How This Multi-Millionaire Investor Spots Billion-Dollar Ideas | Term Sheet

13:40Josh Wolfe:But the virtue of this is a prior, you don't know what companies are going to work. If somebody would have said in the late 90s, you know what, we have too many search engines. Like this is enough already. look smart and AltaVista and Yahoo, and then all of a sudden Google comes along. But the people that funded Google not only helped change the benefit of society for searching through these needles in a haystack, but made fortunes for everybody that was involved. So I think that it's a feature that we get an abundance of companies, lots of experiments tried, 99 % of them are going to fail, and the 1 % that survive go on to make fortunes and change industries.

14:16Josh Wolfe:So that will persist forever. It's one of the great things about capital markets. It's why the United States has the wealth that it does compared to Europe, embracing failure. And nobody wants to be on the side of failure, but it's an inevitable feature, not a bug of the system.

14:31Allie Garfinkle:I was going to say, so basically what you're saying is my exasperation is in response to something that is ultimately a feature and not a bug, and the cycle will evolve.

14:37Josh Wolfe:The cycle will crash. You pick your sector that you're like, oh my God, I'm so sick of hearing about the 12th drone company or the 12th, you know, 200 ,000th AI company. Only 200 ,000? Yeah, 99 % of them will fail. But the 1 % that survive, and you're already seeing that, by the way, the concentration of capital in what people are deeming the winners, particularly in AI, is very large. So it's a feature, not a bug.

14:59Allie Garfinkle:I was going to say, where do you think the AI bubble goes? What does it look like on the other side of this? Because it sounds like you're like, yes, of course, it's a bubble. It's going to pop. That's the way it works.

15:06Josh Wolfe:Yes. You've got your a handful of major labs, number one. And I think that they are reaching sort of peak utility to the average consumer, meaning the incremental spend that they will have to make ChatGPT 5 or 6 or Sonus 4.7 or Opus 4.5.1, it's marginal. There are new sectors that you can go into, particularly in scientific productivity, things in aerospace and defense, which have a lot of controversy around them, which are giant markets. So I think the push of AI into industrial is going to be a big thing. The bigger contrarian take is that I believe there are going to be enormous hundreds of billions of dollars lost and trillions of dollars of market cap that will be wiped out.

15:50Josh Wolfe:Because the presumption that we need endless demand of data centers and hundreds of thousands of H100 chips, the NVIDIA chips inside of each of these, and the high bandwidth memory that's attached to that, and the power that is needed for that, including the speculative sources of power, I believe that that is wrongheaded. And what the wise do in the beginning, the fools do in the end. And what one company does rationally, collectively becomes irrational. And that's what happens when you have a bubble. What comes next is I believe, and we've backed a company in this that will come out of stealth too.

16:19Josh Wolfe:So I'm talking my book here, but it is a bias rooted in a directional arrow of progress. And it's actually a founder that we backed in the past. And his insight was, We can't compete to make the next LLM that is going to do next word prediction or basing an inference or summarization or conjuring your poem or your artist or your video or your image. But let's look at something else which has a very low power envelope. Let's look at a fruit fly. A fruit fly is ridiculous. This little thing can navigate this room, can evade us as a predator or as a threat in trying to kill it, can find a mate, can find a food source, has local memory to go back to that.

16:54Josh Wolfe:and it does it on microwatts of power with a brain and a cortical structure that is the size of the tip of a pencil. So instead of going larger and larger in scaling laws and more compute and more power, why don't we go the other direction and not do this for an expert prediction, but for the future of embodied intelligence in drones and robots and whatever new devices are going to come that can navigate the real world. They are able to detect shadows and quickly evade. They don't put a prompt in and wait four seconds for it. It's instantaneous. And so if you look at the design of the cortical structure of all kinds of animals through evolution, from an insect, from a fruit fly to a mouse, to a dog, to a cat, to a human, there's 20 watts of power.

17:34Josh Wolfe:And this thing doesn't flex up on top of my head when it has to be pre-trained or post-trained. It's just 20 watts of constant power. And we are the ultimate one-shot learners. So I think there's going to be a big push into computational neuroscientists that are reverse engineering the cortical structure of the brain to come up with more efficient chips and algorithms for the next wave of AI. And it's going to blow people's minds.

