Scott Nolan on Uranium Enrichment, Founders Fund, and Reindustrialising America

Scott Nolan

My framework has always been: just do something that's useful. Do something where you feel like you're making a real contribution and using your talents to make a positive impact. What important problem is there that's not going to get solved otherwise, that somehow I can contribute to? I think all three major things I've done — SpaceX, Founders Fund, and now General Matter — fit that in some way.

Taking them one at a time: SpaceX, I was just an engineer coming out of undergrad. I had worked at Boeing during college, saw what the incumbent aerospace industry was like, didn't want to work in that, didn't believe it was going to change anything. I had an aerospace background and always wanted to work on rockets and aircraft. I asked myself, okay, what's the most exciting thing to do? I knew the incumbents were not going to make an impact, but there was this new company, SpaceX, that was going to ultimately own the entire space-launch industry — which I believed even when it was thirty people. It was a no-brainer to go work there right after college. I'd interned during college and seen what it was: these guys are going to win, I want to be a part of that.

That was an industry that had stagnated for decades. No incumbent was doing anything interesting — they were all just riding government cost-plus contracts. The US assumed that space launch was a nation-state capability that would never be commercially interesting, and so it had to be subsidised forever. The result was cost-plus contracts, layers of subcontractors dozens deep, and no ability for anyone to do something really novel. It was going to take a new company.

That led me to SpaceX early on, and then I found my way to Founders Fund in 2011. I was at Stanford in business school, started in 2010, was quickly voted most likely to drop out — I wanted to get to work, I just wanted to do stuff. I thought about dropping out in the first or second month of business school to join Square. Keith, who was at Founders Fund, was trying to recruit me to drop out and go to Square. In the meantime, I met Peter — I was sitting in a class he was teaching at the law school, something called technology sovereignty and globalisation, with readings on theory of government and how technology would change the power dynamics between government and industry. He convinced me to join his startup in the venture capital space.

The basic premise was that VC needed innovation, and the incumbents wouldn't do it. Circa 2005, the concept was founder-friendly: if you looked at the most successful companies, they were founder-run all the way to the end, so the premise was to give founders back control of their companies and unilaterally support them in building that. That was the 2005 genesis of Founders Fund. By 2010, when I was talking to Peter, it was more this contrarian thing of what important companies is no one funding, and how can we be the capital for that.

The thing I focused on when I joined in 2011 was: what set of companies are really promising that people underappreciate? I had just come from SpaceX, which wasn't yet appreciated in 2011 — four years later they were landing rockets and it was obvious from the inside that all that stuff was going to work, but the whole world didn't understand it yet. My thesis was that there was a huge set of physical-world companies, hardware companies, that could be really valuable — spanning biotech, computer chips, satellites, space launch, transportation, infrastructure, almost anything that was not digital — and that this was a huge opportunity area everybody was ignoring. And then I came across the problem of uranium enrichment and the US's total lack of capacity in the space, which essentially forced me to go start General Matter. We'll come back to General Matter in great detail.

Scott Nolan

There are probably many layers to this. Number one was just avoiding trends, avoiding the herd, thinking for yourself. That was probably layer one. The second part was that Peter always took a very orthogonal view to most people whenever we looked at any company. There would be layers of abstraction — instead of just doing a spreadsheet and trying to analyse the investment, why don't we think about why we're even seeing this investment, how we should think about it from a very different perspective than everyone else. Sometimes there would be many layers of abstraction, and you'd end up with a really different view on things. It's natural to just dive in and start trying to understand the business, but trying to develop a very different perspective that would yield some alpha was almost always the approach.

I think around 2010 he was already talking a lot about how we'd made all this progress in the world of bits but not in the world of atoms — you could be on your cell phone and it was interesting, and then you'd look around and nothing had changed in fifty years. I think he was starting to think about which companies were doing this well, and where would someone who could understand the business world, the investing world, and the startup world come from.

Scott Nolan

Some of our hardware investments turned out to be pretty good — I think they exceeded all expectations for everybody. SpaceX, Founders Fund invested in 2008, and what it is today, I don't think many people would have predicted. Maybe Elon could have seen it going to this length — the ultimate purpose is to colonise Mars, so inherently it has to become the scale company to do that. But I doubt anyone would have expected this sort of outcome this quickly.

Scott Nolan

There are two layers of competition, and avoiding competition was a huge part of this. The competition piece is typically understood at the company level: if there's a trend, inherently many companies go after the same trend, you get new entrants, it becomes a thing about the theme rather than one company. And if it's about many companies, how is it not the case that competition drives profits down to the economic equilibrium of perfect competition? So there's that piece — avoid a trend for that reason. But layer two is: if there's a trend, probably many investors are looking at it and pricing it up, and then where's your advantage? You want to avoid competition on both fronts.

Scott Nolan

Unfortunately, or for better or worse, they lead you to a potentially brute-force approach. In the first few years at Founders Fund, I'm looking for great founders, looking for under-explored ideas — it's just lots and lots of meetings. From the investor perspective, unless you have things you're into that you think are underappreciated by the world, that you've maybe been excited about for a long time, and you keep digging on why no one else thinks about this, then it's mostly meeting a lot of people and trying to find what's interesting, what sounds really different, what makes sense.

