Lux and the Art of Startup Maintenance

Danny Crichton:
So last quarter, we had a theme in our quarterly letter around maintenance, which I think for most folks, people don't think of maintenance as something that's very exciting. We think of the bold frontiers of sciences, rocketry, putting a space telescope into orbit, deep-sea mining, nuclear fusion, all these frontiers we've never done before. And so when you say the word maintenance, it really triggers us like, "Well, that's boring. That's uninteresting. Oh, we've done maintenance for thousands of years." Why is that a theme for us now in 2024 when, just as we're recording this, SpaceX launched a rocket into space?

‍

Josh Wolfe:
Okay, a few things. First, there's a timeless aspect of that, and that is rooted in physics, which is the idea of entropy. Maybe we'll go to that in a moment. Second is, years ago, I remember seeing a banner that Walmart had hoisted up on one of their factories at a time when Amazon was ascendant, and it said, "You can out-Amazon Amazon," and it sort of dawned on me, in this low interest rate environment that we have come out of of a decade of an excess of excesses, that low interest rate serves as a tractor beam for these far out 20-year ideas that become these 20-month frenzy projects, and you do get not only rocketry and some of the amazing things that we fund, but hyperloop, and then you also get things like quantum computing and fusion, of which I think a vast majority of those things are fraud.
When you get those things and everybody is racing, and when I say everybody, a substantial number of venture capitalists have reinvented themselves from SaaS and enterprise investors into deep tech investors, and they're racing to the future. So it's sort of like you mentioned Dante's Inferno before, you know, Dante, like everybody's rushing. And I started to look at this crowd of people that are going after the stuff that in many cases us and a handful of others have been funding for a long time, and abundance and scarcity are always the key to anticipate returns, and when there is a scarcity of the investors that are going after the future, it's a valuable thing. But when there's an abundance of people that are going after the future, suddenly you start to look and say, "Well, what are people passing by?" And what they're passing by is they race so far into the future having forgotten that interest rates have risen, that the cost of capital is higher, that the incremental growth investor is more scrutinizing is maintenance.
So the sexiest thing that I started to have an idea around in this contrarian idea is while everybody's racing to the far future, the stuff that's here and now is unattended. There are trillions of dollars, I don't say that exaggeratively, trillions of dollars of assets in infrastructure, shipping, electricity, power production, software itself, of the maintenance of software, and all of these things need to be maintained. If you think about it in just simple finance or accounting terms, think about something like CapEx. There's two elements of CapEx, growth and maintenance. Everybody has been funding growth, growth, growth, growth, growth. Who is talking about maintenance? So the scarcity of the discussion around maintenance is what got me excited to say, "Wait a second, there's opportunities for incredible new technologies that can maintain incredibly important and valuable old technologies."

‍

Danny Crichton:
You look at enterprise investors, one of the best parts of, I think, our position here is watching folks who used to do SaaS go into quantum or fusion and apply SaaS metrics to these industries where you didn't have kind of the capital commitment to build those SaaS companies. But you go into fusion, you're talking billions of dollars to get to maybe a scientific experiment which may prove net positive for energy production, which might be able to be commercialized 20, 30, 40 years down the line.

‍

Josh Wolfe:
Well, you know, Danny, the LTV to CAC of a femtosecond quantum annealing is... No. It's true, it's true.

‍

Danny Crichton:
And so, obviously, with interest rates rising, the concern is obviously now we don't have the ability to replace.

‍

Josh Wolfe:
Yes.

‍

Danny Crichton:
And so much, particularly if you look at software, the argument was, "Let's go in banking." You get rid of the infrastructure, everything's built on COBOL. But the mentality of the last couple of years was like, "Well, let's throw everything away in terms of software that was produced in this industry over the last 40, 50 years, all this old code that's been reliable but is decades past its prime with new code," which was cheap at a time when engineers were plentiful, interest rates were low, you had the ability to just do this capital improvement. Now, people are looking back, CFOs, and are saying, "Well, look, this is working. What can we do to maintain it? It saves money, it fits better into our capital profile given the interest rates in the market."

