that's doing its bureaucratic motions and what's really driving it, what's making sure that this is achieving the mission that it set out to do, and who's really going to hold it accountable and ask the really hard questions and make the really unpopular choices that put it back on the right track when it gets off track? Yeah, I think only someone who knows the whole history, who's really willing to sacrifice a lot before it's all proven, clearly has that motivation. And so the best bet is just let the founder run it and try and work with great companies. From the VC perspective, what was the

38:12

most unintuitive thing that you learned from Peter?

38:53

Probably coming from the business consulting world, which I'd done right before Founders Fund and been in that world, that world thinks pretty differently than I would say, Peter would think. And so oftentimes it's a lot of conventional frameworks, more conventional, straightforward, linear thinking to arrive at hopefully the right answer. Yeah, more linear thinking, I would say, in that world, which can result in more conventional answers and more conventional ways of doing things that are often first principle. But I think the big difference with what I first saw at Founders Fund was that it wasn't this linear thinking approach. It was often very abstracted. And so you would often ask questions that were not obvious. I think it went back a lot more actually to asking the right questions than just following the breadcrumbs of. Okay, let's dig into this more. Let's dig into this more. Okay, we don't have much of an insight there. Let's go over here and let's try and figure something out there. But it was taking a step back and trying to figure out, okay, how do we look at this from a totally different angle and try and realize something that could surface one problem or company or solution above the others? And which of these angles do we think is actually correct? And sometimes it would involve multiple layers of abstraction to generate new insights.

38:56

Do you have an example of that?

40:33

One company we invested in? Probably my first or second year, there was a cancer therapeutics company. And this cancer therapeutics company was really focused on a Slightly different or not articulated theory of cancer, which was that ultimately most cancers were really driven by stem cells. And you just had to really focus on the stem cells and how to differentiate them from the rest and how to specifically try and target those cancer cells. And you would say, okay, and maybe by today people realize this and it's not unconventional, but 15 years ago it was not popular and this company was uniquely focused on that and figuring out, okay, how do we go after those, how do we develop targets for those? And you might say, okay, let's spend a lot of time looking into the mouse studies, which we did, and let's spend a lot of time looking into the biology of this. But I think the conclusion was yes, this all looks good, but we founders fund don't have any really special insight into biology. More than you know, people in academia do. They know much more than we do. And so these mouse studies can look good to us. Mouse has been. Cancer has been cured many times in mice before. What's different about this? We can try and figure that out and decode that and come up with a really high conviction answer on why that's the case. But then there's this other layer of okay, how do we know that this stem cell theory is correct? And I think at the time there was, there was one insight of at the same time there had been a very pro stem cell argument in the scientific and government community about how these were actually useful for many things, which they were. And so there was the conclusion that hey, maybe this is a really under explored area because it's maybe unpopular to say that stem cells are involved in cancer because we're supposed to be looking at only positive things about stem cells. This was maybe yeah. And then that company ended up succeeding from a business perspective. And then many of their targets that they developed are now in cancer therapeutics that are coming to market. So it turned out to be correct. So I think that's one example in the biotech space. I think in the, if we go back to the, the example of SpaceX at the time the launch market was not a huge market, it was single digit billions. I guess this was another example of the abstraction going to a social, political level and saying what things maybe for the first time could be done well by a private company as opposed to a government organization. And so space launch had historically always been done by the government or government contractors. And a lot of people accepted that that was just the way that the industry had always been and would always be and that maybe it could never be Highly commercial and it would always need to be subsidized by the government. And we were never going to get a ton of growth in space launch. It was just a nation state capability that you had to have. And maybe I think the thinking or the different questioning was around, yeah, is this true and is it generally the case that private companies can do things better? And if so, how far does that span? And so SpaceX has proven that one,

40:34

I think one of the least understood things is it's very important to ask the right questions. And I think prior to SpaceX people were not asking the right questions as to how to get mass to orbit. And Elon was, from what I can tell the first person, it basically just optimized for this one key variable, which is how do you get mass orbit instead of, you know, what does the US government want us to do? And then with that, he basically just went all in on this idea of just lowering launch costs with the understanding that hopefully there was going to be a market afterwards. And it seems that, you know, with what he's done recently, it's effectively just scaling his own network of satellites because he's able to fill his own supply. Um, how did you figure out what the right question to ask was for in like key KPI to think about when you were starting general matter?