The attributes are something like: you meet a founder, they seem really smart, they're talking about something no one's really talking about, they're telling you why everyone who's thought about this problem thinks it's impossible or is going about it the wrong way, and if you adjust and come at it from a different point of view it results in a really different solution with really different business characteristics. When you meet a founder like that, usually they're not giving you superficial answers to convince you to give them money. The conversations with great companies always felt like: this person is really into this thing for some reason, and when I ask them a question they don't just bounce back to the surface — they say, here's the answer, here's the next question you're going to ask, let's take you all the way down the rabbit hole. They like showing you around the space. That's how it felt from the investor seat.

If you think about the attributes of industries where this is the case, I think a huge portion of them are industries that somehow just stagnated. The thing most linked to stagnation is being a cost-plus industry, where there's very little incentive for progress, not much incentive to bring the cost structure down, and you end up with a fixed market size that never takes off because you get into a stalemate where all the companies get to oligopoly status. The equation for max profits is just make pricing high enough to the breaking point, collect your cost-plus revenue and margins, and it never becomes a compelling thing. Space launch, to some extent defence — which you see with Anduril trying to break that — to some extent infrastructure, like the Boring Company, which is their prime thesis. I think incumbent, stagnated, oligopolistic, cost-plus industries are prime for this.

Scott Nolan

Airbnb is a classic huge example. When Founders Fund invested, it was still crossing from a weird backpacking, couch-surfing, air-mattress-in-someone's-living-room thing to what it is today. Not that many people were interested in sleeping on air mattresses in people's houses, but that team was really into turning it into something much larger, where people could meet each other and have a really authentic experience instead of just staying in a hotel. That wasn't something many people were thinking about, and yet when you actually looked at it you realised how big the market could be if they could cross over to a mainstream thing.

Sean Parker was on the Founders Fund team right when I joined, and led the Spotify investment. The internal memo on that was so well reasoned, because of his history of understanding music, doing the Napster thing, and then years of trying to find the right company with the right formula — which led him to see the potential of Spotify, why it was the perfect geography to start in, and the perfect licensing strategy. I think it's often a really deep, personal interest in something a founder believes should exist — everyone's done it the wrong way, and here's the right way. Sometimes people sit with those ideas for five or ten years.

Scott Nolan

It probably wasn't the meeting itself so much as trusting your judgement more. On day one it was: I don't know anything, I'm going to take a lot of meetings, some seem good, some don't, but we need to do the work because what do I know? Early on, intuition was all you had to go on, and I think it's usually correct. Then you get a little better at the job over the next couple of years, better at analysing things, and that might actually lead you astray — you start doing the analysis when you kind of already know which ones you like. You should just concentrate into the fewest number of companies possible and not dilute your average returns by indexing. Over time you get better at asking the right questions to harness the intuition — okay, my gut's telling me we should dig into this.

Scott Nolan

Airbnb was one of these. We did a lot of work on it. Founders Fund did a small angel cheque early on, then a much bigger cheque in the next round. At the angel cheque it was still very informal — air mattresses, and I think there had been an incident where a guest completely destroyed a home and there was controversy around that. The company took a hard stance: we will reimburse the hosts, we are professionalising this. That was maybe the moment you could tell it was going to go mainstream, and then Founders Fund made a huge investment. If you just looked at the past you might say this seems like a niche thing, but if you did the work you could see the demographics shifting to a slightly older crowd, not just backpackers out of college as some had perceived, and market share in different markets increasing a lot. We looked at every single market, sliced market share and marketing spend, and you could see they were just taking share and becoming the dominant thing. This is it, they're going to win.

Scott Nolan

The low-competition thing typically would be associated with lower valuations, but I think trying to find value deals in venture is a dumb idea — it's not the right plan. Maybe if you have a very small fund and can pick up some interesting IP, or this company may never be huge but it's a really good deal, it starts to look more like PE. That could be okay, but for true venture I think it's dangerous, because either it says something about the company's ability to raise capital at market prices, and unless this is the last round they need or they're going to be genuinely capital-efficient, that's a risky thing. Maybe the team is incredible at debt financing but terrible at venture financing, so they'll switch to debt — you could imagine that, but typically if you're meeting a company and it's a crazy value deal, it's probably just not going to end up being that good. That's what I've observed empirically.

Scott Nolan

Yeah, that may be the case — the whole strategy is probably more concentrated into the winners. I think the way that can still be a good strategy even if the company's popular is: A, it's popular but not as popular as it will be; or B, this idea that up-rounds are almost like anchoring on the past versus looking to the future. Peter's said this a lot and has guided the Founders Fund team to think this way, more so in the early 2010s. If you've got a company growing steadily and there's a big up-round, it's probably the case that that up-round isn't even enough — if it's a 2x up-round, maybe it should be a 4x up-round.

Scott Nolan

People are anchoring on the past — last round was this, I guess it should be reasonable compared to last round's price. In reality, all that matters is the next round's price: how do we make sure the next round's an up-round, what are the catalysts for further increases? You don't get paid as an investor based on how close you were to last round's price — it's ultimately against exit price. But the only data you have that's empirical is the past, so people anchor on that.