17:55Allie Garfinkle:I was going to say, that actually doesn't sound crazy to me at all, especially because fruit flies, the byproduct of the research results of many Nobel prizes, right? Also share a lot of genetic material with humans. And I think if you, like, I don't think it's an accident that someone like Demis Asabas, for example, has a PhD in neuroscience. So does

18:10Josh Wolfe:Dario and so does Reardon, who we backed in the prior company.

18:14Allie Garfinkle:Who is known only as Reardon.

18:16Josh Wolfe:Only as Reardon and is the founder of this new company that'll come out soon. But yes.

18:20Allie Garfinkle:Tell him to call me.

18:21Josh Wolfe:Yes, I will. If you're listening, call me. He will. But that is to me the directional arrow. There's another big trend that I think is really interesting that I've dubbed life courting. And this is one of those directional arrows of progress. Again, it also has an age and demographic discomfort and comfort with it. Life courting is what I've dubbed for all of the devices that are going to come. At first, it started as a speculation, but again, it's an observation now.

18:45Allie Garfinkle:C-O-R-D, I understand.

18:47Josh Wolfe:Yeah, like recording your life. So it's an eilogism of life courting. It is the thing like the pendant from Limitless. That was a small device. It basically was a magnetic clip. Half of it was a magnet and the battery. Half of it was a microphone. So that got me interested. Okay, what is this thing doing? You're able to passively record your entire day. It's got enough power that it can last passively for 12 hours and can record everything throughout your day. My general counsel and most older people are really discomforted by this idea. Young people who are digitally always on, it presents really interesting moral questions of how you behave and whether this information will be revealed and that kind of stuff.

19:23Josh Wolfe:Amazon then bought a company called The Bee, which was a wrist-borne device that did effectively the same thing. There are about four other companies that have started. And then you have the big labs. So Johnny Ive at IO, which is OpenAI's hardware element division, they are most likely going to have a always-on set of recording devices that can integrate with OpenAI and ChatGPT to tell insights about your day. Who did you see? What did you talk about? Who mentioned that book? Or what meeting was I in when people said that? People are already comfortable with the laptop-based or phone-based Otter or Granola for recording and transcribing and summarizing meetings.

19:58Josh Wolfe:There will be this trade-off between privacy and productivity, which is a perennial trade-off in technology, whether you give yourself over to the information gods, whether that stuff can be used for discoverability and litigation. It will have complex set of considerations, like all technology does as it evolves, but this is an inevitability. Where does that lead me? The computational components that go into these devices, as you have hundreds of companies that are going to be producing and competing them. And again, 99 % of them will fail. But who will be the beneficiaries? The people that are making MEMS microphones, digital signal processors, novel capacitors, and new batteries.

20:30Josh Wolfe:It's a new segment of technology where one or more of these companies, some of them publicly traded, will get a demand shock because Meta will say, we need 50 ,000, or we need 100 ,000, they'll say, oh my god, we can't make that. And they're going to get pricing power. Somebody's going to do a teardown of one of these devices, post it online. Stocks are going to rocket. It's going to become a new phase. So I'm calling it life courting. I'm of high conviction and Lux has been making investments in the technological and scientific roots of this for this next phase.

20:56Allie Garfinkle:I want to pivot to defense because you were an early backer of Andrel in 2017. What was that conversation like? Because I understand you got a lot of pushback at the time.

21:07Josh Wolfe:Well, the conversation it started, and again, I want to be intellectually honest, did not start with an analysis or a DCF or some consultant report about, you know, this hockey stick growth in defense. It started with a brunch. Actually, we debate whether it was brunch or breakfast, but it definitely was at Balthazar. And it was me and Trey Stevens, an amazing friend and a partner at Founders Fund. And he said, I'm going to be starting this company with Palmer Luckey, and it's going to be a more kinetic hardware driven style of defense instead of Palantir. And I said, on the spot, I'm in. We're in.

21:33Josh Wolfe:And so we became founding investors and it was controversial because at the time Trump was talking about let's build the wall One of the first products that they wanted to do was take a whole suite of technologies and put them into a form factor That was basically a virtual wall that would give border protection with the century towers Which were actually using became century correct And it was using technologies from one of the other companies that we had back that was doing radar Using this crazy technology that itself was out of sci-fi called meta materials A company that we had started with Bill Gates and intellectual ventures and they were integrating that into the device It was the first company that we had to contend with the moral question.