‍

Josh Wolfe:
So you hit on something, by the way, both in COBOL, and I want to talk about Fortran on the software side, but one of the greatest capital allocators of all time, Warren Buffett, when interest rates would rise, he would start to look at these systems exactly as you were saying, that the replacement costs would be much higher. So rail, and trucking, and heavy asset industries, these were interesting. Why? Because the incremental entrant into that field was suddenly going to have to pay a lot more money. So the people that had these assets already that had basically deployed them, either when the cost of capital was cheap or following a speculative excess, you were able to capture all of that value. So there's no reason that venture investors, and Lux is certainly, I think, at the forefront of this, are thinking about the same thing, how do you take asset-heavy industries where all the groundwork has been laid, that there are these assets, there's embedded customer bases, and just service them?
And it's everything. This doesn't mean unsexy industries. It's cutting-edge software applied to old software, and it's things in space. So let's talk about those real quick as two bookends. 80% of financial transactions use Fortran or COBOL. You've got 60-year-old programming languages, and nearly half of the banking system is running on COBOL. 95% of the time that you swipe your ATM card, it is running on COBOL. So there's an opportunity for, yes, the maintenance of those systems, but also to maintain the infrastructure but change the underlying software. So that's one.
Take space. We, unfortunately, as you and I talk a lot about, are in a state of not just war, but many wars, lots of conflagrations. Those conflagrations actually have a visible signature. There are heat signatures when people are launching rockets. There are heat signatures when missiles are flying, whether it's in Israel-Gaza, or Russia-Ukraine. We have a triangulated satellite system in the US that helps us to identify at the split second when a missile is launched that heat signature so that that can be communicated down to a ground station and/or a counterattack so that counter artillery can be sent. Those systems need maintenance.
One of the great entrepreneurs that we had the opportunity to back was Tom Mueller. Tom was, as many people may know or not know, Elon Musk's right-hand guy at SpaceX for nearly 20 years. He is the main rocketeer responsible for the solid fuel rocket engines for some of the greatest breakthroughs that they had. After 20 years of working with Elon, which is an amazing feat in itself, he decided that he was going to start a business, and we funded him in impulse space, and he's focused on how do you service all of the assets that are in space today, how do you put thrusters, how do you have vessels to do maintenance so that you can delay the depreciation of these very expensive assets up in space.
So whether it's cutting-edge software applied to old software-driven infrastructure, or whether it is cutting-edge new technologies in space to maintain these very expensive long-life billion-dollar-plus satellite systems for triangulating to be able to do missile detection and defense, critically important things where maintenance is the underlying theme.

‍

Danny Crichton:
Now, one of the themes that have come out in a lot of these high-capital industries is sort of this democratization. So we look at space, it used to be very expensive to put stuff into orbit. That's why we cared so much about these satellites. We look at far distant probes, like Cassini, which lasted for decades far past their useful lives. Well, we invested in that because it takes so long to get that out there, it was very expensive to get it up, and then we democratize it so that the cost goes down, and so this idea of like, "Well, we should just replace it." There'll be new technology. Starlink don't have six satellites like the old Iridium network but now have tens of thousands of satellites, you could get them all into orbit. Or even look in the consumer markets like a Shein or a Temu, where fast fashion, I just buy a new pair of clothes every week because they break all the time, but I can drop ship them from China, they come in right away. There's this mentality of just replace, replace, replace.
I mean, just beyond interest rates, why should people care about that longevity and maintenance of these capital assets, whether in the consumer markets, homes and consumer goods, kitchen appliances, et cetera. I'm thinking of like an Impulse Labs oven or something like that, or a stove top, as well in space and other large capital assets?

‍

Josh Wolfe:
Well, a lot of this, again, as you were just saying, is the availability of capital. When consumers are flush with it and they're able to spend, then the replacement cycle is very rapid, you know? People are changing their phones every six months or every year, they're updating things, the availability of cash.
When cost of capital gets tight, when credit gets tight, when the cost of credit gets high, then suddenly people are spending less. So the replacement cycles get longer just because people can't afford. It's not that the demand might not be there, but their ability to actually pay is not there. So you have elongated cycles of product development, even in software, but certainly in hard asset goods. You might've been replacing something every four years or six years, you might've been getting a new car every three or four years, and now the life of those things is extending. That demands more maintenance of those systems. Automobiles especially, whether you're talking about old-school cars or you're talking about new cutting-edge cars, we have a company Kinetic that is focused on how do you service electric vehicles that need calibration and you're not going to your traditional Jiffy Lube to do that kind of stuff. So all of that just creates increasing demand for systems that are atrophying and entropy.

‍

Danny Crichton:
Well, and this is one of the themes we've talked about with Apex, which is general chaos is increasing across the world. Whether that's through war or whether that's through economic dislocations, there's just sort of this increased chaos which wears down on all these different systems. So the intensity of hurricanes, for instance, has increased dramatically. There was just a recent article in the last week talking about cat bonds or catastrophic bonds was the single most profitable trade last year at hedge funds in 2023.