Scott Nolan

From the investor seat, I don't think you want to be in love with an idea — that's risky, because you'll try to find a way to express that by investing, and you'll maybe make compromises on the team because the idea is just so good, its time has come. But if the team isn't there, ninety per cent of the time it doesn't work, and then there's the whole horse-and-jockey question of can you swap out the team. The whole Founders Fund thesis was always no, you need the founder to run the company, to have the vision and see it through. So from the investor side, being in love with an idea is dangerous — it can cause you to make compromises that come back to haunt you, and to put good money after bad despite the writing on the wall.

But on the company side, you have to be in love with the idea. It's not that rational to start a company — there are easier, more comfortable ways to make money if that's the goal. Smart people who want to make money have plenty of good jobs in finance; people who just want to build have lots of places to build. But if you want to actually start a company, it had better be something you're really passionate about, or you think the problem is really important.

For me, enriching uranium was never something I was specifically excited about — I was always into nuclear energy, always thought it was a no-brainer. Probably the two industries we were supposed to have, from sci-fi from the '60s and from what our country thought — we're going to be doing things in space, and we're obviously going to have nuclear energy. We went from burning wood to chemical bonds, and now atomic energy was clearly so much more energy-dense, powerful, and should be lower cost. Those were always things I was excited about. I never had a specific interest in uranium enrichment, but through the course of investing at Founders Fund — from 2010 looking at all sorts of hardware companies, first investment was a satellite company, Planet Labs, then a lot of things outside pure software, and in the last couple of years drifted back towards almost pure hardware and then really energy.

Scott Nolan

We invested in Crusoe Energy and understood the whole stranded supply of flare gas and what you could do with that. Then invested in a company called Radiant, which was the inverse — stranded demand. How could you serve that demand, maybe with a small microreactor? Even if that microreactor's output was expensive, the stranded demand had to pay crazy rates anyway for diesel generators in a remote Alaskan village, or an army base. That's a good starting point, fitting the whole Founders Fund thesis of start with a really small market and grow into a bigger one — don't worry about your TAM, worry about owning that market and then grow from there.

My path to understanding the bottlenecks in nuclear energy came from having invested in Radiant and met so many other advanced-reactor companies along the way. All of them said the exact same thing: we're going to make nuclear affordable, we're going to make it scalable, we're going to take this from huge construction projects to factory built. And yet the hardest thing isn't licensing — everyone thinks the NRC is impossible to get through, but no, it's not that. They told us: we cannot get the fuel. The fuel comes from Russia, only Russia makes it, we have to import it. That's quite challenging, and this was even pre-Russia-ban. We just need some source of fuel.

Scott Nolan

I spent all of 2023 looking into this, trying to understand: of the five steps of making fuel, which one — or is it all of them — is stopping this? Is there not enough uranium? Is it something about the process? It was the enrichment step. I looked for a company in the enrichment step to debottleneck nuclear and get the nuclear future we want, and could not find anything for an entire year. Finally I decided if this is going to exist, it's got to be a new company — not an incumbent, not a government, a new private company. This was the important company nobody was building, the important problem nobody was solving that I could somehow contribute to.

Scott Nolan

My experience through SpaceX was that the government wants to work with private industry and wants to solve problems. At SpaceX I worked on the engine systems under Tom Mueller, did a lot of the structural and thermal work on those, made sure they'd stand up to the environments and be low-cost and high-performance — that's ultimately what we were optimising for. Once those engine systems were working well, I moved over to the Dragon capsule. On Dragon we were on the NASA COTS programme, Commercial Orbital Transportation Services — a multi-hundred-million-dollar programme to bring back the capability of launch to the space station, first a cargo vehicle and then ultimately a crewed vehicle that could dock with the station.

The last year or two I was there was focused on Dragon, working very closely with NASA, because we were going to dock with the most expensive asset mankind had ever developed. The last thing anybody wants is any sort of issue with a private company's cargo capsule docking with the space station — colliding with it, or worse. Actually, given how orbits work, there's some ability to avoid a collision even if things are off track. The harder thing to get a handle on was thermal and pressure — you've got this vehicle with solar panels, what's the heating on the vehicle, what thermal load does that drive back to the space station, can it handle it? There are all these interface requirements, and we were working closely with NASA on those. These were incredibly smart people who believed in space, who had been working at NASA for decades despite not a ton of growth in space activity — they were there because they believed it and loved it, and the opportunity to work with a company was: how do we get you to the next step on the milestones, how do we collaborate to make sure this is safe but that it actually happens. Super collaborative, very positive. My takeaway was that in industries where you have true believers in the government agencies who've been doing this for a long time, there's a lot of openness and excitement to working together if the company is credible and cares as much about safety and performance as the agency does.

Scott Nolan

There's all the classic things — the Elon algorithm, though it wasn't explicit back then. It was: we have to get this rocket launched, let's make sure it works, let's not over-optimise or have analysis paralysis. Define the goal, come up with a good solution, this looks like a good solution, that's a good plan, let's run with it. Don't deliberate for months over things — decide and move forward, and if we're wrong we can always go back, because there was extensive testing along the way. Use good engineering principles, think from first principles, move fast, get it to ninety-five, ninety-nine, ninety-nine-point-nine per cent, get it operational, and you can make it better later. But if we never launch, if we never get operational, this isn't going to work.

