Why the Future of Venture Is Hard Tech, Drones & Physical AI | Ian Rountree

If you find yourself wanting to hang out with a founder too much, they might not be good to invest in. I used to harp on this a lot: the question in software was not, “Can you build it?” It was, “Will they come?” In deep tech, it was the inverse. If you have something magical like fusion energy, people will buy it. The question is, can you build it?

1. Why Invest in Venture Capital?

Hi, I’m José María Macedo, and I’m really excited to be hosting this episode of the Delphi podcast. We’ve been doing a series speaking to emerging managers, and today my guest is Ian Rountree, who’s the founder and managing partner of KTOS Ventures. They’ve just announced their fourth fund, ready to go. It’s an overnight success 10 years in the making.

Yeah, exactly.

You guys started back in 2016, right? I think your first fund was still a mix of everything, but you pretty quickly pivoted into deep tech, into the stuff that no one was looking at back then. We were in crypto back then, and we were definitely very familiar with the venture climate. Everyone was looking at software and consumer stuff, like a marketplace for dog walkers or whatever—all this kind of stuff. You went very orthogonal into frontier deep tech, and you have some bangers in your portfolio. You have Neros Technologies, which we definitely want to talk about, and a bunch of others.

I’m really excited to dig in and speak to you. By the way, I’m an investor in KTOS—a proud investor in KTOS. I’m really excited to dig in with you. The first question I actually want to ask—you’ll have fun with this.

Yeah.

The first question I’d love to ask is about something you’ve said. I know Roelof Botha recently, or semi-recently, referred to venture as “return-free risk,” and you’ve echoed this to some extent. You’ve said that most of the industry underperforms T-bills once you account for the missing hurdle rate. I’m curious because I know you put a hurdle rate in your fund, which you didn’t need to do, and you did it because I think you’re a stand-up guy, which is actually rare in this industry.

Why do venture at all? I think there’s been a lot of soul-searching about this over the last year. There are some people who are asking, “Why do venture at all?” Then the LLMs came about and returned more money to the LPs, showing that there are banger venture investments. Now people are saying, “Maybe you don’t need to do seed and pre-seed. You can just pile into the winners,” which I know is a bit of hindsight bias—deciding which ones are the winners after the fact. But what do you think? Why venture, and why seed and pre-seed investing at all in this climate?

I’ll give you my answer, and I think the first question, in your career as in life, is: know thyself. I’m drawn to venture for 2 reasons. First, I grew up an only child reading too much science fiction because there was nothing else to do and thinking about the future. Once I realized there was a job—investing—where you’re paid to think about the future, and I’m also a little competitive and like that there’s a scoreboard, I thought, “Oh my gosh, I need to be doing that.” Investing is basically answering who is best at predicting the future.

That’s fascinating. There’s a job where you can just sit there and think about what’s going to happen in the future, and then you go monetize that. I’m drawn to that obsession with the future, but directionality is also important. Hopefully, the future is better than the past. In a societal way, an anthropological way, technology is the measure of time. The reason that today looks different from 100 years ago is mostly that our technology has evolved.

Sometimes you get giant societal or political shifts, but those are often correlated to techno-economic progression. I want to be part of making the future better, and what better way is there than to back entrepreneurs who are doing that? That’s part of the reason I ended up getting into hard tech in the first place, but we can talk about that in a bit.

That’s a good answer for why you should do venture. What would be your answer for why an LP should allocate to venture? Why is it interesting right now? What Roelof is getting at in the idea that venture is a return-free-risk asset class is that the median fund underperforms public equities over long periods of time and, in certain vintages, underperforms the risk-free rate.

There’s a huge caveat there, which makes me think of that old expression: How do you know a VC is talking their book? Their mouth is moving. If you’re in the outlier funds, then you’re an extreme outlier. You’re not investing for the median. You’re trying to invest in a fund that’s at least top quartile, ideally top decile. The top 1%–2% of venture capital puts up mind-boggling numbers.

There’s more persistence in venture capital as an asset class than in other asset classes. If you look at the top 10 performing hedge funds in 1 year, there’s not much correlation to the top 10 the next year. If you really account for the Sharpe ratio, then the pod shops maybe belie that rule, but on average, there’s reversion to the mean. In venture, the top 10 firms in 1 year are highly correlated to the top 10 the next year.

What’s going on there? Venture is a really interesting asset class in that a firm like Roelof’s firm, Sequoia, has been extremely persistent. I suspect it’s largely to do with the fact that the asset class picks the investor rather than the other way around to a greater extent than in any other asset class. If you want to go invest in Micron, no one’s going to stop you. But if you find some amazing, pedigreed founder in a hot space right now and they have their pick of the litter in terms of venture capitalists, Sequoia is going to be on their shortlist, in addition to a handful of other firms.

There’s been some evolution in that over the years, but it’s highly persistent from 1 vintage to the next. Roelof is basically saying, “Don’t invest in other venture funds. Invest in Sequoia,” right?

2. The Edge of Emerging Managers

That’s a good segue into where you think edge comes from in venture, and maybe specifically for you, where your edge comes from. Sequoia is very compelling, so you have to make an argument for why people should allocate to you over Sequoia to some extent. We have the same thesis as you—we’re investing in a bunch of emerging managers, and we think the edge is there.

The big platforms have never been as compelling as they are today. They can lead multistage rounds, they have massive media teams, huge brand equity, and lobbying teams. It’s pretty daunting. I’m curious: What is the edge of an emerging manager, and what is the edge of KTOS specifically?

I differ from some of my peers, and even my partner Grant, a little in this way. I’m a very competitive person, but I’m lazy enough to want to be unfairly competitive. I want unfair advantages. Some of the persistence that happens in venture capital comes from structural advantages and positioning.

You’re right: The big platforms, with more capital than they’ve ever had, bigger teams, and more capabilities to bring to bear for the entrepreneur than they’ve ever had, are exceedingly powerful. If you’re investing in something that’s selling to the frontier labs, the hyperscalers, or the government, there are people at these big firms who are 1 or 2 calls away from the very top. That’s going to be extraordinarily compelling versus a seed fund.

What I’m trying to do with KTOS is keep us the most performant firm that stays small enough to be aligned with founders at that first stage. When we write a $1.5 million to $4 million check, everyone on the investing team can look a founder in the eye and say, “Your company is a meaningful percentage of my net worth at this point.” I’m not looking at you as some call option that only matters if I get to invest $50–$100 million-plus in a later round.

Entrepreneurs are smart enough to know what the incentives of the investor are. That’s one of the reasons we kept the fund small. We went from $50 million to just $70 million this time because we like where we’re at. We think we fly under the floor of economic rationality for the big funds.

Not that they won’t dip down, but when they do, it’s really easy to have a conversation with the founder about who’s more aligned with you at this stage. If you get all of their time, I’ll grant you that maybe they’re going to be a little more useful than we are, but what percentage of their time are you going to get versus ours? That’s what we’re trying to do at KTOS, in part as an acknowledgment of how powerful the big platforms are.

If you ask an entrepreneur today, especially a younger entrepreneur starting a company for the first time, they’re going to want Sequoia, Founders Fund, Thrive, Andreessen, and maybe they’ve got a favorite fifth firm.

That’s really tough to play in. I think if you’re a new entrant in Series A or B and you don’t have some really unique advantage, like you were the former management team from Palantir or SpaceX and you’re investing in a bunch of your old colleagues, then it’s going to be really hard to compete with the big boys.

3. From Software to Hard Tech

Yeah, agreed. One of the ways in which I think you’ve competed is by being early to a very specific thesis. I’d love to talk about your deep-tech thesis generally, what gave you the conviction to pursue it, and how important that was to you. When did you start investing in deep tech, and what gave you the conviction to do it? I’m definitely going to have some follow-ups.

Yes. It was more gradual than binary. Ex post, it’s easy to put labels on things, but getting back to my psychology, when I founded KTOS almost exactly 10 years ago, I had moved to San Francisco in 2012. I joined SoFi, the fintech startup, very early, and even back then, fintech was weird. I question my own memory because it’s so dissonant with the present. I almost think I’m misremembering, but fintech was a new word, and the idea that you could apply software to industries that weren’t software was novel back then.