22:07Josh Wolfe:Are you comfortable funding something that will be involved in a kill chain? We had been investors for over 15 years in all kinds of intelligence community facing things, defense things that were dual use, but never something that said, are you comfortable being involved in a kill chain? We had two partners that were opposed to this. We ended up having to off-ramp these partners. It was sort of a tragic thing. But morally, we had this debate internally. And I felt very strongly, the greater your technological precision in any kind of military affairs, the greater your moral precision. If you have the ability to discriminate between a guy that is coming home from aerial data or drone, a guy that is coming home from a Mine or a farm with a pickaxe on a shoulder and a guy that is coming with an AK-47 to shoot up a school of girls That is a clear moral discrimination that you want to be able to make to in one case prevent human suffering And in both cases to prevent human suffering to not make the wrong action and to make the right action And it's unfortunate there are bad people in the world But you do want good people with better technology fighting the bad people with ideally worse technology And so we became comfortable with a moral question that the good guys should have the best technology.

23:09Josh Wolfe:And if we sat back in that debate or argument, then the bad guys would continue funding, you know, sort of asymmetric threats. And so, yeah, we became founding investors in Andrel. And then from there, SailDrone, Hadrian, Varda, ImpulseSpace, every domain that you can imagine, air, land, sea, space, subsea, and now even subterranean, underground warfare. We have also funded UK, Israel, Bulgaria, Japan. We haven't yet done South Korea or India, but I want to. And then we've also done one in the Middle East, particularly in sync with the United Emirates, which I think are amazing and very bullish on Western defense in support of Western values, which are both being kinetically and ideologically attacked.

23:50Allie Garfinkle:Well, you know, for me, it took me a while to get comfortable, to your point, covering defense tech because I was very comfortable with, say, reading about World War II, but was I comfortable thinking about war and technology that could kill? I've thought about this a lot. And I actually asked, I recently profiled Brian Schimpf, CEO of Anderil.

24:07Josh Wolfe:Amazing, amazing leader.

24:09Allie Garfinkle:And I actually asked him what I'm going to ask you. How do you explain what some of these companies do to your kids?

24:15Josh Wolfe:Not only do I explain it, I show them videos. And I am so bullish on this next generation that wants to work on some of these hard problems. If you look outside in your lobby here at Fortune. The covers on the magazines as you enter this floor are 1942, 1945, 1951. I mean, extraordinary legacy of Fortune. But the covers are celebrating science and military and technology because the way that people were able to make fortunes over the ensuing five decades or six decades since were because we had this arsenal of democracy, because we were celebrating the people that were inventing the technologies to protect this country, our economy, our way of life, our markets.

24:57Josh Wolfe:So I feel very strongly that if anything, this past generation, because we had one war of choice and one war of necessity, you could argue, as Richard Haas might call it between Iraq, one in Afghanistan, the war on terror. There's a lot of people that rightly have been disenamored, people that were disenamored post World War II and Cold War and Vietnam. But the forces that we are fighting today, even if you look at Iran, these are not good people. We want to see positive change and security in America. And you need cutting-edge technologies. And I believe that is on the force of moral good.

25:31Allie Garfinkle:Well, and I will say the thing that really was the deciding factor for me was when I realized that a lot of men and women in uniform were going into conflict zones with technology less advanced than they used when they ordered an Uber. That really blew my mind. And I think it blew a lot of minds of investors, as evidenced by the fact that defense went from something that was deeply controversial that you didn't really do, and now it feels like everybody wants to do it. In my crazy notebook, I have written defense tech bubble, question mark. Yes. It is a question asked in sincerity, but do you think there's a bubble?

26:06Josh Wolfe:Not yet, but soon. And I will tell you, by the way, 11th commander of SOCOM, Tony Thomas, who we ended up bringing on to our team, sent me out to Indo-PACOM for nearly two weeks and embedded with Special Operations Forces, six different countries. And this was right before COVID. And he said, I want you to come back to the Pentagon and report back to me. And to your exact point, I remember as simple as five operators in the Philippines that had the inability to do what they call blue dot tracking, what we're able to pull up our iPhone and basically see where your friends are, you know, find my friends.