44:03

Yeah, so the, the way we realized the importance of enrichment was talking to the advanced reactor companies and they not only said, hey, we don't have a source of fuel, but they said, even if we can get it, the fuel's too expensive at these fuel costs. Look at my economics, look at how much of this fuel cost is driven by the fuel. And then if you decompose the fuel into the different steps, enrichment was the biggest step. And so it was just by digging into the problem and understanding where the cost structure was that for us we realized that the key cost variable was cost per unit of enrichment drove the cost of fuel, which would drive the cost of nuclear energy going forward. I think this was similar to the example that you just gave at SpaceX, where you go back to the 1990s. Satellites cost billions of dollars. Therefore all that mattered about launch was that it was highly reliable and launch costs therefore didn't matter that much. If you're paying 100, $200 million, $300 million for a launch, it is costly, but it is not the difference maker. What you need to ensure is that your launch succeeds. Fast Forward to the 2000s. People were talking about microsatellites, smallsats, those all of a sudden were much, much cheaper, $10 million, $1 million, not a billion dollars. And so now this $100 million launch cost really, really mattered. And so if you want much more of these small satellites, yeah, we have to bring the launch cost down. And launch cost is really dollars per kilo, but it's dollars per kilo to an orbit, which is really like dollars per kilo times Delta V. And so as commonly understood, it was dollars per kilo to low Earth orbit, but that was really dollars per kilo times some Delta V. And so for us, the version of that that makes nuclear energy more affordable, makes advanced reactors really succeed economically, and makes their fuel lower cost, is bringing down the cost of enrichment, which is dollars per unit of enrichment per kilo. And so you can take that one step further and make it dollars per. It's called swu is the unit of enrichment, really dollars per swoo per kilogram. And so that's the metric that we really focus on, which is colloquially known as dollars per swoo. Dollars per separate of work unit. And so it's basically how much does it cost us to reduce the entropy of a kilo of raw uranium? And the amount that you reduce the entropy is the amount that you're enriching it.

44:54

What was the India Index for enriching uranium?

47:32

That is a great question. We haven't really thought of it as much in those terms in space, the Idiot index. So the Idiot index is generally understood to be what's the cost of a thing compared to the cost of the raw material?

47:36

So if a rocket costs $100 million, but the raw materials are 1 million, then it's 100 million index.

47:54

Yeah, that's right, yeah. So that's when you're specifically making hardware that then goes and launches and performs a service. In our case, the service that we're performing is enrichment. I would say it's commonly understood to be an energy input problem. So a little bit harder to get the exact Idiot index on that. How efficiently can you use energy to enrich? I think in the gaseous diffusion world it was a lot of energy cost. And so that was the focus of how do we bring this down. But if you actually look at a facility that's doing enrichment, there's a tremendous amount of CAPEX upfront spend in fixed equipment that does the enrichment. And actually the energy cost is relatively low. And so the version of your question is the Idiot Index, as popularized by SpaceX, is a guide on how much should we be able to bring down the cost. And SpaceX has always said we can bring it down 10 times an order of magnitude over time, maybe two orders of magnitude. Get it all the way down from that 100x100 million dollars, not to $1 million, but maybe even lower if we can reuse this hardware and not throw it out every time. For us, the focus has been how do we make the most economical enrichment service in the world. How do we make it extremely safe and reliable, but primarily scalable and lower cost? And that comes through really thinking through how do you design a much, much better facility, how do you do better building construction, how do you make the guts of it better, how do you site in a better location, so on and so forth. And so we think there's easily potential for a 2x reduction in the cost of enrichment versus where it sits today and likely more. And there's a lot of ways you get there. You get there through designing your own technology, not using off the shelf. You get there through building your own in house EPC rather than outsourcing everything to a gc, instead manage your own subcontractors and run your own project and maybe even site in a place with relative to do those final optimizations. Pick locations that have available energy at

47:59

reasonable cost when you were starting out. Because there's literally only one operational enrichment facility that's just not R and D focused in the United States. And you know that it's going to take multiple years, maybe half a decade in order to get this thing operational so you're actually producing your rich uranium. How much were you thinking? We're going to vertically integrate and try to do all in house versus go find other suppliers and things and take learnings from other people that are currently doing it.

50:17

I think the intuition was that we would have to do it in house. That was almost out of the gates and almost an assumption. And we did question this a little bit, but it became very obvious that a lot of things we would have to do ourselves in house. Really drawing on the lessons that we saw in the early days of SpaceX. And so our team comes from a whole mixture of places. There are people who worked at SpaceX and then went to work at other companies and wanted to get back into something very mission oriented. But places like SpaceX and Tesla and Anduril, the whole tech world, who has actually broken into incumbent dominated high capital, government partnered industries. So that's the tech world. We also have people from national labs who have seen the full range of the most advanced technologies and then even people from other parts of the nuclear fuel Supply chain who are very aware of how this works. And so combining that experience and just talking through, how should we do this? Really, in the first months of the company, it was pretty obvious that there were certain things you could use existing supply chains for, but some things you would just have to do yourself, those were pretty straightforward, I'd say. The thing that we learned in the first year of the company that maybe is more surprising is that not only did you have to do that, but you really should design and build the buildings yourself. Not with every single piece of expertise, not with every trade inside the company, but you need to project, manage your own project, not just for cost, but for schedule. And the minute that you throw your building design over the wall and you say, hey, some outsourced firm, can you figure out what we want and design a really great building for us, there's so much that has to be explained in that process that it's unreasonable for someone else to figure that out. Unless you're. If you're building a warehouse, if you're building like an Amazon warehouse, maybe it is so standard that it's just. It could be done by someone else in a cookie cutter fashion. But when you're doing something more bespoke,

50:42

this is a little bit more of the product is the factory, and that's what you have to get great at building.