Again, it sounds crazy to say today, but 10 years ago, that was the state of things. We went to one of the biggest finance conferences, but all the banks were there. This was the ABS conference in Las Vegas, immortalized in The Big Short, and we were the only startup there. Wild.

I thought, “I don’t know if it’s because I came from Florida and Tennessee and had insight into other industries, but let’s do the Silicon Valley thing and apply software initially, while bringing technology to the other 90% of the global economy.” I thought that was more important because it’s the food we eat, the buildings we live in, our financial infrastructure, our manufacturing capacity, healthcare, and all of that.

That seemed so obvious to me, and again, this was just software. I started angel investing, and Fund I was really focused on vertical software going into these other industries. That was weird again. There were maybe 10 or 12 of us investing outside of Silicon Valley.

As we did that and entered these industries that are largely more physical in nature, the entrepreneurs were saying things like, “My software is going to be a lot better if I have a sensor at the edge. If I have a sensor at the edge, it’s going to be better if I have a GPU at the edge as well. If we’re going to put a GPU at the edge with these sensors, maybe we hack a KUKA or Yaskawa robot arm and actually start moving and manipulating things at the edge.”

That was how it happened. It was more gradual. Things went—

Three-dimensional.

It wasn’t from fintech to fusion. It started getting, as you said earlier, orthogonal. Almost a third dimension got created, and over time these things got more and more vertical. One day, they were so physical in nature that people started calling it deep tech. It was more of a labeling shift than a decision for me.

Interesting. You were following that thread because you were seeing interesting things being built in your portfolio companies, and your interests were leading you down there? Or did you think it was a place where you could get an edge specifically? What was your—

Yeah, like—

You were seeing great founders move into it?

No. No one else was doing it. Some were, but it wasn’t competitive back then because it was—

So if you were—

If you were a big fund and had a big brand, you maybe wouldn’t have gone there. Would you say it was a question of, “I need to be”—like you were saying before, you want to be lazy and not compete too hard, so you found a sector that not many people were looking at that you thought was exciting?

Right, back then.

Back then. I’ve never worked at a big firm, so I don’t know the counterfactual well, but I suspect it would be some combination of the fact that it would be so anathema to the senior partners that I wouldn’t have been able to do it.

When you’re in a bigger organization, there are 2 jobs. One of them is investing, and the other one is playing politics in your own organization.

Yeah. At bigger firms, you’ll sometimes see someone who’s much better at the latter than the former, and they might get elevated over someone who’s the inverse, or vice versa.

I think I’m bad at that second thing. I probably would have had the conviction but been unable to politic and articulate it in a way to the senior partners that would have gotten me stopped. I just had to go do my own thing and started backing entrepreneurs initially in vertical software, then IoT and robotics at a light level.

At some point, the lid just came off and I thought, “This is ridiculous. Silicon Valley is living in flatland. Of course we need to invest in 3 dimensions and fully embrace this deep-tech thing.”

And how important do you think that is? You’re still doing deep tech, right? But right now, deep tech is arguably not that—

Yeah, yeah, okay. You’ll correct me.

The evolution is that we now don’t love being called deep-tech investors because, to us, it denotes a lot of science risk and puts technology first. I’ve learned this largely the hard way. I’ve come to question the myth of the breakthrough.

Academically, I was taught that there are sustaining innovations and disruptive innovations, and Silicon Valley is all about disruptive innovation. I don’t know that that’s true. Occasionally, you get something transformational, like the Transformer model, and GPT-3 comes out and the world changes. You’d be remiss if you weren’t thinking about that in some capacity ahead of time or fighting to get exposure early.

Most of the time, though, it’s just 1,000 little things compounding. You do one thing better, then the next thing better. Each individual thing doesn’t look very interesting, but in aggregate, you have this compounding effect that we look for now. We would frame it more as engineering and execution than technological risk.

To us, the canonical deep-tech investment is some PhD who’s done their thesis in an amazing field and is a hammer looking for nails. That’s not what a business is. We prefer hard tech to deep tech nowadays.

Okay, hard tech. Another of the funds we invested in, Fondo, which I think you’re also a fan of, has this framing of technical risk—science risk versus engineering risk. You want something where the science has been proven out and you’re underwriting engineering execution rather than new breakthroughs, which I think is an interesting framing.

I also like this framing of hard tech, where you can think of some of the risks you take in venture as market risk and technical risk. For most of the software stuff—not all of it, but most of it—you’re taking on a lot of market risk but not that much technical risk. The canonical examples might be social media or the dog-walking app. These things don’t have a lot of technical risk, but they have a lot of market risk. You have to get someone to use it, get the network effect, and so on.

For most of the hard-tech things that you do, it’s more weighted toward technical risk. If this thing exists, it’s probably going to be extremely useful, right? Fusion, for example—you’re not going to have a shortage of buyers. Is that still a good framing for what you do, or is it too simplistic? I’ve heard you talk a lot about market risk for hard tech, actually.

I think it’s too simplistic. I used to harp on this a lot: the question in software was not, “Can you build it?” It was, “Will they come?” In deep tech, it was the inverse. You have some magical thing like fusion energy; people will buy it. The question is, “Can you build it?”

You don’t want to take too many risks, but there’s a difference in risks in kind and in magnitude. If you have a little bit of both, that’s better than a ginormous amount of one or the other. It’s more the aggregate of the 2.

I like the framing.

Of course, the most important thing is execution. It’s not just whether you can build it from a technical perspective. Do you have the wherewithal to run a team, recruit amazing people, keep them when other people are trying to poach them and pay them more money, build a culture, make strategic decisions, raise capital, and bring it to market?

That is a much bigger focus today. I also don't know that, in the fullness of time—and time is, of course, an important factor too, but I'll come back to this—in the case of fusion energy, you have these eye-popping valuations for things that might take 20 years to come to fruition. Everyone's like, “Oh, it's still worth it because these are going to be 10 trillion-dollar companies because it's cheap, clean energy.” I'm like, “Okay, let me ask you a question: What happens if you have 2 fusion energy companies succeed? Do you understand that competitive arbitrage is a thing?”

Why are you both going to be worth 10 trillion? You're going to be in a price war, potentially.

Are you building and developing these plants yourself? Are you just licensing the technology? If you're just licensing it, then you're going to have a high-margin business. But—

Are you really building a 10 trillion-dollar opportunity?

I think people are a little flippant sometimes about the breakthrough because a lot of us are sci-fi nuts and we like to think about the future. But in a business context, I'm not totally convinced that the breakthrough actually monetizes to the degree that you can take that kind of duration and capital risk.

The other thing is, the timeline is sort of uncertain. You know, the joke about fusion energy is it's 5 years away and always will be, or 15 years away and always will be. I've been hearing about it forever, but as a venture capital investment, my mentor invested in a fusion energy company in 2002. It's still working on it and making progress, and it's really exciting, and it might come to fruition at some point. But what do you do with that venture fund? You dissolve that. Are the LPs holding the shares of this still-private company?

You have to know—you have to have some idea of how long it's going to take. If I see a team that's knocking down risks every single day and it's a less mind-stretching technology than fusion energy or quantum computing, but I can see them making progress and I can just sort of connect the dots and pull out my stopwatch that's clocked to 10-year intervals, then that's a little more compelling to me than, “We'll figure it out one day.”

4. Contrarian Investing, Valuations, and Venture Mistakes

I definitely want to spend some more time on team with you because I think it's a super interesting topic. But on the hard thing, even if it's as hard, it's still, I would say, definitely much more consensus now than it was when you started investing in it, and the valuations sort of reflect that. How important does that mean it is to move on to finding the next niche?

How important is being contrarian in venture? There's almost this tension between being contrarian and adverse selection. To some extent, to find a good deal, you need to see something that someone else didn't pick or didn't see, or you're picking from a pool that others are overlooking in some way. But that's also the recipe to get adverse selection, right? So you need to find this superset of deals that people aren't looking at but that are good, or do you just need to be in the good ones and pay up?

You're alluding to the Howard Marks matrix, and I don't think it applies to early-stage venture.

Interesting.

Right? A lot of people tend to use the right-wrong, consensus-non-consensus framework, which comes from Howard Marks of Oaktree Capital Management, a firm that trades largely in debt, where there's a lot more information. It's a lot more quantitative. It's a much more liquid market. There, you have to have some insight that everybody else is missing because you know that most other investors are looking at the same asset.