Scott Nolan

There's the onshoring piece, but it's maybe even simpler than that — it's vertical integration. Even domestically, you have companies in the nuclear space, the aerospace industry, defence, many sectors where subcontracting is the norm — you subcontract a subsystem to somebody who subcontracts a component, and that component has different inputs and they subcontract that all the way down. In aerospace, at the SpaceX days, there were something like thirty layers of subcontractors in one system — I think that was the case for the space shuttle. In the nuclear space recently, there was an analyst call where a company was bragging about having something like nine hundred subcontractors, so many that they needed regional organisers of the subcontractors.

On one hand, clearly what you're doing is really complicated and there are deep barriers to entry, but every interface that crosses another company is typically fixed and won't move quickly, so you have to treat it as fixed. Everyone designs their piece against preconceived interface requirements, and you end up with a calcified architecture at the system-design level — you can't optimise across layers for the overall goal. If you bring all that in-house and don't have a lot of subcontractors, you can, as an engineering team, optimise and trade off interface requirements with every iteration — hey, this thing you asked of me over in the electrical team is really hard for me on the mechanical side, do you mind giving me a little breathing room and I'll make it up somewhere else. You can have those conversations much more easily when two people are sitting side by side at a desk than when you have two separate companies across the country. At a minimum, you need to pull the engineering together, vertically integrate, and sign up for doing more of the engineering yourself rather than outsourcing it.

Part two is manufacturing. If you're actually going to make the thing and you're designing for manufacturing, you want your manufacturing co-located with engineering. That's what all the great hardware companies do now — at least your first-of-a-kind manufacturing co-located with engineering under the same roof, so you can talk out why something that was just designed is hard to make. If you can make it a different way, maybe you don't need a six-axis CNC, you can use a laser cutter and get ten times the throughput at a tenth of the cost. So you need at a minimum to put your first-of-a-kind manufacturing co-located with engineering, and then small-scale and mid-scale manufacturing too — maybe large-scale you can push somewhere else, to a lower-cost centre than Southern California, where so many of these companies are. But bringing it all together is mandatory.

Scott Nolan

There are two of my favourite charts on this topic. One is GDP per capita versus energy consumption per capita — the R-squared is probably over 0.8. You cluster every country on Earth, plot GDP per capita, plot energy consumption per capita, and there's a very obvious line through them. There are outliers, but it's so predictive. Energy use and production is the ultimate proxy for human prosperity, for economic activity — I think this has been pretty well understood for a while but completely under-discussed.

The other chart I always look at is the US grid: we grew it for a long time, and then starting around the '90s, essentially no growth until today. China was growing a long time — around 2010 we were neck and neck, and I think this year they'll be triple the US on total energy production. The US just needs to do something if it wants to continue to be relevant economically. You need more energy production if you want to grow your economy, and purely outsourcing stuff overseas just gives up capability, especially as it involves new manufacturing techniques, scaling manufacturing rapidly — all these things that are so important.

Scott Nolan

It's curious to imagine — if I air-dropped three times the US per-capita capacity of energy on some random country, would that country necessarily become prosperous? The argument from the '90s almost till 2020 was something like: even if it would increase the economy, would you want to? Do I want dozens of aluminium-smelting facilities in my country, or is that better done overseas, low-margin and not interesting, and why do you have people working on that when they could be doing more interesting things in services? That was certainly an argument, but I think we're seeing right now what happens when you're not proactive about it, when you don't have the capability to expand rapidly — you don't want to be caught flat-footed when there's large demand for energy production, and now we have it with AI and data centres. Some of these curves look like data centres could consume the entire grid by 2030 at this growth rate if allowed to.

In theory you'd say energy production will come online as we need it. But in practice, if you don't develop it, you may bias yourself towards other things because our energy costs are kind of high compared to other countries because we're tighter on supply — and they've over-built and are subsidising, but then they take those industries, take the front of the supply chain, and start marching forward to where they can dictate a lot about the economy. So I think that was the classic view — economies are efficient, markets are efficient, if there's demand for energy we'll bring it online when we need to. But physical-world stuff has a timeline, and in the US it can be very hard to get things permitted. So when there's unexpected demand that increases very rapidly, you might be caught completely off guard — which I think is the situation in the US now with data centres, scrambling to find power. A few years ago, if you looked at where a data centre would go, the answer was wherever there's stranded power — stranded wind in West Texas, stranded flare gas in the Dakotas. Now those stranded assets are gone, and it's time to build new capacity, and data centres are backed up on getting natural-gas turbines, and nuclear takes a few years to install. From an energy-stability, economic-robustness point of view, you want to have the capability, at least the ability to stand it up as you need to.

Scott Nolan

The state's been fine until now, and the direction's completely flat. The issue is we haven't increased at all in decades on the supply side, so the direction doesn't feel great. On the state, I wish we had more — I wish we weren't a third of China. I think we could be a lot more if energy weren't just abundant but cheap. Just making more high-cost energy also won't bring back certain industries, won't cause us to start doing things we used to do.

Scott Nolan

Labour is actually the one that's probably most responsive — if there are really compelling jobs in a field, people will shift over to it. The classic example the last couple of years has been electricians — massively undersupplied, can't get enough to build all the data centres, commanding very high wages, can make more than people who got a master's. It's like an incredible career, and now you're probably going to get a lot of people shifting into it. I think labour responds — it probably takes a few years, but it's faster than building infrastructure. So infrastructure is probably the bottleneck, and it's probably not just energy — probably permitting too, which can take a while. I think there's a bunch of different bottlenecks to getting things online.