52:50

That's right, yeah. And that extends to obviously the guts of the factory. But the thing that was unexpected or that we learned over time, and that's now obvious in hindsight, is you should design your building the way that you need it to be and you should manage that project just like every other piece.

52:56

Did you try to outsource this initially?

53:12

We didn't go too far down that path, but it's. What's typically done in the industry is you hire a GC and they run your project and they do all that. And we felt like it was important to do that in house. And that's something we saw and learned by talking with people who had been part of the Tesla Gigafactory builds and just hearing what their experiences were when they tried to do certain things with third parties with, who are going to need more guidance than you can provide versus just doing it in house.

53:15

So we were just talking about the IDYA index in reference to basically the cost of the finished good versus the raw material inputs. I also think something that's interesting to think about is how long a conventional timeline is versus what the actual timeline is. If you kind of eliminated all the unnecessary requirements and all these other things. How do you think about kind of the idiot index for timelines?

53:44

Yeah, the idiot index is a pretty harsh way of saying it because often, I mean, often things are the way they are and the people working on them are really smart. They're really smart people. They're not idiots. They. They're trying pretty hard to do their best. But people are often working on something quite narrow or their piece of the puzzle is just what they're working on. And the thing you really have to do is just rethink the puzzle from big picture. And so, yeah, the schedule piece, I think commonly understood, it's a cost thing. Like on a rocket. Yeah. It costs $100 million for a launch. You throw the rocket away every time. Why is it that $100 million comes from only $1 million parts? How can we make this thing cheaper? There's the same thing on the schedule side where if you just take the conventional approach and you say, okay, this is how it's been done, let's continue doing it that way. We do this step first, we do this step next, and we keep working our way down the chain, it's going to take five years, and that's just how we should do it. A lot of times you can just question, do you really need to do everything in sequence? Which things can we do in parallel? Maybe we're taking some financial risk to do it in parallel. Maybe we will have to go back and redesign something that's inconsequential for some other paths, but to do it better. But how do we just get going? And how do we get something started or online or launched as quickly as we can because time's wasting? And in our case, the industry needs us to do this. And so how do we go quickly in a safe, responsible way that may end up being slightly higher cost, but may actually be cheaper because we did it in half the time. And so the cost side is pretty well discussed at this point. On the schedule side, I think there's really probably two main variables. One is, which things are you doing sequentially versus in parallel? And then how long does each step take? And both have to be interrogated to really get to get the shortest amount of time. In our case, that means everything from at the same time doing engineering work, tech development, standing at manufacturing capability, finding the location of the facility, preparing the grounds of the facility for construction, working on licensing, all these things at the exact same time. So. So that you're not just waiting for something to be completely done and perfected, and then Going to the next step, which could make your project take 20 years versus five, three years. So one big dimension is parallelization. I think that's pretty obvious. The second dimension that's probably not fully appreciated is one that wasn't even, to me evident at SpaceX in the early days. It's how much are you really questioning each block of time? What's the resolution of your Gantt chart? Is it one month, Is it one week, is it one day, is it an hour? Can we question these tasks of like, oh, yeah, that'll take me a week. Okay, is it really a week or is it like four days? And how do we start really pushing on the limits of what can we do? How quickly can we get this stuff done? Why are we going to wait for a week from now when someone's waiting on the decision from us to go design this other component? When we're all in a room right now, no one's leaving this room. Let's just decide right now, in the next 15 minutes based on what we know. If we can't get to a decision because we need more information, let's reconvene tonight, Please go get that piece of information. Let's decide, because that other team over here is waiting on that. We need to get them that information or they're going to just be burning time and we could be moving forward. I think it's really those two things of how do you parallel path things? Second, how do you bring in the time or bring in the resolution of this? I guess there's a military saying about this, that minutes make hours, hours make days, days make months. How do we just really focus on the minimum amount of time that we can get something done? There's probably a third piece that I did see at SpaceX, even from the earliest days I was there, which was this idea that instead of deliberating over something to get to 99% certainty, let's just get to where we think it's probably the right choice with some degree of certainty.

54:06

More than 50%, more than 90%.

58:34

Maybe 80, maybe in some cases, like 50 maybe. There's cases where, hey, these two things we really understand, they're both pretty good paths. Instead of figuring out which one is 10% better, let's just pick one and not waste three months and let's just go and let's make forward progress. Because that momentum, every day we're waiting, we're spending money, there's salaries, there's other costs. The opportunity might be going away, the industry might be struggling because we're not there to help. And so it's this concept of if something is reversible, if it's easily reversible, decision just go. And so I think that can shave months as well.

58:36

I remember Elon said that he had, or like in Walter Isaacson biography, they had this mental model where he would say every day that they didn't get to orbit or some other metric was basically a lost Future Day of 10 million of revenue. And so if there was some thing that they could just move one day sooner by spending $100,000, that may make sense, like flying one part from like LA to quad or something like that. How do you kind of think about taking risk and what things are you willing to basically take risk on in order to move the timeline just slightly faster or maybe move the probability of the timeline happening faster?