I think we overapply that to venture, which is not a liquid market. There's far from perfect information, and you can get yourself in a trap because if you're funding an unprofitable—oh, the other important distinction is that, for the vast majority of the companies issuing the debt that Oaktree is buying, they're solvent. They're profitable.

If you're investing in unprofitable businesses, then by definition, you need to be somewhat consensus by the time they run out of money so that someone else with a bigger checkbook backs them. I think it's generally useful because you have to be doing things that are different, but not every VC is looking at the same asset at the same time. That's a really important distinction.

If you can be in the right networks and you can move fast enough with conviction, then we often write the first check to a pre-seed round on a SAFE, which means we can shake hands, sign, and wire the money the same day if we meet them before 2 p.m. Pacific. If another investor meets them next week, sorry, we beat you.

Yeah, it's much more about speed and conviction than it is about the Howard Marks matrix.

That's really interesting because people have such different views on this, actually, even VCs. So the alpha for you is the access—basically having access and then being quick. To some extent, picking quickly and having access and the sourcing—

At inception, I think it's different, by the way.

At Series A and Series B, the company is a going concern. They have some marketing and social media presence, they have a website, and people have been talking about them. You have pressure building around the company, and when they fire the starting gun for their raise, everybody's competing.

Yeah. At inception, we might be talking to a founder whose company isn't incorporated yet.

Yeah. That's cool. How do you think about that inception valuation, actually, and the importance of it? We have this GP that has a framework where polarized valuations are interesting. You want the seed-to-pre-seed deals that are very expensive because they're great deals and everyone wants to get into them, so they price up. You should do those because they're the top 0.1% of founders or whatever, and they're priced in.

Then you have these deals that are extremely cheap, where it's a founder that you've done the work to source. They're in Africa or India or somewhere, and you've managed to source them. Maybe you can do it at a $4 million to $10 million cap or something. The middle ground is the danger zone.

Again, you're going to tell me it's overly simplistic, which I agree it is, but do you have a—

I find it interesting. Going back to the consensus-non-consensus framework, in liquid asset classes, you'll typically see something like the consensus investments being less risky but offering less return, and the inverse being true for the non-consensus investments.

In venture, we have a good friend who runs data science at AngelList and puts out these really interesting studies. His name is Abe Othman. If you look up Abe Othman and AngelList, you'll see some of these. He asked this question, looking at AngelList data: Is there any correlation between valuation with respect to the stage and performance? He found there is actually a positive correlation between price and performance.

Which is a little counterintuitive.

Yeah, it's a little dangerous to follow this in the limit, but for the most part, it seems VCs are actually kind of good at pricing things. If it's more expensive, it's probably going to do better. If the whole market starts acting with that information, obviously it can distort things, but I just think it's important to remember that from the top-down perspective.

My own experience is that a lot of my biggest misses were saying no when I liked everything except the price.

Same here.

Yeah. It's hard when you're having a conversation with one founder and maybe they will have listened to this podcast and they say, “Well, Ian, you said sometimes you pay up for founders. What about us?” And what you're saying is, “Yeah, but you're—I'm not willing to pay up for you.” Very awkward conversation. Founders listening to this should know that that is often part of it.

So for founders that you've passed on and told them the valuation is too high, that was cope. It's not actually the case.

Yes. Although, again, there are some that I have said no to because of valuation. I've regretted it. Whenever I say that, I tell a founder, “I genuinely hope that you prove you make me regret this.”

Same. Yeah. What are your biggest misses, out of curiosity? And maybe your biggest learnings from the misses? Is there a certain category? In poker, you might have leaks in your game. In your venture game, there might be a certain category of mistake that you've tended to make. What are those, out of curiosity?

Yeah. My biggest misses are kind of 2 categories, which are both under the same header. There's saying no when I liked everything except the price, and then there's passing when I saw a deck and thought the pitch seemed stupid, but the signals about the founders were such that I should have taken the meeting.

Mhm. I made a judgment: I didn’t think this technology was going to work, I didn’t like the market, or this business model didn’t make sense to me. I forgot that I was pretending I was underwriting a business that didn’t exist yet. I was looking at a 7- to 20-page pitch deck, pretending there was an existing business to underwrite, and then either the entrepreneurs were so compelling that they represented most of the information I wasn’t seeing because I said no to the deck and didn’t take the meeting, or they were smart enough to figure out that the thing wasn’t right, tack and turn, and land on something that works.

Yeah, that’s my biggest mistake. One of our—and again, this is coming from a pre-seed and seed investor perspective—we like to say, “If the company has a website, it’s too late for us.” We want to introduce the company to its bank and its legal counsel. That’s how early we like to invest, so take everything I’m saying in that context.

I’ll give you an example from our own portfolio. One of our largest investments is in a company called Castelion that is furthering deterrence through the mass production of hypersonic missiles. We think having more of these things deters conflict and is a net peace effect. My partner Grant, when he was still at In-Q-Tel, led that seed investment when he was there.

I had seen the deck for Castelion at pre-seed, and this was before defense tech had gotten cool. I thought, “I don’t know. You’re making missiles. I don’t know about this. You seem cool, but who’s going to do the next round?” We had another investment in hypersonics that was more focused on making engines—liquid rocket engines—which are a totally different mission profile. I thought, “They’re both hypersonics. Maybe they’ll be competitive. I shouldn’t talk to them.” That was a profound miss.

Those guys are so incredible that, had I spoken to them in the first 5 minutes, I probably would have been banging down their door trying to give them a SAFE. Grant met them back then and wanted to do it, but it took them until the seed round to convince the partnership. Thankfully, we were now investors, because when Grant joined Cantos, he said, “Look, I know it’s past our typical stage, but I think we should invest in the Series A.” We paid a heck of a lot more than we would have if we had done the pre-seed.

That’s definitely my biggest class of mistake: overthinking a business. I think particularly overthinking a business model or a sector in a case where the founder is clearly excellent.

Yeah, it’s tough because you have to look at the counterargument. You have a filter. You have to apply some kind of filter. You have limited time for how many companies you can see, right? You can’t take too many meetings, so you do have to filter out some decks. Those are the hard trade-offs of venture.

5. What Makes an Exceptional Founder?

Totally. You and Pierce see it, right? Sometimes you’ll send us something that we haven’t seen, and we get into this back-and-forth on WhatsApp or email debating the business. This is a little dangerous to tell you, but if you guys are ever like, “I know all that, but this team is truly exceptional,” then we’ll take the call despite our reservations. But there’s only a certain number of people that you can give that ace to.

So let’s get into that. Everyone knows that you should invest in the best founders. It’s a trope, right? But there are very few people who really do in-depth thinking around what that means and can describe it. I’m curious: What do you look for? What does it mean for someone to be an amazing founder? Do you have different archetypes that you look at for different types of businesses? Do you feel it in the first meeting? What are your signals?

Do you have a north star internally that you use? We’ve just started putting ours together and have learned a lot from the Hummingbird team on this. I think they’re the best I’ve seen at talking about it, but I’m curious how you think about it.

It is funny to say, “We’ve had this incredible insight after 10 years of investing in startups: invest in amazing people.”

Yeah, exactly. Why didn’t I think of that before?

Yeah. Part of it is that my partner Grant likes to say, “Your bar for excellence is the best you’ve ever seen.” Implicit in that is, “What’s the best you’ve ever seen?”

Yeah.

Not everyone has worked hand in hand with Elon, so everyone else’s bar is going to be a little bit lower. The calibration is extremely important in this game. When I was just some junior startup fintech guy starting to write angel checks into startups, and then when I had my first $4 million demo fund dabbling in deep tech, I didn’t really know what good looked like. Everybody seemed smart to me. Over time, we’ve gotten more vocabulary around it. Our bar has gone up.

What is that vocabulary?

We score founders along 6 attributes. These are always evolving, and how you tease out these attributes is how you put more flesh on the bones. We’re constantly adding and tweaking questions that we ask in diligence.

The 6 attributes we look for are velocity or drive: what’s their impulse to move forward and make progress? What is their talent gravity? A big portion of getting things done is recruiting the people who get those things done. Are they a narrative builder and storyteller? Do they have a growth mindset such that they’re improving upon their baseline?