‍

Josh Wolfe:
One specific hedge fund, by the way, and there's an amazing article, I think the Wall Street Journal just published this one, about Fermat Capital, which is up in Westport, Connecticut. It's something that I followed because it must've been 15, almost 20 years ago, Michael Lewis wrote a piece called In Nature's Casino, and I was obsessed with this because the idea of making probabilistic bets on insurance or reinsurance on catastrophe bonds is uncorrelated to the markets. Doesn't matter what's happening in the markets. It's a different kind of market that they were creating.
But what's really interesting is these low-probability, high-magnitude events. Venture capital, low-probability, high-magnitude events, where effectively, if you were to be very pejorative about our asset class, we are buying lottery tickets on founders who are inventing the future. The inverse is insurance. You are basically getting paid to help people pay for low-probability, highly-negative events, the catastrophic thing. And there were always these really interesting oddsmakers. There was a guy whose name escapes me at the moment but was a close colleague of both Charlie Munger and Warren Buffett who would ensure these things that had no priors, no precedence before, so Mike Tyson's life in a fight, some famous, I think it was Barbara Walters, newscaster's legs, you know, some supermodel, like these things that were very valuable-

‍

Danny Crichton:
The hand model for the iPhone was like 10 million insured.

‍

Josh Wolfe:
Was that actually true?

‍

Danny Crichton:
Yes, yes.

‍

Josh Wolfe:
So those things are very interesting low-probability, high-magnitude events, what people [inaudible 00:11:05] what are called black swans, positive or negative, depending on the convexity of how it's going to play out. That to me is fascinating, and these cat bonds, to your point about the chaos of the world, is also a really interesting asset class, and I'm sure that we'll find other things that are low probabilities that have no priors that people might create markets in. But the broader point about what we've called the entropic apex before, this goes back to the poet Yeats and The Second Coming, "Things fall apart, the center cannot hold, mere anarchy is loosed upon the world." We're seeing that, whether it's in geopolitics, whether it's in much of the democratization of the technologies that allow for a bit more chaos, the more expression we have. We generally think that's a good thing, but it creates a lot of divisiveness and a lot of chaos.
And so again, all of these systems, the more chaos, the more entropy, just think about it in very simple terms. I've got three kids. If they don't maintain their rooms, it turns to total entropy. And then what happens? Dad or mom is coming in and saying, "Clean your room, get the stuff off the floor, and clean up the bathroom," or whatever. It requires energy into a system to reduce the entropy and the disorder. Same thing in relationships. You and I are friends, we spend a lot of time, we have wonderful thought-provoking lunches and coffees, but it requires energy to maintain a relationship. Same thing in a marriage, in a friendship, in a partnership, in a business, in a startup, in a democracy. These things require energy to maintain the stability of these systems. That doesn't mean keep the status quo. You want growth, but you also need maintenance, the two pieces of that CapEx equation.

‍

Danny Crichton:
Well, this brings us to, I mean you brought up Yeats, but it also goes back to almost the Greeks. So there's a concept of a Ship of Theseus, which is sort of a thought experiment. I don't believe it's a real ship, but this concept of extensive periods of time, a ship that's being maintained, you replace wood, you replace the doors, the panels, the hull. Over time, the ship is no longer the original ship, nothing from the original ship.

‍

Josh Wolfe:
Plank by plank, yes.

‍

Danny Crichton:
Plank by plank. You, I think, referenced in the letter the idea of Salisbury Cathedral, this idea of keepers of the fabric, which was one that I had not heard of before but also was very interesting. What was that about?

‍

Josh Wolfe:
Well, it literally is this bloodline or this long line that goes back where they have one of the oldest clocks in the world and it still runs, and it's an example of individuals that are passing on, in this case, you could argue it's both technological maintenance and cultural maintenance, of these maintainers. And maintainers, you find not only in physical infrastructure like Salisbury Cathedral, but you see it in the maintainers in open-source software systems.
The idea that people are, whether out of their own incentive, out of a sense of moral duty, out of a value-laden motivation, they are leaders that could be cultural leaders who are preserving heritage through oral, or culinary, or costume traditions. It could be board members at startups who are taking everything that they know from the past 10 or 20 years of successes and failures, and leadership, and capital allocation, but it could be senior cabinet leaders advising our counseling prime ministers or presidents, it could be high-tech product managers who are navigating supply chain components. All of these people possess this sort of tacit knowledge from experience of how do you maintain systems. And if you just have people that are totally revolutionaries, either because they understand the system and want to upthrow it, or they are totally naive about how the system works, there are so many systems that we depend on today where the people that maintain those systems are so underpaid and undervalued.
We lived in a building in Tribeca. There was a guy that had been there longer than any tenant. He was the super, the superintendent. Lived in the building. The institutional knowledge that he had about this building, where every water pipe was and all the things through the seasons of freezing, and thawing, and water main breaks, and why this thing didn't work, and the bank in the basement, it's just like that institutional knowledge, and when he dies, for all the people that live in that building, they've lost their maintainer. And you have that same sort of thing. There was an old sci-fi book that I'm sure you've read, A Canticle for Leibowitz, where the world basically goes through a nuclear explosion kind of thing and there's this one person who's basically keeping the knowledge of humanity.
So I just think that we undervalue massively the people who are maintaining these systems, and this gets into the political and the social values. Bill Conway, founder of Carlyle, who put us in business, has spent a lot of philanthropic dollars on something that a lot of people have neglected, and that's the best use of philanthropic dollars, right? Not the thing that everybody's funding because it's good social signaling. He looked and identified a problem, and his particular problem was, "My God, there's a shortage of nurses." There's so few nurses in the country, and so he's been funding just thousands of nurse schools to increase the labor force of nurses, and he said, "You know what? If people are against immigration," because there are so many immigrants that become nurses, "don't get sick."