59:14

Yeah, yeah, there's two versions of that metric. One version is the tail end. Like, okay, what date do we go online to your point, how much, how many dollars per day of launch revenue or whatever business you're in? So that's the once you're online date. And if you're building a pharmaceutical plant, okay, a million dollars a day of revenue, every day we can bring that in is worth a million dollars. The other version of it, if you're a very early stage company, imagine three people in a garage working and they say, well, our product, once it's out there, we'll be making hundreds of thousands a day. We need to spend $100,000 today to go one day faster. You could run out of money. There's a current burn version of this, there's a future revenue version, and there's a current burn version. The current burden version is basically how much are we spending today on salaries? If we can bring our schedule in by one day, we know it's at least worth that. There's two bounding versions of it. Probably the, the big picture one to really focus on is the once you're online version and not be myopic and just worry about today's spend because you're trying to really get to that end target, but you've got to survive to get there. And so the right answer is typically somewhere in between those two things, subject to one requirement, which is you're not increasing your cost structure of the company with that decision on an ongoing basis. So if you said, hey, my burn today as a company is 10 grand a day, let's bump it to 20 grand a day to bring in schedule one day, the next day when you Evaluate this question. You're going to say, our burn Today is 20 grand per day. We should bring it in by day. You can get into sort of escalating spend, but when they're one off decisions, usually there's opportunities that are not marginal. It's not like, oh, we're going to earn an additional million in revenue by spending 900 grand. It's usually much more like, hey everybody, take a step back. We can shave this amount of time off the schedule and increase revenue like total revenue that we ever collect by a million dollars by spending five grand. I think that's a good trade. Let's do that. And people are very reluctant to in a big company propose those things. Propose we're going to spend a lot more money on this project to get it there faster. But I think that's one of the tools that's in the toolkit to accelerate. And you see the same thing even in the film industry. Sometimes they will have to wrap up filming of a movie faster to actually save money. A lot of people think, going back to the speed question, people associate going faster with spending more. Sometimes you can actually go faster and spend less and get a better product. So I think that's, that's not commonly appreciated. And remembering that's really important.

59:51

What's been a moment over the past two years while you've been working on this where you've been able to make one of those rapid decisions that was able to kind of bring in timelines or you spend a little bit of money and shave some massive hurdle off your checklist?

1:02:45

Yeah, maybe one of the decisions that we made over the last couple years that in hindsight is kind of obvious, but that we could have spent years and years trying to think about was where to put our facility. And so the first year of the company, we spent a lot of time going around looking for the right sites to put an enrichment facility. And one direction we approached this one from was which states, really thinking about this from a commercial perspective, which states and areas are very supportive of energy production in general, and which ones have some history of nuclear energy or uranium mining or different parts of the supply chain. And let's go to those places and let's look for vacant land that could potentially work. And so we spent about a year visiting something like, I think it was 11 different states looking at over 1,000 pieces of land, not in person, but one way or another, and certainly hundreds of in person. And we could have really boiled the ocean on that and just tried to come up with complicated frameworks for evaluating these things and trading them all off each other. And after some time of looking at it this way, we came across Paducah, Kentucky and we realized this community is extremely supportive of enrichment. It's the last place in the country that it was actually performed at scale. It was shut down in 2013. So people remembered what it was like to have the facility and understood that it was safe and an important part of the community. Yeah. And so really this question of what is the right place to go, you could do a really exhaustive search and do community surveys and look at costs and look at tax incentives and look at the cost of land and look at your supply chain and shipment and all these things. Or you could just say, where's the last place this was done in the country that we know people are happy to have it in their community and appreciate it and understand how important it is? And would point you straight to Paducah, Kentucky. And so once we realized that that community, you know, once we had visited that community, we understood their support for nuclear and for enrichment and then were able to work out a lease with the Department of Energy on a part of the land on that site. It was, it was a very obvious choice that we could have otherwise spent years trying to figure out and which other companies have spent years trying to figure out looking around the US So we made a really well informed choice because we'd done all that work in other areas. But it was very obvious to us once we finally picked Kentucky and it's been awesome since. And that allowed us to really get started on all the different pieces required to bring the facility online.

1:02:59

I think with this sort of company, this is going to be a multi decade journey. And even setting out, I think your guys's timeline is trying to have the first facility online in the next five years producing enriched uranium. How do you kind of think about maintaining this like maniacal urgency while also understanding that these are like multi year timelines just to turn on the first facility and then scale up as quickly as possible?