What is their financial mindedness? Do they understand that the point of a business, at the end of the day, is to produce profit margins above and beyond its cost of capital and build some immunity through competitive arbitrage such that those margins sustain over time? Do they know that’s their job, or do they just think having a startup is cool and that they should raise at the highest valuation possible because then their friends give them more pats on the back?

The 6th is whether they have an encyclopedic knowledge of their industry. Those are the 6. We’re not quantitative, but it at least gives us a framework to discuss things. We can say, “Oh, hey, we forgot. Let’s get more data on this point.”

You have this framework, which sounds very rigorous, but you told me that with the Castelion founders, you would have known in the first 5 minutes. That’s certainly not enough time to go through these 6 attributes. What would it be about meeting them in the first 5 minutes? Is it just a pattern-matching thing, or was it something specific, like they had one of those qualities to such an extent that you’d say, “Yes, I have to do this”?

More generally, how quickly do you know? What’s the latest you’ve ever known? Has it been on the 3rd meeting when you realized, “These guys are actually great. I was wrong”? I’m curious about this.

I’m querying the database for that.

Yeah.

No, I think you know really early. Anytime we’ve second-guessed our first negative impression, I don’t think we’ve ever come back to someone and said, “You know what? They are great. We were wrong.” The one exception would be—and this is why growth mindset is one of the 6—you meet someone who has a very low baseline, but their slope is incredibly high.

They don’t seem great, and then you realize they’ve grown so much in 6 to 9 months that you think, “Oh man, we missed this. We should have plotted a couple of dots so we got some indication of slope.” If you can, from an evolutionary perspective, invest in a primate that isn’t evolving at all, and you’ve got a bacterium that’s evolving every few seconds, you should probably bet on the latter, because over the long arc of time, they’re going to catch up.

We see that sometimes in entrepreneurs. Maybe they’re a Thiel Fellow who didn’t even finish high school, and they seem really green, but they’re learning so fast. Then you’ve got someone who’s experienced but very set in their ways. We’ll bet on the first all day, every day.

Now, a team as veteran as the Castelion guys—all 3 of whom, Bryan, Andrew, and Sean, were at SpaceX in senior roles for a while—does make me question how important growth-mindset weighting is. If your y-intercept is already exceedingly high, do I really care whether Bryan Hargus is growing or not at this stage? No, I think he is. Bryan’s always learning, very curious, and looking to improve.

But does it matter as much for a team like that? I don't think so. That's the one of the six that I always bring up: do we need this one?

Yeah. Interesting.

Yeah, we have a similar one.

To the first-time founders. There are a few things I'm curious about. How important is it to you that they have shown courage and a sort of first-principles thinking—independent thinking? I think courage is the ability to burn all your bridges and do something completely nonlinear, where you're doing something completely revolutionary that requires a lot of pain to execute.

Every company that has become great—or most of them, when you study their history—has these moments where the founder had to do something completely nonlinear. It wasn't just an adjacent pivot; it required immense courage. A lot of founders can execute on a clear mission, but do they have the courage to make those tough decisions? How important is that to you?

Yeah, we've talked about this. We feel it's mostly embedded. The attribute I call financial-mindedness is—

Also kind of like strategic thinking.

Okay. We'd probably say that this is reflected in some combination of that attribute, contextualizing the decision in an encyclopedic knowledge of the industry, and how good you are at story-crafting to bring others along.

Mm-hm. But if we were going to add another, it would probably be to call out courage, as you put it.

Exactly.

When were you right when everybody thought you were wrong? Or, conversely, what's your greatest mistake? What did you learn from that? Those are twins of the same thing. You have to be able to rest in your own convictions and learn from them, rather than being blindly egotistical or zealous about something. Or is it too hard to describe?

You get glimpses of everything. Our framework gives us a way to—

Score, compare, and discuss these attributes. But this all comes together in the way we plot these six attributes: on a spider chart. I like to think it implies a graph like you'd see if you were looking at the bending of space-time. Are these entrepreneurs creating such intense energy density that they bend the space-time continuum, like gravity does?

If we don't feel ourselves pulled toward them in the meeting—as much as we love our jobs at Cantos—if there's not some small part of us that's thinking, “Oh my gosh, I kind of want to quit and leave this company,” then—

You know, do they have this reality-distortion field such that they're inspiring, borderline altering your reality? If not, then they don't have it. That's talent, gravity, and narrative-building, but also their velocity and drive.

In terms of relativistic speeds, there start to be contortions to space-time as well. They can't just have that; they need to really understand their industry and have a strategic mindset. But sometimes, because of their background or who introduced you, you're giving them a checkbox on that.

Bryan Hargus led the Starshield initiative, the government version of Starlink within SpaceX. Can you lead a team, make business decisions, and sell customers? Going into that meeting, it was a checkbox.

How much time do you spend on their business versus their background and their lives, digging into that? What does that look like for you?

I think we get to their psychology probably a third of the time, and that informs most of the scoring.

What do you mean by a third of the time? A third of the time we spend with the founders in diligence—

—is on psychology. That's understanding them, and that's most of the score.

And then the other two-thirds is the industry and the business. Are you doing that because you want to know about the industry and their business, and you might disagree with them—“Actually, this isn't an interesting sector; it's not going to have a moat”—or is it more just a vehicle to understand the founder? Is it another vehicle to understand how the founder thinks?

Both. That conversation will inform really two and a half of the six attributes. If you're talking about the business, then you're assessing their knowledge of the industry and their financial and strategic mindset in that context. You get a little bit of the narrative-building: how do they articulate their understanding of the industry, the problems, and all that?

But we're not making the decision on those six attributes alone. There's the founder score, and then there's the trend score—the market score. We sometimes talk about this as rider and wave: if you have the best surfer in the world but there's no swell, you aren't going anywhere.

Yeah. And so, a conversation we've been having more recently is, “Okay, if you have the biggest swell in the world off Portugal, then obviously you need a capable enough rider to ride the wave, but do you need the world's best surfer?” Probably not.

Mm-hm. And so it's an awkward conversation to have. We've recently been asking, “Okay, how much should we lower our bar if you've got the biggest wave ever behind you?”

Yeah.

Yeah, we have the same discussion, because I think investing in crypto in particular teaches you how important the wave is. You could have backed pretty mediocre surfers and still done very well in crypto if you invested at the right time, as long as you managed to sell before the wipeout, which is another skill altogether.

I think the wave is really important. We have similar discussions, and that goes into venture right now. I feel like people get this right: they understand that if you're in these big waves, you don't need a top-decile founder to do okay. To some extent, people are tweeting, “Is this security going to sell at a higher price later on?” rather than asking whether the business is going to do well.

6. Bringing Technology Into the Physical Economy

You see it with Palmer Luckey companies, or whatever ex-founder companies, which raise at a very high valuation. Project Prometheus right now is a good example—the Bezos company where people think, “He's going to be able to raise at a higher valuation later. I won't lose money here.” I'm curious about your thoughts on this in general, because it's a different calculus. It's more of a left-curve calculus. I see Bezos; I see Palmer Luckey. I don't think about valuation, moat, or defensibility as much.

I mean, in recent history, it's been a rational strategy. I might roll my eyes at it, but I can't blame them. I used to have this view that venture is a long game. It takes 10 or 12 years to know, and you just have to stick with these companies. Sometimes they pull that off, but again, these are insolvent businesses. If they run out of money—

Yeah, you can't afford to be non-consensus for more than a few years if you're really pinching pennies. To have a shot at becoming valuable in the future, you have to make it through each of the gates.

It's path dependent—

Right? And so the pejorative of momentum investing in venture is, in some way, an acknowledgment that it's a multiround game. You have to make it through each of the nodes in the decision tree, or you're not going to be valuable at all.

I can't figure out if this is an aspect of the game that we're finally acknowledging rationally or if it's a mind virus. And I'm like—

Yeah, we'll find out in a few years.

Yeah. I find myself thinking, “Okay.”

Well, I have a bunch of questions, but I know so-and-so is really [in] a hot category, or so-and-so’s investing, and that changes it a little bit for me. I kind of hate it, but I also know that’s the game on the field. I’d say where we settle on this is: let’s be half stubborn and contrarian, and maybe half acknowledge that this is the game on the field.