Scott Nolan

Statistically, it's always been the safest, cleanest base load. For the economy, what you really want is base load.

Scott Nolan

Always on — very reliable supply of electricity, not intermittent, something you can actually design an industrial process around as the foundation of the economy. Even data centres want base load. Solar with enough storage could be base load, but typically it's not anywhere near that amount of storage — usually it's more intermittent, more for peak shaving. If you really want base load, something businesses and industry can rely on, it has to be on almost all the time, highly reliable. Nuclear is the most reliable — it checks the base-load box.

People have only started caring about climate as much as they do now in the last ten to twenty years, and if you care about carbon, you probably also care about particulates, and nuclear is the cleanest by far on those dimensions. So on the environmental side, nuclear wins for base load. On safety — this is the part people question — it's the safest by far of base load too. A fossil-fuel plant has a known impact on human wellbeing from emissions; nuclear has none of the carbon emissions, and there have been high-profile accidents — Chernobyl, Three Mile Island, Fukushima — but the actual risk associated with nuclear, despite those incidents, is still far lower than any other form of base load energy. If you ask me what's the safest, cleanest form of base load, it's nuclear, absolutely.

What about cost? This is where it gets exciting, because over time we did less and less nuclear, probably starting in the '70s, and the cost of nuclear went up as we did less of it. Today it's more expensive than fossil fuels. Unless you really care about safety or carbon emissions, are you really going to do nuclear when it can take ten to fifteen years to build a reactor and it can be double the expected budget? That's really hard for a utility to stand up to and say I want to do more of this — they need to bring cost down for ratepayers, need predictability in when they'll bring capacity online. So on those dimensions, nuclear hasn't been the best — it's not highly responsive new energy generation, and it hasn't been the lowest cost.

So why care about nuclear? Because on first principles it should be one of the lowest cost. You've got much more powerful physics, potentially much smaller reactors outputting a lot of energy — a pellet of nuclear fuel is equivalent to a ton of coal. I can take all that space and package it into so much smaller — I don't need as much material for my reactor, I'm not going through as much mined product, it's an order of magnitude different. On first principles, nuclear should be an extremely affordable thing, potentially much more affordable than fossil fuels. It hasn't been at all, but that's the goal — and now you see this whole wave of advanced reactors trying to make that true on the reactor side, and we're trying to do it on the fuel side.

Scott Nolan

On the reactor side, everyone thinks of the cooling towers — that's all steam or hot water cooling off so it can be reused as coolant at the start of the cycle; lots of industrial processes do that, but nuclear is known for it. That's typically about a gigawatt-scale reactor, often in the US an AP1000 Westinghouse gigawatt-scale design. The thing to think about is that's a big reactor. I think the interesting buckets are around size, because size links to applications and markets. Gigawatt scale is for the grid — you need that scale to move the needle and bring cost down as low as possible. On the other end, you have microreactors: a stranded community running diesel generators, both bad for the environment and not cheap — we can beat them on cost, so you go microreactor, maybe one-megawatt scale, not gigawatt, megawatt. And then the middle, a hundred to three hundred total megawatts, SMR — small modular reactor. All three buckets have different technology approaches, and I think all three are going to be important. Gigawatt scale is for the grid, megawatt scale is for remote communities or government installations, and the middle scale I think will find its niche behind the meter with data centres over the next five to ten years — one unit dedicated to one data centre, cordoned off from the rate payer in whatever district it's in.

Scott Nolan

Exactly. Think of it as the ultimate behind-the-meter thing — an island. Imagine a fence around a few hundred acres, you've got a data centre and its nuclear power plant, or peaker plants, natural-gas turbines, maybe even solar. That's all cordoned off — it does not touch the grid, doesn't impact rate costs for communities, completely separate. I actually think there's a huge opportunity to improve communities through this. If a data centre is putting billions into power production and compute, even fifty-fifty, can we increase that power production ten per cent, which is only a five per cent project increase, and bring, in the case of a gigawatt-scale data centre, an extra hundred megawatts to a community? That's huge — it would plummet utility prices for that community. At a minimum we're going to see people going behind the meter, and ideally we see them actually feeding the grid and feeding base load.

Scott Nolan

The concept was: we're doing all this investment in the grid for the first time in decades, private companies are making the investment, and they'll make that investment with bring your own power, bring your own energy — BYOE. Their biggest worry is probably will I be allowed to build this data centre in this community. If they want the right answer to that, they can tell a community: we're going to be over here on unused land, are you okay with us running a bunch of compute and being net positive to your grid — can we just plug in and give you some extra power in return for inviting us to be a neighbour? To me that's a complete no-brainer for everybody. At a slight cost increase, during a time when hyperscalers building data centres are all about speed to power because it's a contest for who can get to the greatest scale first, that's a very low cost compared to the advantage of deploying quickly in a community and being invited there versus being rebuffed. And modern data centres can do completely closed-loop cooling, so the water issue isn't an issue — it's really just this power thing.

Scott Nolan

Number one is they don't have fuel to operate. That's a showstopper — we're trying to eliminate that risk. Part two is they're too expensive. I don't think the physics will be a problem — you have dozens of incredible companies with great engineers trying to solve the advanced-reactor problem, and typically they'll pick a form factor of the right size for a certain market, whether microreactor for remote, SMR for data centres, or very large gigawatt scale for the grid. SMRs I think ultimately have the plan of being cheaper than gigawatt-scale construction for the grid, but there are many reactors trying to solve the cost problem, and the question is whether their energy-production cost is low enough. For these advanced reactors it's not just a fuel-availability question, it's a cost question — the fuel can be half of their total cost, and long-term I think enrichment will be half of the fuel cost.