‍

Danny Crichton:
Right.

‍

Josh Wolfe:
And it's such a good point because those are the people who maintain the healthcare system and help you maintain your health.

‍

Danny Crichton:
Right. This dovetails with Matthew Crawford, who's a political theorist, but one of his book... I mean, Shop Class as Soulcraft is his most famous. The second one he wrote, The World Beyond Your Head, really gets at this notion of not just the importance of maintaining systems, but the sort of gratification that comes from protecting a system, realizing that you're part of this project that will outlast your own life. So he really emphasizes organs, the actual instrument of organs, and these people who come to, say, Yale or different locations where these organs need repair every 100, 150 years. They open them up. In some cases, these organs have been around hundreds of years, and so you can actually see the handiwork of all the people who came before you, the quality they did, things that they chose in terms of glues or in terms of ways of repairing this instrument that only one other person in the next a hundred years will ever know the quality that you did.
And so this idea of not just maintenance for demonstration or for signaling, as you're sort of getting at with Bill Conway, but this idea that you know yourself, that you've done something satisfactory, that you've done great work, and you don't need to put on Twitter, there's no way to actually demonstrate it. Although I will say, amazingly, with a lot of this maintenance, if you go to YouTube now or some specific channels on X, you do see more and more folks showing like, "Look. Look how amazing these pipes are. We've organized it in such a specific way. Clearly this is a plumber who really loves their job. They've done an absolute extraordinary job compared to the junk that you normally see in a lot of these constructions." So some demonstration here is actually interesting as well.

‍

Josh Wolfe:
Totally. And if you looked at my YouTube history, one of the most frequently-visited sites are these anonymous individuals. I don't know who they are, but something has broken in my house or some piece of technology is not working and I need to figure out a fix, or a kid's toy needs to be put together in a certain way and is missing a part, and what do you do? You go to the people that are basically like, "Here's how you fix it. Here's how you keep it. Here's how you maintain it." So we just vastly underestimate the maintainers, therefore they're important.
On the point about plumbers, there's an old corny dad joke which was about a plumber who comes, and there's this banging pipe, the thing's not working or whatever, and the plumber comes, and he goes in, and he bangs on the pipe, and he fixes it, and he sends a bill. He says, "It's a thousand bucks." And the guy says, "A thousand bucks? All you did was bang on the pipe." He's like, "No, no, no, the work was $5. Knowing where to bang on the pipe was 995."

‍

Danny Crichton:
Right. Well, I think that's always one of the challenges with tacit knowledge, right? It's very hard to demonstrate this sort of knowledge. You look at what universities teach, you look at computer science, you learn a set of theory, but when it actually comes to software engineering, understanding how to architect, what are the trade-offs, that professionalization that comes in from 20, 30 years of experience, that is why you're hiring a specific person.
It applies just as much to professions as it does to everyone in the trades, and obviously, over the last 30, 40 years, the US has really guided folks towards professions, towards college, away from the trades, but we're actually seeing that now we're looking at EVs, I mean, just across everything in deep tech, it's mostly trades. Anduril is a trade-driven enterprise. You're building actual machines, drones that are going into battle. That is not being programmed by a couple of engineers, like there are process engineers building the factories that actually get that stuff out there. You look at EVs, you look at Impulse Labs, you're building a stove, and the key to actually making that work is manufacturing engineers who actually know how to connect, automate, and scale that up so that you'd both get the right price that works for consumers but it's also reliable that it doesn't need repair constantly.