1:05:43

Yeah. I think from a, from a motivational standpoint, how do we keep the urgency? I think that's simply a consequence of the market as it stands today. You asked about the Russia uranium ban that goes into full effect January 1, 2028. We know that we need to be online as quickly as we can so that there's a reliable backstop for that, a reliable source to really offset that loss of US supply as soon as possible. If that's happening in 28, we want to be online before end of decade to help support that. Then you look, that's just for low enriched uranium. Then you look at Haylou for the advanced reactors and they have no reliable source of fuel other than the DOE today, who only has so much to give out to reactors. And you realize, okay, these are primarily venture backed companies. If they can't show deployment and scale up in a few years, that's an issue. So from the market's perspective, the motivation is very obvious. And so we're trying to serve that market as quickly as we can for its sake. And then how do you do it internally? I think it comes down to it's probably two things. One, let's look at that finish line and how quickly we have to actually do certain things to hit it by that date. And let's break the schedule into big blocks and think internally in terms of those big milestones of to hit that date, when do we have to begin construction, when do we have to begin scale manufacturing, et cetera. Um, and then it's decomposing that into like a very detailed, almost like day by day schedule that we call the master schedule, and keeping track of literally every major item that's either under work today or coming up very soon. And so we actually had a review meeting today, which we do once a week on each of these items. Where does this item that's active right now that someone's working on, where does it stand, what's its required completion, what's its expected completion date, how are we doing in that? Are we ahead of schedule? Are we behind schedule? What does that drive? It's looking at every piece of work that's going on, understanding how that's either pulling in or pushing out schedule based on completion and then using that as a tool to understand where should we put more resources and how can we accelerate that thing that is the primary schedule driver, which we call the long pole in the schedule. And so these things stack up. Then you want to look at all those and figure out how do we resource that primary schedule driver and then how do we watch out for the ones behind it? So if we take care of that path that's driving schedule, what's the next thing that's going to drive schedule and how do we address that ahead of time and over resource it so we can just keep pace, going as fast as possible. And so the more and more things you're doing in parallel, the more complicated this gets. But yeah, taking that big picture, decomposing it, breaking it down to all the line items, understanding what your Schedule drivers are and then resourcing them and addressing areas where we're stuck or need to make a decision. That's how you pull it in.

1:06:08

I think we like to prepare for the last disaster. And so if you see some major earthquake or a fire burn down your house, suddenly everyone wants to buy earthquake insurance or fire insurance. I saw Elon tweet a while back where he's like, war with China is basically virtually guaranteed. It's just a question of when. And like Andrew has this internal policy where they want, you know, most done by like 2027 or something like this in a very like unfortunate scenario where we were to go to war with China even earlier than 2028. How does that impact like our ability

1:09:12

to produce energy in the nuclear supply chain? It doesn't, fortunately. So right now all of our enriched we can talk about the different steps in the supply chain. But from mining to conversion into gas, to enrichment, to deconversion and then fuel fabrication, these five steps currently in the US don't involve China. So mining is in the US in places like Wyoming and Texas, but also Canada, Australia, other places in Europe like Kazakhstan. Yep. So uranium comes from a bunch of sources that are not China. Conversion today is done in the US and Canada and also in Europe. And most US utilities purchase from US and Canada. And then enrichment is a European and Russia supplied step of the supply chain. And then finally field fabrication is primarily done in the US by a mix of US and European companies. And so in nuclear there really is not much impact. Other sources of energy there would be like solar panels. Yeah, like solar, yeah, probably batteries, range of things. You know, if we rewind like five years ago, China was very, very tiny part of global enrichment capacity. If you look today, they're roughly 15%. Next decade, if they keep growing, they will be 30%. So to your question of conflict with China or China's role in the nuclear supply chain, it's growing extremely rapidly to where they could be neck and neck with Russia and China next decade if the US does nothing. But the same thing was true of launch. If we're going back to the SpaceX story again where 2002 US was below 30%, I think of global launch capacity and now is over 90% of global launch capacity with China as number two. And if the US had stayed where it was, China would be number one. So I think on this question of the US's role and China's role in this industry, the US has primarily just stayed flat for decades. China's growing rapidly. This is not the only sector where that's true, but it's one of these situations where if the US doesn't do something and bring back its own capability and expand that capability rapidly, China will dictate where the industry goes. That's certainly part of our mission, is to bring back the US capability, but not just for an onshoring thing, but for a move the whole industry forward thing. It's not just is the US capable, can we supply our own power plants? But can we do that in a way that's so low cost that advanced reactors become much, much more viable and we get them everywhere and we get even safer baseload power, even cleaner baseload power at much lower cost. And so you look at the stats and nuclear energy statistically has been the safest, cleanest form of baseload for a very long time by a wide margin. But it has not been the cheapest form. It's been more expensive than other forms.

1:09:44

And so is that mostly like a regulatory issue?

1:12:58

I think it's a bunch of issues. A lot of people like to blame the regulator and say the regulator makes it too hard to do things. In our experience, that's not true at all. I think first industry, I worked in space launch people would have said why is space launch so expensive? And you'd say, well, it's really hard and you need to get this reliability. And maybe the regulations are hard. And then it just turns out that a lot of the organizations were just doing things the way that they'd always done them, that they were cost plus organizations, that they didn't have a lot of incentive to actually push the limits, to actually innovate and question why they were doing things, question the requirements. And I think what we've seen now in space launch, for example, is it's even safer, even more reliable, even lower cost. And that's just through getting iterations. So, and by scaling up. So yeah, our goal is not just bring back US capability, it's not just make it lower cost, it's also make it more scalable so we can power all the things the US wants to do.