Scott Nolan

This goes back to understanding the fuel supply chain for nuclear. There are five steps. You mine uranium out of the ground, convert it into a gas, enrich it, deconvert it back into a solid, and form fuel pellets. Enrichment is really a refining step, a separation step. Five steps to go from ore in the ground to a little pellet, into a fuel rod, into a reactor. The US has good capability across the board except in enrichment — that's the one area we have no commercial-scale capability in. There's some R&D capability, but nothing commercial-scale or commercially competitive with Russia and Europe. So we decided to hone in on the enrichment step — that's the bottleneck, the thing leading to what we call the nuclear fuel cliffs, of which there are three.

Cliff one is the HALEU supply chain — the same problem we heard from all the SMR companies. HALEU is uranium enriched to about twenty per cent U-235, the fissile material — enriched a bit higher to help the reactor be smaller. There's no reliable supply of HALEU in the next few years, which will make it very difficult for advanced reactors to scale up. So far they have small amounts from the DOE to prove their reactors work and do first deployments, but not to scale to where it moves the needle on the grid.

The second cliff is 2028, when the US ban on Russian uranium imports goes into full effect, 1 January 2028 — an act of Congress means we'll no longer be allowed to import Russian uranium.

Scott Nolan

We will not be importing enriched uranium from Russia, or uranium from Russia. The only place we get it enriched is Europe and Russia, and today Russia is about twenty-five per cent of US imports — so twenty-five per cent immediately goes away in 2028. Utilities in the US will start eating into their inventories and trying to work out deals to import more from Europe. Cliff one is HALEU, cliff two is LEU — low-enriched uranium at three to five per cent. And then at some point in the future, the US stockpile of enriched uranium for Navy propulsion eventually runs out too. We're focused on the nuclear-energy side, LEU and HALEU, and that's what we're going to be producing.

We realised the most urgent cliff to address as soon as possible is HALEU for advanced reactors. Coincidentally, this is the small, emergent market that we can go after that we don't think any incumbent will go after on the same timeframe as us, and we can serve the advanced-reactor customers we've known for years and make sure they have the HALEU they need to deploy and scale. That's where we're starting, and phase two is going to LEU production, for the ninety-four reactors on the grid today, which today in the US is a two-to-two-and-a-half-billion-dollar market.

Scott Nolan

All enrichment is for the most part the same — it's a refining or distillation process, you're just refining further. The product we make is enrichment services, and you can run material through from natural uranium to any arbitrary level of enrichment. The commercially relevant ones are LEU, three to five per cent for existing reactors, or HALEU, twenty per cent for advanced reactors. When we set out to do this — pre-Russia-ban, so there was no LEU cliff yet — we said HALEU is what the market needs most right now, and we're going to develop enrichment capability to serve that market. The fundamental unit of enrichment capability is the ability to do this refining, measured as separative work units in the field. We provide that service and sell it to utilities and advanced reactors to get them the fuel they need. Ultimately the product we developed is enrichment, applicable to either level.

On how that's different from weapons grade: when countries go to weapons grade, they're trying to go north of twenty per cent, the internationally agreed limit for non-proliferation reasons — there's strong international consensus that we should stay below twenty, because the risk isn't worth the reward to let people go higher. You see countries going to sixty per cent saying it's for their nuclear energy industry, which is pretty suspect — clearly they're trying to get as close as they can to weapons grade, well over ninety per cent, and trying to develop weapons. It's not different technology, but it's applied in a different way, in a completely non-international-consensus setup.

Scott Nolan

Yeah. There's U-238, there's U-235 as it comes out of the ground, and other isotopes too, but the main one is U-238, and you basically want to filter that out to get U-235 — the fissile isotope, the one that wants to react.

Scott Nolan

We've certainly got hundreds of years of supply in the ground. The US has supply, we have active mining in the US. Canada has even more supply, higher-grade ores — enough to supply us for a very long time. Same with Australia. Kazakhstan has a huge deposit, and the US imports from there as well. On the ore side, it's not an issue.

Scott Nolan

I wouldn't say it's easy at all. The latent demand hasn't been obvious for a long time, until the last couple of years — any large latent demand in the last twenty-four months is linked to the thought that there's going to be a lot more data centres doing AI inference or training. Before that there wasn't a lot of new nuclear being talked about — there was need for HALEU, but that's a pretty small market. If you go back to the 2010s, there was really no new capacity being built anywhere, certainly not for LEU or HALEU production, since it's been so recently emergent. So the fact that this is a growing market — we're still betting on that in many ways. It's not the case that we're currently deploying tons of new reactors in the US, but we think that's coming, and we believe in the market deeply. I don't know that the entire industry believes the demand is there — they've been through nuclear renaissances before that didn't pan out, and their point of view is we'll build it when we know there's a need. At least that's been the attitude of many people we've talked to.