Interesting. I like that. Yeah, I think liquidity—

Liquidity changes this a lot, right? Liquidity makes this look a lot more like crypto. The fact that you have good secondary liquidity for these venture names means that you can afford to be a bit more momentum-driven and think less about the endgame of whether this company will IPO or be acquired, which is how you had to treat venture before, right?

This is a 10-year hold. I need to underwrite this thing being acquired or going public, and if it doesn’t, I can’t do it. Whereas now, if you have these liquidity gates with secondary liquidity, I think it changes the rational strategy quite a bit. It starts to look more like crypto, which is actually what you’re seeing, right? There are these meme companies—

Which—yeah, which I think is quite interesting.

I often remind people that it’s tempting to think that we live in objective reality because that makes it more intelligible, but that’s not true. At least in finance, we know from behavioral economics that we live in subjective reality.

Yeah. And so, for something to be more valuable, this greater-fool theory—which is itself kind of a pejorative—is, in a way, an acknowledgment that an asset is worth whatever people will buy it for. And the more people—

Believe in the asset, the more it’s worth. And so you have to perpetuate and expand belief in something. If you’re contrarian forever and everybody in the world disagrees with you forever, then there are going to be no buyers for the asset.

Now, you can overdo this and just be a pure meme with no substance, and that’s not productive for the world. I would argue that the thing that keeps me centered is that I don’t want to exit a position in 10, 12, or 15 years, make a bunch of money, dump it on retail, and have the stock plummet. That would not make me feel good.

I’d feel bad sending my kid to school knowing that what paid for it was dumping a crappy asset on less-informed investors. I want to invest in something that just keeps going forever. I’m tempted, when we distribute shares and I take my carry in kind, to say, “I don’t want to sell most of this because I love the company so much.” That’s the endgame.

In the short term, we know you have to get the next round done to have a chance to pursue the endgame, but the endgame is something enduring that perpetuates and compounds forever. If you look at Kantos.vc, our website is “Visions of the Future,” where some of these companies are engraved in stone and metal in 1,000 years.

I want your website. It’s so dope. It’s like this future-archeology theme.

You know? Yeah, I really like it.

Okay. I wanted one last thing on process, but maybe we could just make it short. Then I want to move on to some of your companies and some of your theses, because you guys have really unique takes, which I always learn from.

You seem, from the outside, interacting with you guys, very structured, rigorous, and process-driven. You write great memos and produce great research—these long decks about where you see the future going. Is that how it is internally when you’re doing a deal? Are you producing a memo that’s very well-referenced and discussing it, or are you sometimes committing in the first 5 minutes for a Castelion-like deal?

I have a rule that we don’t commit the same day.

Okay.

We might say something to the effect of, “Look, we’re really interested. Let us sleep on this and get back to you tomorrow.” There are 3 investors, so at the very least, we need to go have a conversation with the other 2, although each of us has an ace we can use if the other 2 disagree. You just know that if you’re doing that, your neck’s on the line.

We usually end up making a decision within a week, sometimes 2. The process is such that all the information is already there. We’re having the conversation between the 3 of us. It might be calls and texts, Slack threads, and Granola notes. We dump everything we can into a Slack channel, and later we’ll have Claude turn that information into a memo. Very rarely do we have the memo actually written before we make the investment.

Okay, I like that.

I view our Slack channel, where we input all this information, as the time series. That is the raw data. Then we articulate the decision, typically when we announce it to our LPs and in the subsequent quarterly update.

Nice. That makes sense. I’d love to go into some of the companies and one of your theses that I think you told me about and that I really liked. Maybe you can recreate it here.

The S&P 500 is, like, 40% tech right now. I think when we spoke, maybe it was 30% or something like that. In the future, it’ll be 100%. There are all these boring industries, some of which you have investments in, like mining, lumber, and fisheries, that technology maybe hasn’t touched at all. There’s this really interesting opportunity in picking amazing founders to take technology into industries that are very much stuck in the past.

I’d love for you to riff on that thesis and maybe talk about some of the investments you’ve made there, because it’s one that I haven’t heard many people articulate in the way you have.

Again, this is an extension of the original inspiration of bringing technology into these big, important industries that matter for the world. Silicon Valley is like living in Flatland. Why are you focused on 2 dimensions? We live in 3—4 if you add time. We need to be investing in these other industries.

We don’t want technology to be this boogeyman that’s only for the elite and steals jobs. We want it to create value, for there to be technological deflation in necessary asset prices, for us to invest in abundance and resilience, and, from a geopolitical perspective, in advantage versus more nefarious jurisdictions. You can only do that if you break into the 3rd dimension.

We look for companies that aren’t just operating in these industries and bringing technology to them. Because these industries are often staid and set in their ways and haven’t innovated in a long time, selling technology to the incumbents is rarely enough to change the world. You have to compete with them.

This is intimidating because they have incredible advantages—switching costs, IP, relationships, and all of that. But the most important companies in the world will displace the incumbent. It will be Tesla to Ford and Toyota. It will be SpaceX to United Launch Alliance. It will be Anduril, Castelion, and Neros to Northrop, Lockheed, and Boeing.

It’s not, “Please, Mrs. Boeing executive, buy my actuator.” It’s vertically integrating to compete with them. In some cases, you do get something horizontal. I used actuators because they’re top of mind and so valuable, not just for aerospace and defense but also for robotics. Something like 50% to 70% of the BOM cost of a humanoid robot is motors and actuators.

Despite saying that, I think that’s maybe one exception. For the most part, you want to go compete with them. If we’re going into an industry like mining, I don’t want to sell AI-driven leads to a mining company. I want to buy a drill, like our portfolio company Earth AI, apply for the mineral rights myself, and find metal deposits that I own with my technology, rather than licensing it for some small sliver to someone else who’s going to find literal gold.

The same goes for our portfolio company Shinkei, which makes robots that process and slaughter fish. Soon, they’ll also do the breakdown and packaging as well. Rather than selling those robots to fish-processing companies, they’re vertically integrating where they’re selling fish. They have a brand, Shinkei—I’ve got my hat right here. If you go on that website, you won’t see anything about the robots, but they’re able to give you a better product because they’re using technology on the back end.

That’s much more our preference than trying to sell the robot into an old, maybe even family-owned business.

Yeah, it’s a great thesis. I love that thesis. If your technology is so good, why don’t you compete, right? To some extent, the SaaS—there was a time when we thought this productivity software was really important, or sales software, but it wasn’t. If it was so important, you could have used it to compete with the incumbents rather than selling to them.

I think we’re seeing this a lot with AI. These companies are saying, “I don’t want to go sell to a mining company. I can compete with them and reinvent the whole process from first principles.” I’d love for you to talk about the lumber company as part of this, because I remember you mentioning how there are these machines to cut trees. There are these really big trees that have the most wood, and they’re like, “Why don’t you cut those down?” The response is, “It doesn’t fit in our machine.” They’re mind-blown by this inefficiency.

But yeah, it was something like this. Maybe you can tell the story.

Yeah, no, you’re remembering right. There have now been iterations on the machine, but it’s an illustrative process. We invested in a company called Lumber Manufacturing, and its operators and engineers are out of Blue Origin, Planet Labs, SpaceX, Anduril, and, importantly, the lumber industry.

The founder got really obsessed with the lumber industry. He was opening a facility for a prior defense-tech startup in Mississippi because of local subsidies there, and he started meeting people who were forest owners. This was after the pandemic, when lumber prices went through the roof because everybody was at home working on their renovations. He was like, “Oh, you guys must have made a killing, right?”

They were like, “No, actually, the lumber mills made a killing. But those of us who just grow trees and sell to them, we didn’t get that passed on because lumber mills have so much power in their local geography that they’re monopsonies. You’re the only buyer. What are you going to do? Ship the logs another 300 miles to another lumber mill? It eats all of your profit.”

They started studying the industry, asking why there was such a concentration in the mills and whether you could use technology to break that up and bring power back to the landowners. As an example of sustaining innovation—it’s 100 little things, 1,000 little things, not disruptive innovation—they were looking at some of the trends and noticed that the price of a tree goes up to a point and then comes down. You’re like, “Well, hold on. There’s more wood in that tree. Why is it cheaper?”