1:13:01

What have been the biggest conventional wisdoms in turning on this sort of operation that were actually wrong when you looked at the facts.

1:14:08

Maybe the most, the least obvious thing on question requirements around a company like us, or even even some of the nuclear companies in general, reactor companies, people would think that the entire team must be comprised of people with nuclear PhDs. And then when you actually, and I think that's the, that's the natural Intuition, you would say, hey, this is a nuclear energy company, therefore it's a nuclear science company. Therefore, you need nuclear engineers. And you do, and you need people who understand the licensing, safety in the case of reactors, nuclear reactions. And so it's a part of the team, but it's not the whole team. And so now you look at some nuclear reactor companies, and yes, they have a handful of nuclear engineers, but it's mostly mechanical, aerospace, electrical, software, building the other systems that go into it with only a core group that's really looking at the nuclear reactions and making sure that you're going to get the power production, you think, out of a reactor. In our case, it's even one step further where, yes, there's nuclear engineers on our team, and they're a really important part for making sure that our system's safe and that there are not nuclear reactions. But you look at an enrichment facility, and there's actually no nuclear reactions occurring. There's no chemical reactions occurring. It's simply phase change and separation. And so the team that you need to do that again, yes, we'll have some nuclear engineers, we'll have some chemical engineers, not for the sake of chemical reactions, but like process development, making sure you're getting the performance of the system that you need, making sure that your pumps, power systems, everything, are what you think they should be, what you think they're going to do. But it's mostly software, electrical, mechanical, aerospace engineers. And which is why we chose to locate the company in Los Angeles, really the. The hub for that. So your number one assumption would be, oh, we need a lot of nuclear engineers. We need to be located in or at a national lab or right next door to one in a historically nuclear community. And after thinking about that at the very beginning of the company and okay, should we locate the company in Oak Ridge? There's an incredible amount of talent there in the nuclear industry. Should we locate it somewhere else? Los Alamos? We realized, no, we need to locate the company where most of the employees are going to be coming from and where we think we can recruit great nuclear engineers, too, which is la.

1:14:17

Yeah. This is similar to what Elon has kind of done, where Dorakesh was asking about, like, do you try to, like, optimize for PhDs or something? And he was like, no, I just want great engineers. And great engineers don't always have PhDs.

1:16:51

Yeah. I mean, great engineers even. Even might not even have a master's degree or even an undergrad degree. There's people who are incredible technicians who are basically Very hands on engineers who might not have gone to college, but maybe they have worked on all sorts of mechanical systems for decades, even from when they were a kid because they were tuning up their shifter car or doing car engine rebuilds or they were working on hot rods. And so I've worked with people before who are some of the best engineers who don't have a formal degree at all, but have really built a lot of systems and have really smart ways of doing things that someone who maybe spent four years in university wouldn't realize.

1:17:03

They don't have a traditional proof of work, but they have real, real world stuff.

1:17:46

Yeah.

1:17:49

I think one thing that's very important to cover is basically the idea that we've got this massive tsunami which is AI and you have to turn on all these data centers. And I think a lot of data centers you need nuclear power.

1:17:50

Yeah.

1:18:00

But if you can't actually produce power because you don't have enriched uranium, it doesn't work.

1:18:00

Yeah. When we started the company, the whole AI data center wave was not really top of mind for people or even for us. We were just very focused on bringing back capability as quickly as possible. So there wasn't the AI data center boom, there wasn't the Russia ban. It was, we need to do this and we need to do it for the advanced reactors as quickly as we can so that they can succeed. Now it's become obvious that not only do we need to do that, but we need to massively increase the amount of nuclear fuel produced to power all the reactors that people are working on that the hyperscalers are planning to put at their sites behind the meter. It's one of the most urgent things we think about is how do we build for capacity. Right now, as we build enrichment capacity in Kentucky, we're sizing that to completely satisfy the domestic industry all the way through 2040. We want to get way ahead of it and make sure there's more than enough capacity to power this first in decades. Net meaningful expansion of the US grid. So I mean, the way we think about it is not just cost, it's not just schedule, it's how do we do these things at scale and allow ourselves to scale up capacity both responsively as we see the demand coming so on short timelines and then in a way that really moves the needle. And so yeah, that's something we think about constantly.

1:18:04

How did that change the way that you think about building the company and your need to go produce way more uranium?

1:19:25

The AI boom makes it even more obvious how Quickly, we need to expand grid capacity. So if we go back to something I said earlier, It's 2010. The US and China were exactly at the same place on the grid. Today they've doubled it. So in about 15 years it is the case that if you really want to, you can double your grid from where we're sitting today. And then if you look at the AI demand, it's dramatic. It's like by 2030, I believe the latest projections are that the AI data center demand for electricity is going to completely consume or equal today's grid eclipse.