On the ease of doing this — it's a proven thing, people have done this before, but it's still not easy. You see some of these retro technologies that were done in the past, and then we lost a lot of the capability because we didn't do it for so long. For enrichment, it's been an industry without a lot of change since really the '90s — not a lot of progress, not a lot of leveraging new technology. There's a lot of hard engineering work to get back up to modern speed, and then even once that's done, you're building facilities that are a million square feet, large, multi-billion-dollar infrastructure projects. In the same way that maybe building the first Tesla Roadster was hard, the real challenge was how do you scale this up. I think it's extremely underappreciated how much goes into standing up a huge industrial facility — I think that's actually the hard part.

Scott Nolan

They had cost per kilo to orbit, specific to a particular orbit, and you can actually decompose the orbital piece into a velocity — almost cost per kilo to some velocity. Our version is cost per kilo of uranium, not of a payload, to some enrichment level — cost per kilogram to three to five per cent, or to nineteen per cent — which can be described as cost per separative work unit, where a separative work unit is the industry measure of enrichment, basically kilos times some entropy reduction or separation of the material. It's typically referred to as kg SWU, kilos times separative work units. Our North Star metric is dollars per kilo SWU.

Scott Nolan

It is, but it depends on what type of reactor you're running. If you're doing a gigawatt-scale classic AP1000 light-water reactor that uses three-to-five-per-cent enriched LEU, the cost of that fuel as a percentage of overall cost is quite low — it doesn't really matter. What you want is availability, reliability, diversity of suppliers, knowing they'll be there, not overly concentrating and creating supply risk. Most of the cost is CAPEX for the huge, ten-billion-dollar project. But for advanced reactors it matters a lot — for some, the cost of the HALEU fuel, enriched further to 19.75 per cent, which requires more input material to refine all the way there, can be more than half of their energy-production cost.

The way that fuel has been purchased in the industry so far, mapping directly onto our North Star metric, is five different services. A utility purchases uranium and then purchases upgrades to it as it goes through the supply chain — it's really a tolling business, they own licence to the material, it's book transfers all the way through, and they pay different people per unit of service provided. The service we provide is priced in dollars per SWU, and if you look at the actual cost structure of producing fuel, for low-enriched, enrichment is maybe a little less than a third but one of the largest cost segments; for HALEU it's even more, and we think it will be the dominant cost driver of fuel cost.

Scott Nolan

If you want to put it in investment or business-strategy terms, that would be it. But I think if you go back to the very beginning, it just turned out this was the thing that was most necessary — this step of enrichment was why we didn't have the more enriched fuel advanced reactors needed. You realise the US had lost the ability, despite being number one in the world during the '80s by far — it was something we were extremely good at, and we completely stopped doing it, and this bottleneck was going to be the bottleneck to all of nuclear energy if you believed in advanced nuclear. This is the thing we had to solve. We did it because no one was doing it and it was extremely important and urgent, and I think that ended up corresponding to a very good entry point for the market. I don't think it was coincidence — if you're solving a problem that's urgent, that no one else is doing, on a small but emerging market, it will completely fit that framework.

Scott Nolan

It comes back to the team. Naval has said this a bunch of times — the team you build is the company you build. To us it's team DNA. We thought about where to put the company. You look at who you actually need on the team — of course nuclear engineers, but also great mechanical, electrical, software, chemical engineers, every type. So where do you put the company? It turns out the percentage of the team that's nuclear engineers is single digits — we're not doing a reactor, there's actually no nuclear reactions in our process. We need to make sure there are no nuclear reactions, but you don't need a huge number of nuclear engineers to do that. So there was a question of do you go to where the nuclear engineers are, or where all the other types of engineers are. On the nuclear-engineer dimension you don't need that many, though you need great ones — and they're scattered everywhere, there's no one place to go. So we went with the other option: go to where all the other engineers are, specifically hardware and aerospace, which is Southern California.

The team DNA we wanted was: this is an engineering-driven company. We're not doing a science project, no new physics, no multi-year uncertain R&D path — we need to get this operational as fast as possible. Everything has to be thought of as engineering: engineering cost out of the system, engineering performance up, engineering cost of capital down, engineering schedule to be as tight as possible. We're even engineering our own buildings — taken from the Tesla playbook, don't hire a GC and outsource everything, you lose total control, the schedule's not in your hands. You need to build your own in-house engineer-procure-construct firm to run your construction projects, because one of the hardest things about this isn't just engineering or manufacturing but actual construction of millions of square feet, on schedule, on budget. The whole DNA of the company is oriented around delivering for the industry as quickly as we can while being safe and reliable — go quickly because the industry needs it, don't have analysis paralysis, get a service live that can deliver for the US industry, then bring the cost down over time. Everything's oriented around schedule and cost.

Scott Nolan

It changes over time. My strong preference is internal-facing, to be there helping do the work internally. Right now we're in a phase of company-building and team-building, so a lot of my days are interviewing people — I'm the last step in the interview, and I'm screening for whether this person, beyond having the skills already screened for, has the attitude that they care about this problem, that there's probably places they could work on an easier job with a more predictable schedule and higher salary, but we're trying to do a mission that requires us to go as quickly as possible, to deliver before end of decade for both the advanced-reactor industry and the existing reactors on the grid, which is twenty per cent of our grid. That means a lot of late nights and weekends, working as hard as we possibly can. Is that something they've done before, something they like doing, do they know what they're signing up for, are they really motivated by this — those are the things I screen for, subject to everything else already being verified. A lot of my day is making sure we have the right early people to build the right culture, so they can then go recruit and hire the next people. It's not just the team you build is the company you build — the early team you build is the company you build. That's what drives a lot of my external-facing work: getting the word out to the right people, so they know what we're working on, why this is important, why this is the bottleneck to scaling nuclear and energy in the country, and why they should join our team.