They were like, “It doesn’t fit in the saw.” And he was like, “Well, can’t you just build a bigger saw? Or can you buy a bigger saw?” They were like, “They don’t make them bigger. There are only 2 companies that make these saws, and it’s the biggest ones they’ve got.” And they’re like, “Well, how hard is it to make a saw? Can you just design one that’s a little bigger and kind of arbitrage the fact that there are bigger trees for less?”

They were like, “Huh. Never thought of that.” They had been doing it for decades; they just never questioned it.

And if you can’t—not everybody can design it. It turns out it’s really hard, even for SpaceX engineers, to have blades under that much tension. Not everybody can do it.

But if you have a capable enough team, then you can start to make observations, question things, innovate, and compound this again 1,000 little ways through the industry. You can only do that if you’re vertically integrated. If your business is doing everything, you have that in-house. You have the team, the culture, and the ability to make strategic decisions that are counterintuitive to the current industry logic.

We have a fully operational lumber mill in Mississippi, and soon they’ll break ground on numbers 2 and 3. That’s the idea: vertically integrate and sell lumber. Our end product is a literal commodity. We’re not selling technology; we’re using our technology to have a preferential cost structure for making a commodity.

And how do you think about this? With hard tech and vertical integration, the question of China always comes up, right? In theory, it’s much easier to vertically integrate in China. They have these Shenzhen supply chains that really bring down your innovation cycles, and there are certain things that America just doesn’t produce enough of to vertically integrate, like these rare metals and magnets and stuff like this.

Western world.

Yeah. So how do you think about this? Can American companies actually compete with China in these hard-tech sectors when it comes to scaling? The pattern is that America innovates, right, and then—and maybe this is a little bit unfair—the trope is that America innovates and then China goes and out-executes. Is that true anymore?

Yeah. Look, I’m a patriot. For those in audio, I’m wearing my USA jersey.

I don’t think it’s true either. That’s why I corrected it mid-sentence to the trope. Yeah.

I think it was true for a while, but we don’t have a monopoly on intelligence. If you outsource things that are thought of as lesser industries, you can get surprised when those pieces of the stack that underpin more important parts of the stack are no longer controlled.

I’m very concerned going into a world where we might get physical AGI. LLMs are cool and all, but when AI breaks into the 3rd dimension, that’s a lot more interesting to me. It’s very exciting. But as I said earlier, if 50% to 70% of the BOM cost of a humanoid robot is motors, then we’re potentially in trouble, because we don’t really make cheap motors in the Western world.

If you look at drones, sometimes you’ll meet a defense-tech startup that’s got a little drone. Go look at the motor. I guarantee you it will say “Made in China.” This is one of the things our portfolio company Neros is bringing in-house, and they’re onshoring—or at least friend-shoring—all parts of their military drones.

But DJI drones, you know, those are made in China. That’s fine. It’s not just the motors; it’s the whole stack, as you said. You have the entire industry and city of Shenzhen that, if you look at it from above, might look like a motherboard: everything is right next to one another where it needs to be and optimized.

You’ve got the rare earths mined in China and processed in China, and the chemicals to process the rare-earth minerals are in China. Then those motors—you know, the copper is also produced in China, and the copper and rare earths get put together in motors. Those motors go into robots, electric vehicles, and everything else.

There’s a bunch of that stack that we don’t even control in the Western world, and that’s going to underpin robotics as they go into our homes and industries and potentially onto the battlefield.

Sorry, do you want to finish up, actually?

I hope that we all can get along geopolitically, because that’s a hell of a disadvantage we’ve got.

Well, follow-up question: Why aren’t you investing in China if that’s the case? Is it just patriotism, bandwidth, or something else? It sounds like you’re making a case for investing in Chinese hardware, or Chinese hard tech, to some extent.

We’re an American venture capital firm. We found that our best returns are in our backyard. Again, I am a patriot. I think we have to build these capabilities in the U.S. and with our allies in the Western world. I really hope we get along, but you have to be prepared for if we don’t.

There’s going to be incredible asymmetry. We gave up on our sanctions on China because they basically flexed their control of rare earths and Trump had to back down. If you think that was bad for where rare earths go today—into electric vehicles, smaller devices, smartphones, and some military technologies—if we get humanoid robots everywhere and China can flip off a switch, then they’ve got all the power.

I don’t think that’s healthy for the world. So we’re investing in the States. I don’t have any exposure to China in my personal portfolio. Again, not that that’s sort of a protest against the current regime in China.

I lived there one summer in college back in 2008, and it was a very different place. I haven’t been back since, and I don’t know if I should go back given some of our investments, but I had an amazing time there. It was really opening up. Xi Jinping’s running a very different CCP, and I find that concerning.

So I’m not investing in China in protest. If I didn’t have that principle and I wanted an investment in robotics, I would find a way to invest in Unitree.

7. Defense Tech, Drones, and the Future of Warfare

Yeah, I agree with that. I want to move on to defense, because you guys have this investment in Neros, which has been, I think, a defining investment for you. I’d love to hear about it. This is one of the areas we’re most interested in.

I think physical AGI—people talk a lot about robotics. I feel like no one talks about drones, even though it’s the thing that has product-market fit right now. Drones have had their ChatGPT moment, right? In some sense, it was the Ukraine war, you could say, or maybe even earlier.

How did you start investing in defense? What’s your thesis? Maybe tell us a bit about Neros—wherever you want to take it. I’d be very curious.

I first invested in Neros when it was just Soren. We were their first investor and led their pre-seed round in August 2023. His co-founder, Olaf, who’s also incredible, had not officially joined full-time yet. So this was very, very early.

We met Soren, and he was building drones while still living with his parents. Soren had dropped out of high school to pursue a career in professional drone racing. He won the world championship, and at that level, you’re not just racing DJI drones; you’re tweaking and building your own.

So he really was an engineer as well as a pilot. He'd become very concerned because the war in Ukraine had broken out, and it was being fought with consumer-style quadcopters. He was the world champion drone racer and was building drones, and he thought he could bring something to bear for this conflict to help Ukraine defend its sovereign territory from an authoritarian regime.

We backed him, and he started building drones. 2 weeks later, he sent us a video of the first prototype doing 130 miles an hour and pulling 5–6 Gs in his parents' backyard. Then they were like, “We booked flights to Warsaw, and we're going to Kyiv to see if the Ukrainians need these.” We wrote an investment into the company—the first money in—and 2 months later, they were in Ukraine meeting with the Ministry of Defense there.

They talked about speed, and that got the attention of some other investors. We introduced him to our friend Shaun Maguire at Sequoia Capital, who gave him a term sheet and led the seed round. We invested alongside him again there. Now, this is not even 3 years later, and they are the highest-volume American drone manufacturer. They just crossed 1,000 drones a week. Their goal is 1 million drones a year.

You need something like that if you're going to deter conflict and defend territory. Yeah, it's been incredible to watch, because again, it's been less than 3 years since you wrote that first check. To see them at this scale and speeding up is just mind-boggling.

And how do you win? What do you think matters in terms of building a huge drone company? One founder we spoke to framed it as, “There are these 3 problems”—and I know you're going to say there are 1,000 different things—but the 3 problems were autonomy, fleet software orchestration, and mass manufacturing. You have to solve all 3: the future is these autonomous flocks of drones that can use advanced AI software to move, can't be jammed, and can be manufactured at scale.

Is that how you think about it? I'm really curious how you see it. What kind of investments would you be making right now? Do you think it's still going to be mostly FPVs that get built out? How do you see it?

This gets back to the earlier point: it's not necessarily an innovation; it's execution. I was speaking to someone in the Pentagon recently who is an assistant undersecretary in the Department of Defense. They're really funny, and they had talked to another friend of mine. When he met this friend, he said, “Hey, Mr. VC, let's play a game. You tell me what you think I want, and I'll tell you how wrong you are.”

My friend was like, “Oh, okay. Well, you mean you really need attritable mass, affordable mass, drones, and hypersonics?” He listed off some of the things that the Pentagon's interested in. The person said, “Yeah, yeah, we need all that. I'll tell you what: we would give $1 billion to someone who could make 1 million of something.”