1:19:33

Yeah.

1:20:10

And so clearly we do need to double. Can we double by 2030? I think that's a big feat. But what all the data centers are planning is, okay, bringing new transmission online that quickly and going for centralized nuclear plants is going to be too hard. We're going to go behind the meter and so they're going to bring their own power. They're going to be deploying SMRs. I think we see that really ramp in the 2000-30s. It'll take a couple of years before they're through licensing and really scaling. But we're planning for ramping in the2030s. And so as we think ahead to how are we doing things where everything that we're doing is designed to scale. So how do we unlock things like mass manufacturing, all the progress in the last really 30 plus years since anyone has completely re architected enrichment, how do we leverage all those things along with the supply chains for them to unlock rapid scaling next decade Once we're online.

1:20:11

Do you have any really great stories from working at SpaceX that kind of shifted the way that you think to

1:21:04

your question on schedule and how do you drive schedule? The dimensions I mentioned were put things in parallel that can be put in parallel. Question your granularity on timelines, make decisions without belaboring them. Once you're at a pretty good level of confidence, move forward. I think the fourth dimension that may be very hard for people to do sometimes is don't beat your head against the wall if you're working on something. And it's just a huge roadblock to the question of make a decision when you're pretty confident. Also, don't be afraid to go back and make a different decision. Early days when I was there of working on the Merlin engine system alongside some incredible people, just as a kid out of school, a couple of years out of school, I was working on a lot of the structural thermal analysis and design around those engine systems. And I remember one night at the Time we had an ablative chamber. And an ablative chamber is basically a composite wrapped chamber that burns away as you operate the rocket, as you operate the engine. And so this ablative chamber was having a few issues. We were operating at probably higher pressures and temperatures than most ablatives had been run previously. They'd been proven extremely reliable and low cost before. But we were running into these persistent issues with them. And late one night a few people had gotten together to talk about this. And I remember the question came up like, what if we just switch? What if this is a fool's errand to get this ablative working at the thrust levels we need, at the pressures and temperatures we need, what if we just bite the bullet and go as fast as we can rethinking this, accepting that maybe this was not the right path, and going to a regeneratively cooled all metal engine. And that's what we did. And now obviously the Merlin engine and Raptor engines work this way. And none that SpaceX flies are a blade of engines. And so one lesson I learned was, yeah, try and move fast, try and make the right decisions, don't make irreversible decisions, but also don't get completely stuck. And if you're just going down a path that doesn't feel like it's going anywhere, don't have ego around it, like even enforce a meritocracy in your own mind and just go back and yes, there's going to be some moving targets, yes, things are going to change. There might be additional work to do that you're hoping to not do. Maybe you've got to take a hit on, spend some more money to go fix this, but don't get too hung up on things and don't have any sort of sacred ideas that you cannot question. Question everything, question requirements, question the path, and just try and find the best path forward at all times, while obviously trying to somewhat minimize chaos. And so, yeah, this almost gets back to the the founder question that you asked before. Inside a company, when there's big changes and big shifts in direction in response new information, it's going to mean people have to do more work. And ideally that wouldn't be the case. But sometimes the right path is to change and adjust. And I think it can take someone who really has the big picture in mind, whether that's a founder who might be the CEO or a founder who's not the CEO, really saying, hey, let's think more about why we're doing this. We've all been going down this path together, but it's okay to change direction. That was a lesson that I wouldn't have expected to learn, especially given SpaceX being so known for just the classic things. Fast schedule, hard charging, cost. But sometimes you have to actually go back and change the way you're doing stuff.

1:21:11

Another thing that I've heard consistently is that Elon will make decisions, think through a problem, and make decisions that no one else kind of sees. And then it seems like a really hard, like the wrong decision short term. And then later on, maybe even a year or two later, it becomes the right decision. Was there any moment at SpaceX where you saw Elon make a call that everyone else thought was wrong that ended up being right?