Scott Nolan

You're pointing at the difference between acute and chronic events. Acute events that are attention-worthy, people remember; low-level background things, people don't think about, even car crashes and all these things. You can think of it in terms of catastrophic events, or in terms of your health — what are we chronically doing that's unhealthy versus acute things that take their toll. People have a strong bias towards the acute, but often the right thing is to look at the data. I don't think most people find that compelling, though.

The more compelling thing for me is: think of a world in which we are not constrained by energy, because it's good — base load, very affordable, brings down your rates, safe, clean — and let's imagine that future world where we have all of that. I think the way to make nuclear really compelling is to check the final box, which is cost. Why haven't we had more nuclear? The schedule to bring the large gigawatt-scale projects online has been uncertain, costs are often above what's projected, and when you look at the total package it's not cheaper than coal, natural gas, or hydro — so why do we want this? If you told someone, try nuclear, it's safe, we can dig into that, but if they're not interested because it's more expensive — why would they be? But what if it's cheaper? What if your utility bill got cut in half? I think people would suddenly find that extremely compelling, for something that could be tens of miles from their house and still powering the grid.

Scott Nolan

It's all case by case, even for Founders Fund — it's not a programmatic thing they've tried to do, it's almost the opposite. Founders Fund recruits people who want to be investors, not entrepreneurs — if they want to be entrepreneurs, they should go be entrepreneurs, because the two roles are extremely different, and people who've done both realise how different operating is from investing. So Founders Fund explicitly selects for people who want to be investors. And yet once in a while a company will start because it almost feels like a disservice not to. In this case, over a decade-plus of investing, seeing all the reactor companies, they all see the same problem — no one's doing it. You look at the final thing: important problem, no one's doing it, maybe I can do it. You look at what it takes, and you realise, wait, my background is really aligned with that. If the goal is to have an impact on the world, and you can have positive impact through investing, but a much greater impact through starting this specific company, then it's actually wrong not to go do that. It's almost that there's a desire to not start companies, full stop, and only with extreme exceptions will we start them — so if it's an extreme exception, it makes sense it doesn't happen that often.

Why don't most people do that? I think the investing life is probably far better than the operator life, in terms of quality of life — any entrepreneur would tell you that. Once you start a company, you're really latching on to the harness, and whatever it is, you have to go do it, it's not optional. If you're an investor and a deal comes across your plate but you're a little too busy, it's easy to let that one go. When you're running a company, stuff comes up that you have to deal with or it's going to be a problem. That's also a function of the time horizons of each job — when you're operating, you see the feedback cycle very quickly. As an investor, the company you might have chosen not to meet with, to preserve quality of life, could be the next great company, and you might not realise you're not doing well until five years from now when you don't have that return. Some people are just wired a certain way — even in the investing role, I would take ten meetings a day, to the point of definitely diminishing returns, but wanting to work hard and see everything.

I think the other thing about investing, properly understood, is that you don't actually need to invest in that many companies. You can brute-force it and boil the ocean trying to meet everyone, but all that ultimately matters is a few companies per vintage, and you just have to be into those. As a VC, much more than private equity or hedge funds, you don't actually have to work that hard meeting everybody — you can dive in as much as you want into specific companies, add value, help them, be their preferred investor. It's not a job of labour, it's really a job of ideas and thinking. The quality of life of being surrounded by smart people talking through ideas can be a lot more comfortable than rolling up the sleeves and diving in.

Scott Nolan

Maybe the recent example, completely linked to this conversation, is the transition from Founders Fund to General Matter. You have to ask, okay, what was the thread that led me here — and it goes all the way back to being recruited into Founders Fund by Peter, and support through a decade of investing, and then, as I wanted to shift focus to this, total support for doing that. Obviously Peter tried to beat up the idea, make sure it's good, but being along for that journey and then ultimately joining our board, as one of very few boards he's on, I really appreciate that.

Scott Nolan

Probably all the abstract, meta-level questions — even going back to the fact that we haven't had any new nuclear in a long time. Why is that? Is nuclear just regulated to death, is the regulation a thing meant to basically make it illegal? All these questions of why is it really the case that we're going to get more nuclear now. That challenging really forced us to ask a lot of hard questions we feel great about our answers to. Even when he asked, are you dependent on every SMR succeeding to make a real market — on the HALEU side, yes, HALEU production ultimately will be SMRs creating that demand, but to make HALEU you have to make LEU, and our technology works on LEU too, and we'll be building LEU capacity. There's a two-billion-plus US market, and a similar-sized market with our allied partners, that we can sell into. So there's this known, good market, and you work through questions like what if nuclear doesn't grow — well, worst case there's an existing market, we can start a business there, and time is then on our side for when and if nuclear grows, which we think it will, very rapidly, and we're in a good position.

A lot of these conversations with Peter were in 2023, before the AI data-centre boom, so now it's extremely obvious why we need this — and we're in a good position.

Scott, this has been so much fun. Thank you so much for your time.

Scott Nolan

Yeah, thank you.