My friend's ears perked up. He was like, “Oh, well, 1 million of what? I'd love to go invest in this.” The person said, “1 million of anything.” He's being a little flippant and funny, of course, but the point is that we need scale. We don't make anything at scale in terms of military technologies, and that is a huge problem.

Sometimes Neros will post something, and you'll have naysayer drone experts saying, “The blade could be better,” or whatever. First of all, they're probably wrong. But second, if you have a drone that's second-best but you can make 10,000 more of them than a drone that's a little better, I know which one I'm picking.

Neros is absolutely pushing the envelope, and their drones will be best in class. You've seen the Drone Dominance program, where the Pentagon is running a literal flyoff competition between drones. Neros is number 2. They're the number 1 American supplier. There's a sort of joint Ukrainian-British company that's just edged them out in the early competition, but notably, on that entire list, Neros is the one who's shipped far and away more drones.

The point is that you need to manufacture, and you need to hit cost. You need to do this at scale. You need to make things cheap. Those 2 things go together. Having one of the best missiles in the world isn't going to do anything for you.

Okay. What do you think the future of—how important are drones in Ukraine? They've been incredibly important, right? It's sort of like there aren't that many humans on the front lines anymore. Do you think the future of conflict is how many drones each country is going to have? Are we talking millions or billions of drones? What percentage of military spending? What's a good way to think about this market?

Is it only nation-states, or are tanker companies and oil infrastructure also going to have their own drone fleets or anti-drone warfare? I'm curious—how big is this market?

When you say drones, you're referring to Class 1 drones, which are quadcopters—DJI-style, almost consumer-looking, first-person-view, or FPV, drones. That's what Neros makes, and they're focused on that. They want to be the best in the world at it and ramp manufacturing there. Those have exploded onto the scene because of the land war in—

Literally.

Which is literally a battlefront where, over the centuries, there have been a lot of wars fought. There are open plains. Class 1 drones are perfectly built to deadly effect for land war. You need them to deter conflict, to deter an invasion, particularly in Europe.

I think it's a little different from the U.S. perspective, where we are isolated and very friendly with our neighbors to the north and south because there are only 2 of them. For, say, Japan or Taiwan, as islands, you're dealing with a different theater with different ranges. You probably will have containers of drones if you have ships coming in close, or to take out other drones that are coming in, but those are probably going to be longer-range, Class 2 or 3 drones.

Then you need to pack Japan, Korea, and Taiwan full of these things to deter some type of—God forbid—land invasion. I don't think it'll get there. I pray that it doesn't, but again, deterrence goes a long way. You need to couple those with other technologies, of course.

Class 1 drones are here to stay. We're going to make millions of them. We have to, to lend to an ally like Ukraine or to deter conflict ourselves. I think there probably will be commercial equivalents for private security. As an investor, that's a smaller market today, so I'm less interested in it. But in the fullness of time, I think we will see commercial vessels have a hidden container on board that deploys, at the very least, counter-UAS systems.

When you're investing in defense, do you want single-product companies like Neros, or are you looking for new primes that are building a portfolio of interesting products? What do you think the most interesting types of investments are in defense?

Our leading positions are Castelion, which makes hypersonic missiles, and Neros, which makes FPV drones. We think there's an advantage in picking 1 thing that's very important and doing it better than anybody else, versus a portfolio approach. But there are only so many of those categories.

Our third investment in defense tech is a company that came out of stealth earlier this year as they closed their Series A, called Heaviside. Heaviside is taking more of a portfolio approach. It's a lot harder. You have to have a team that's much more flexible, and you have to have leadership that's extremely savvy in defense procurement and decision-making, because if you go down a path that isn't working, it's very, very risky.

We think that Heaviside has that team, and so they are working on a portfolio that includes 3 products initially. We'll expand from there. They're doing everything from handheld RF sensors, so that every warfighter can see if a drone's coming in or if there's an enemy troop nearby on their radios—a necessary technology in a post-drone battlefront.

They're also making loitering munitions, so winged drones with farther range, and they are developing submersibles. You'll probably see them do other things beyond that. But that's a much harder approach, and in many ways we made an exception to our own rule because we found a team in Heaviside that's an exception. I would not necessarily recommend that to defense tech upstarts.

Makes sense. Yeah, that was the one you were really excited about that hadn't come out of the other—

The other profile? Yeah, they were so secretive for a long time that we redacted the name of the company in our own LP reports, and even the financials. The other trend that I think is interesting, where we haven't made an investment, is sovereign new primes.

So, you have some in Europe that have raised a lot of money now. You have—I saw one recently that's a New Zealand-based defense tech startup. I think that's interesting because you probably will have a sovereign Anduril–Heaviside equivalent in every jurisdiction.

Again, because our advantage has been in our own backyard, we haven't made an investment in a non-U.S. defense tech company. That said, Heaviside does have a sizable Norway office, so we have some exposure, and Castelion and Neros will also, through ITAR, of course, under oversight from the Pentagon, sell to our allies.

8. Physical AI and Shaping the Future

Nice. In terms of other sectors, you obviously named your firm Cantos. For those who don't know, that's the verse of an epic poem that tells the hero story—the ones who shape the world. I think you said recently that, 10 years in, you actually believe more than when you started that the future can be shaped.

I'm curious: what's being built right now, whether a company or a sector, that makes you the most optimistic about the next decade?

Absolutely. I wanted to name the firm as an ode to founders as the heroes of the modern-day epic. I know, because I've been shown by our portfolio founders and friends, that the world can be improved and the future can be shaped. You can go from an idea to something that has geopolitical consequence, and that's extremely compelling.

If you have that kind of capability, you'd be remiss to work on something as neutral or nefarious as digital slot machines. That's kind of the point. Look at our website: you see things that we want these companies to be etched in stone and have stories written about them in 1,000 years.

We have to be careful about how we steward that technology because, as we've seen, it can go off the rails and have negative externalities, where society is transformed and politics are different because of wanting to connect friends on the internet. Not to throw blame necessarily, because I think a lot of this is unintended consequences, but I find that when you're shaping the real world, at least there's a little more—the vision is a little more definite as to what that's going to do.

Some of the things coming down the pipe that we're excited by—in addition to the fact that I think there's a lot of value to be created in improving existing industries—include bringing more technology and robotics into lumber, which goes into housing costs, and into food and water systems, and, of course, into defense tech.

I'm very excited again about physical AI. I think if there's some ChatGPT-3-equivalent moment with robotics, that's going to transform the world in a way that we've never seen before. Basically, all technologies are producing either energy, information, or movement.

We've done a lot on the energy front. There's a lot more to do there, and some of our portfolio companies, like Radiant, Cache Energy, and Arbor Energy, along with a couple of new, unannounced investments, are improving energy efficiency and generation.

But I think you've got a lot of opportunity in the movement category. If we bring artificial intelligence into the physical world, just think about how your home life and industries can be transformed by this.

If I had a robot that would pick up after my toddler, oh my gosh, that would be life-changing. We're fortunate in that we can afford someone to help with our kids, but most people can't. If you don't live near your family or they can't help, and you have a robot that's cheap and can be a full-time nanny and maid, that is going to be absolutely transformative for society.

So we're investing in everything we can in robotics, everything from the model layer to dexterity and actuation systems that underpin it. I think this is possible. I'm not saying it will happen—the future's uncertain—but it's possible that we have something that is more transformative than what we've seen in LLMs.

Very interesting—robotics. I'm interested in the robotics issue because it's so controversial. Some people are so bullish on it, and other people are very skeptical that this will happen, but I think we can leave that for another time.

One question I was curious about, unless you have something you want to say: I saw a poster, by the way, when I was in the Mission District this weekend in San Francisco. Someone had put up posters with the Terminator from the Terminator movies, with lasers coming out of his eyes, and there was a QR code to join a protest. It was called “Don't Build This.”

You think the movement against data centers is bad? Just wait until we start having artificial general intelligence in the physical world. People are this mad about chatbots? Oh boy, just wait.

There's incredible opportunity, but we also have to steward this in a way that isn't alienating people and making them think that their jobs are at risk or, God forbid, that we're creating some Terminator future. I think Hollywood's done us a disservice by giving us mostly negative sci-fi visions of the future, and people are overcorrecting for things like that.