‍

Josh Wolfe:
There's a few aspects there on the Anduril and many of the defense companies we've backed, Saildrone and in others. The basic idea is you are maintaining the value system, you are maintaining the arsenal of American democracy, right? The firepower to thwart, prevent, and deter bad actors that are growing in the world.
And then the underlying technologies themselves require maintenance of epic proportion. I mean, you're not talking about a single widget that's going out. You're having tens of thousands of parts and components, and the maintenance and manufacturer of all of those things is critically important. Then you have companies like Hadrian that are manufacturing things down on earth trying to maintain this base of manufacturing, looking at the 3,000 mom-and-pop shops that since the Cold War have basically been making a single widget and doing that effectively and having a near monopoly on that specific thing, and as those people age, the knowledge entropies and decays with them, and so Chris Power at Hadrian said, "Why don't we roll all these things up?" And then he had a better idea and said, "No, let's build a machine that builds the machines to be able to maintain this arsenal of democracy."

‍

Danny Crichton:
Well, and I think it doesn't apply just to defense, but when we also look at bio and health systems. I mean, you mentioned nurses. Healthcare perspective, but then there's also the cellular systems and maintaining the quality of our underlying biology, which I think we have a couple of companies that are focused on that as well.

‍

Josh Wolfe:
We do. But the entire reason that you eat and sleep is to maintain your body, of course, when you're a baby up until the time you're about 24, so it's a lot of growth, but then the rest of your life, with good luck, you get to live a long one, it's all about maintenance, it's about achieving homeostasis. You have a fever, you fight off an infection, it comes back down. Your sleep is about restoration and maintenance of your muscles, skeletal, and your nervous system, repairing and restoring brain function, cellular reparation. And so yes, people are constantly looking at the cellular level, at the systematic level how do you slow things from aging for longevity, how do you increase maintenance of systems, and in some cases, the entire field of cancer is that something has broken down in the maintenance, whether it's mitochondria gone wrong, genetic mutation from exposure to pathogen or virus, like how do you maintain the homeostasis of the system.

‍

Danny Crichton:
And when I think about this, I mean, we are seeing this use of maintenance across artificial intelligence, and rebuilding software systems across defense, and thinking about the kind of equipment we're fielding on the front lines, you see this in bio, but the theme of the last 10, 15 years in Silicon Valley was disruption.

‍

Josh Wolfe:
Yes.

‍

Danny Crichton:
I used to work at TechCrunch. Our conference is literally called Disrupt, break systems. Move fast and break things was the ethos that's going on. I mean, do you think we're at a disjuncture, a moment when we are turning from an older paradigm of disrupting old systems to a newer paradigm in which keeping those systems, making them better makes more sense?

‍

Josh Wolfe:
I believe increasingly yes, and if you look at the big wave where a ton of talent and a ton of capital is going and has been going for the past few years, in AI, I was just showing family members cognition and the automated software engineering that is going to displace. So many people have feared that technology would displace the blue collar worker. What we're seeing with AI is a displacement of the white collar worker, of the knowledge worker, of everything that Peter Drucker and every management consultant push people to do. You have spoken brilliantly about this before about the creative class and the people that are writing, and producing, and editing, and drawing, and painting, and now, all of a sudden, a prompt engineer and the average person can come in and generate work that approaches an asymptote of pretty damn good, if not significantly better, at a fraction of the cost and time. The value of the knowledge worker needs to use these tools to continuously stay abreast.
In showing this to my family, they were like, "Oh my god." Software engineers, everybody's been pushing, "You have to be a coder, you have to learn how to code." This is going to become less relevant. I don't want to say obsolete because there's always going to be creativity in code, as there is in many of these fields, but what's not going to be obsolete is the guy that's maintaining the elevator, the guy that is maintaining-

‍

Danny Crichton:
Yes, yes.

‍

Josh Wolfe:
I mean, they're going to become more valuable.

‍

Danny Crichton:
Right.

‍

Josh Wolfe:
And so I actually am really bullish on things that can help the maintainers in this current moment when you zoom out, rising capital costs, lots of assets that need maintenance, increasing technological progress targeting the white collar knowledge worker. If you look around at all the systems and suddenly the people that maintain them, from the lowest of low things that you can imagine, plumbing, and heating, and HVAC, and electricity in your basic home or office, and suddenly those things stopped working, you'd be like, "Oh my God." COVID would look like a joke. And so that to me is the epidemic that I think people are going to wake up to to say, "Oh my God, there's a tremendous economic opportunity," and I don't know, in two or three years, people will start calling it maintenance tech or come up with some cooler name and it'll be a bubble, but I think that now is the time to be investing in technologies and ventures that are addressing the maintenance of the trillions of dollars of assets that are there and need to be maintained.

‍