1:25:11

I mean, the classic Tesla example is LiDAR versus no LiDAR. And I think for a long time the thinking was, hey, a lidar is only a few thousand dollars, maybe it's tens of thousands, but why don't we just put it on the car, like, make life easy on ourselves? And then years later you realize, hey, it's not a perception problem. We can perceive the environment with cameras. It's a controls problem, it's a logic problem. So that's probably the most salient example where Even I, circa 2010 would have said, why not use lidars? They're here, we have them. Yes, they're expensive, but the cost will go down. And then it's like, okay, but how much will the cost really go down? Why don't we just think through this 10 years out and it's going to be cameras at SpaceX. I think that thing might have been going to Falcon 9 sooner than people would have expected. When I was there, at some points, we were working on both Falcon 1 and Falcon 9, and Falcon 1 was still in the process of having its first successful flight. And yet we were working on Falcon 9. And that was okay from a resourcing standpoint because many of us, we'd worked on the Merlin engine, Merlin engine was working, it was working pretty well. And we knew it would be used on Falcon 9 so that product could be reused. But there was a lot more rocket to develop. And I think at one point Falcon 9 was Falcon 5. And then it was realized, hey, for this thing to really cover the market and to give us the performance we need, it needs to be Falcon 9. And so, yeah, I would say there were probably people in the company who were surprised by that. Like, oh, wow, we were doing Falcon 5, now we're already doing Falcon 9. Is this the right choice? And obviously it was. And So I think it can be things as simple as that of when do you launch your next product line, when do you, when do you take on another challenge, when do you work on something that's much sooner than you would have thought while you're still trying to get the first thing successful, but clearly proven correct. Clearly Falcon 1 was serving what turned out to be a very small market for a very long time. And that market didn't emerge very quickly. And the correct thing was to jump straight to the larger satellite market. And I think there are other examples of this too, even on the tech side, really, that switch from ablative rocket nozzle to regen, you were just adding a whole block in the schedule for its development versus hoping to just get there and then that's turned out to be absolutely correct versus the unreliability that would have come through the ablative engine. And there were other decisions like that too. Like hey, instead of this type of nozzle for the vacuum engine, we could use this other more exotic but really exciting other one. And the decision was no, don't worry about that. We need to, we just need to get this working. And don't worry about that potentially higher performance thing. This one is known to be reliable, we can buy it, it's cheap enough and it's going to get us to where we need to go. So let's do that. And it was just many, many practical decisions like that along the way. I would say many more practical decisions that were obvious when you looked at what the long term mission was versus really, really deep tech breakthroughs. And I think seeing SpaceX succeed over the years, it's been evident that it's just make a lot of decisions that are in line with the core metric that you're going after. And if you stay focused on that, you can do pretty well even to that point. On day one of joining SpaceX, we had an employee handbook. I remember the first page of the employee handbook, I think in big bold letters said this is not a science experiment, this is an engineering problem. To that question of what unconventional choices were made that turned out to be right. I think they either followed two categories. One was those choices were pragmatic and very focused on cost and schedule or category two, those, those choices were, they really looked at the big picture and let's not optimize for a dead end that's going to come two years from now. Like in the case of LiDAR, let's think a decade out. What are the trends? What things can we really bet on that we know are going to be true and yeah, how do we just optimize for that?

1:25:38

While you were mainly focused on Founders Fund, what was the biggest thing that you kind of evolved in your own way of company building and your thinking on company building during that huge 10 plus year stint?

1:29:52

Yeah, when I started at Founders fund, this was 2011. There were a lot of at the time, Founders Fund was really focused on two things. One, back founders unconditionally and two, what things can we invest in that are great that otherwise won't receive funding. So where can we make a big impact both on the team side and then market or sector side or technology side? And so there was this belief that technology was what could fuel the next wave of growth in the US and it was systematically underfunded and that we should be trying to help that succeed through capital. When I started in 2011, the way that that showed up was you would. There were a lot of very early stage engineering driven companies. Hard tech, like engineering driven companies were not nearly as popular as they are today. The focus was really things like software and consumer. Early on, for the first few years I was much more focused on I'd say early stage smaller checks, seed stage checks into engineering driven companies, which did remain my focus over the 12 plus years, but over time realized more and more. Some of these companies work at a much larger scale than you ever think that they will. And the correct thing is actually to focus on the very small handful of companies that are going to be the most successful, that are going to be the most important and where you really feel the most conviction in the team and just work with those companies. And yes, that means you're not getting to invest in a seed stage company at necessarily a really attractive valuation. Maybe you're accepting, hey, I'm not investing in a company that's 10 million per year, I'm investing in a company that's a billion valuation. But what are the odds that this goes 100x? I think these things are sometimes scale invariant where the probability of going from 10 employees to 100 employees to 1,000 employees, you might be more likely to go from 100 to 1,000 than from 10 to 100. And so like looked at in a different way, the conclusion I have is that sometimes it's easier to build a really hard company than it is to build an easy company. At the end of the day it gets back to what really important problem are you solving that's not getting solved? And I think if you are focused on that, if your mission as a company is doing something that's really critical, potentially really challenging, which is why it hasn't been done to date. You can just attract the best people. You can usually get the resources that you need to go do this if you're executing well. And in total, it actually becomes a lot easier than trying to start a small business or something that is much more competed over. And so I think one way that people understand this is start companies in areas without much competition, avoid competition, et cetera. I think the more meaningful way to think about that is don't focus on problems that are already getting solved or adjacent to problems currently being solved that you know are going to be tackled by someone else. Focus on the problems that are big and that are just not getting solved. Dedicate yourself to them. If you do that, it will be very challenging, but it may actually turn out to be easier than starting the next small business or the next small software company.

1:30:03

Yeah, it's funny because you can kind of attract different types of talent, like the very best talent. And from what I've seen and experienced, the very best talent wants to work on the hardest problems, even if it basically gives them less money or these other things, like the enjoyment of working on something that's important and valuable is more attractive to them.

1:33:34

Yeah, maybe the problem's twice as hard, but you get 10 times better people, and so net, it's somehow easier.

1:33:54