To ignore that is a huge mistake, as the frontier labs have seen.

Agreed. I think it's run throughout this conversation and my interactions with you that you're very principled, as an investor but also as a person. You also seem like a great guy. I think you've made decisions, like having a hurdle rate in your fund, that don't necessarily—I don't think you're profit-maximizing, or at least you're a repeated-game profit maximizer, a long-term thinker.

I think you're also a man of faith, right? I saw something about this. You don't have to talk about it, but personally I've struggled with this over the years. Venture is quite—well, you're in finance, right? So to some extent, your impact on the world is derivative.

Especially now, with where AI is, there's a sense that these things would get funded anyway, regardless of whether I personally was there. I'm curious if you ever struggle with meaning in terms of uniting faith and your work, because I think the way you've built Cantos is very much about meaning: the website and backing these heroes of epic stories. I'm curious if you struggle with that at all, ever.

I think about it constantly. I'm paranoid that I'm going to invest in something and help bring something forth that has some negative externality we're not foreseeing. I spend a lot of time trying to think about that because I deeply care.

But I also believe that the act of creation is sort of divine in and of itself. It's magical that you can have an idea, turn that into reality, and change the world. This is incredible. I'm fascinated by that, but we also need to be careful about what we bring into the world.

The precautionary principle is so dangerous because it basically is absolute conservatism: because we might one day build something that might be bad in an unexpected way in the future, we shouldn't build anything. That's just nonsensical and defeatist. I can't stand that.

We have to take some risk, build some things, and be as careful as we can, knowing that we're going to make some mistakes. But to not try would be almost sinful in my framework. I spend a lot of time thinking about that, again knowing that there's some risk.

There's almost a subtweet in the name Cantos: we're drawn to stories bigger than ourselves, and I think companies that bring things forth into the world can do that. But again, there's always a greater story.

Nice. This might not make it onto the podcast, but I'm curious about the founders. Having been a founder, and considering the kind of people you're looking for, I feel like this level of intensity isn't necessarily healthy for the person. Often, you're not looking for the most balanced, happy people.

In some way, you're looking for disturbed people, right? People with a chip on their shoulder—traumatized people, to some extent—who are willing to put so much of themselves into this thing and attach so much meaning to it.

How do you think about that? There's this post about Marc Andreessen: “I don't want my founders to do psychedelics,” because—and I don't want them to introspect—and I think the subtext is that they might realize they don't want to be spending 16 hours a day building whatever it is they're building.

Obviously, some things are more meaningful than others, but certainly there's somewhere where it would be easy to make that argument.

Yeah. It was more like there's an off chance that you could have psychedelic-induced psychosis, and that's too risky.

So, I don't want my founders doing it.

I don't know. I think it was more like some of them just don't work as hard once they do psychedelics, right? Maybe they have other priorities. Maybe they start looking at some of the sources of their drive and questioning their ambition—things that are potentially healthy for their self-development, but not really for you as an investor. So I'm curious: how do you think about that?

I have a working hypothesis: if you find yourself wanting to hang out with a founder too much, they might not be good to invest in. I've heard this from another excellent investor: it's a red flag if you feel like you want to hang out with them too much. There's something slightly awkward and off about the very best founders; you don't fully jibe with them.

They're kind of weird, and they push things to an uncomfortable degree. You have not just Elon but Steve Jobs as examples—arguably the 2 greatest entrepreneurs of all time. You could find a lot of negative references about both of them, and I don't know that I would want to work with or for either of them.

Maybe Jensen Huang is a notable counterpoint, and I think he's understudied as an entrepreneur. You can make an argument that he's on par with Steve and Elon, and my understanding is that he's maybe a kinder person. So there isn't just 1 way; you can't overfit to 1 archetype. A lot of the qualities that allow you to change the world are not necessarily positive social adaptations.

There have been some very successful people I've seen over the years where you can take this so far that, okay, I don't want to hang out with all my founders all the time, but we're also signing up to work with them for a decade-plus. Sometimes I meet someone where I'm like, "I just know you're going to be successful, and I don't want to have to put up with you for 10 years." So there have been a couple I've opted out of because of that, but I largely agree with that framework.

I study history's entrepreneurs, not just in technology. David Senra's Founders podcast is an incredible way to do this, and I binge as many of those as I can. One of the reasons I love it is that, having packed in all these biographies, he observes commonalities between them. One constant observation is that at the end of their lives, these great entrepreneurs often regret that they didn't spend enough time with their children.

Yeah.

I don't know that that's avoidable.

I think about this a lot as a father now.

Yeah. I think it's also that I'm told I work harder than other VCs, but I don't think I could start Cantos today and be a present father. I think about that a lot too. I think you sort of owe it to the world to think about this once you've reached a certain level of success, because you're finding these founders with this deep intensity that often comes from trauma, right? Something—some lack. If they succeed, they're going to be powerful people in the world.

Yeah. Sorry.

I want them to have faced adversity. I wouldn't index so far on the Hemingway-esque trauma thing, but I think what they're right about is that you have to have overcome some adversity. I don't necessarily want you to have been traumatized, but I want to know that you're incredibly resilient and ideally antifragile.

You're like—

You see this in very competitive athletes, right? My wife was a very competitive distance runner. She's still very fast, but she thinks she's slow by her own standards. She competed at the Olympic marathon trials and placed in the top 75 in the country—an extremely fast, elite athlete.

And she and all her runner friends are a little bit crazy. There's something in their psychology where they kind of enjoy the pain.

100%.

You have to be antifragile in that way—that the stress actually makes you better and you enjoy it in this, again, maybe socially maladaptive way.

Yeah. No, it's something—I definitely see it in myself. It's an interesting one because often your strength, this drive, can also be something that causes you a lot of pain. It's a weird one when you're looking for that in people you invest in, and you're looking for them not to lose it.

I don't think it's always the case, either. You can take this framework too far. I think it's the case for some people. I think other people find a lot of creative joy out of just building things, and it comes out of them in a very beautiful, natural way.

This has been really interesting. Maybe the last thing, just to end on a more positive note: what's the thing you're most proud of outside of Kantos that you've done?

It might be a bit of a cop-out, but certainly my children. They bring me so much joy. I have a 2-and-a-half-year-old girl and a 5-month-old boy—he turned 5 months today. They're just joy incarnate. I wish I got to spend a little more time with them, but such is life.

Interesting. How do you think about that? If Kantos reached a certain level of success, would you continue working as hard? Let's say you didn't have a financial need to support your family—that was taken care of. Would you still be working as hard on Kantos, or would you prioritize more time with your family?

I feel like I have the best job in the world, and there will come a point where I feel like the LPs don't need to pay me to do it. But as steward of Kantos, I feel a responsibility to build the most elite team I can to back these entrepreneurs.

Just because I happened to be the first guy who put the name on the door and got the ball rolling doesn't mean that I get to have a huge chunk of the carry forever. If I get to the point where I start feeling like a weak link, then I've got to go. Benchmark-style, 100% of the economics will go to my partners. I'll stay on boards, be an adviser, and stay around and help, but be an adjunct member of the team.

I hope this is 20-plus years from now, but I think your obligation is to make sure that you are the worst GP that your firm has ever had. I love building a team. I'm still the only general partner. My partner Grant, my principal, and I act almost like an equal partnership because that's directionally where we want to go.

At a certain point, maybe naturally, I'm like, "All right, guys. You got this. I'm only weighing you down at this point. I'll stay on my boards. I'll be a phone call away. I'll help with fundraising, investor relations, whatever. But you don't need me anymore." Again, I love this enough that I hope that's a long time from now. When that happens, I'll probably keep investing on the side. But that, I think, is the way investment firms should be run.

That's awesome. You've built a great team. Everyone I've met on your team is awesome. Thanks so much for taking the time, Ian. You're one of the best people in venture, and I think you're building an awesome firm. I'm really excited to be an LP and to follow along on some of these investments.

Well, we're going to find out together. Not that I've figured everything out, but it is nice to be able to reflect on 10 years and share some of those lessons and hopefully help other people speedrun it, so they can learn them faster.

Absolutely. Thanks so much.

Go USA. Go Portugal.

Let's do it. [laughter]

Later, José. Later.