26:34Josh Wolfe:They could not do that because they were under triple canopy in the jungle. And I had three different companies that had technologies that could provide secure environments for them to be able to do that. And I remember coming back and saying, if I was your adversary, I would almost be like the guy with the red stapler in Office Space, that old movie. And I would basically do everything I could to keep your systems exactly as they are and make no change because that's how I would beat you. And he wanted to punch me in the face. And he later said that. And then he would join Lux. And he's been absolutely amazing.

27:02Josh Wolfe:But I do believe that it is unfair that the women and men that are on the front lines do not have the best technology, and they should. Some of that is bureaucracy. Some of that is capture from the Beltway Bandits and the large primes. And you are, for the first time, really in 50 years, seeing with Anderil and Palantir and a handful of others in SpaceX soon, the breakthrough with the next generation saying, no, we want the better technologies and better technologists who we trust. So that, I think, is very important. Bubble. I remember being with a very prominent investor who said, you know, you guys were early to defense tech and you will benefit from a tailwind of capital, but you will also benefit from a phenomenon, which I think people don't appreciate that will lead to a bubble.

27:40Josh Wolfe:And that is that some of the allocators, just like you as a journalist, were discomforted covering the space, were discomforted investing in the space. So they would say, we won't invest directly in Andrel. Now they all want to, but back then they wouldn't. We're okay investing in Lux because they were diversified fund, but we wouldn't invest in a defense tech exclusive fund, but we would invest in a fund of funds that invests in a Lux or defense fund. And so you would get demand for pools of capital that would form that a large endowment or a foundation would give money to, who in turn would give money to, who in turn would give money to.

28:11Josh Wolfe:And so that is going to lead to an abundance of capital chasing too few deals. And I do think that even if you look at the Palantir valuation, and Alex Karp is amazing and Palantir is extraordinary. And the people that work there, Mike Gallagher and Jacob Helberg, who was there in USA, these are extraordinary people. But the valuation today, to me, represents the pent-up demand that the market has for the next wave of defense, because their alternative is investing in Lockheed or Boeing or Raytheon or L3, and they don't want to do that. So we will see a bubble, but I'm usually wrong in the timing, and I think it'll probably be two or three years from now.

28:44Allie Garfinkle:In late 2021, you said accurately that 2022 will be a punch in the face. apply that same logic to 2026. What do you think will be true later this year and in 2027?

28:56Josh Wolfe:Well, my crazy speculation, which is not self-serving, is I believe we will see a headline, possibly the end of this year, maybe early next, which is a sad one, but I think a realistic one, which is I think Warren Buffett dies. And I think that the headline says something like, reports of the death of Buffett were not exaggerated, but the reports of the death of value investing were. And what I think happens is we see a 10 % or 15 % decline in the NASDAQ, very heavily concentrated in the MAG-7 or the MAG-10, whatever you want to call it today. And I think that you see those markets down 10 % or 15%.

29:29Josh Wolfe:Not crazy market crash, but enough that the concentration of capital, also driven by a structure of market, largely passive index buying, which has an algorithm in itself, which is dollar in buy indiscriminately in a market cap weighted index dollar out sell indiscriminately. And I think you end up with a handful of value investors, which is not my world, that are up 10 or 15%. And that's stark contrast. And I don't know if it's going to be a Dan Loeb or an Einhorn or my wife, Lauren Taylor Wolfe, or some other value investor, but it might be some new person that people suddenly say, oh my god, how are they outperforming by 300 basis points by 30 % when the market's down 15 and they're up 15?

30:08Josh Wolfe:And I do think that we see a scenario that would be unexpected in that shift from passive growth, which has been dominating for the past 10 plus years, into active value. So that's something that is not self-serving for me as a venture investor, but I feel is a lower probability, high magnitude outcome in market structure that is likely. And the other thing that I think is going to happen is the Strait of Hormuz, which everybody's mostly focused on oil and Iranian control. You could have a scenario that I would put reasonably low odds on, let's call it 10%, which is still maybe high, but huge order of magnitude of salience and importance, where fertilizer scarcity leads to a food crisis on the periphery of Europe, which is unexpected, maybe in a place like Macedonia or Bulgaria or something like that, compounded with migratory pressures, and you get a economic and market and financial crisis in the EU.

31:03Josh Wolfe:And you have decline of currency, you have decline of markets and you have panic and then I think you get a a rise of sort of far-right strongmen that typically are there that was the most chilling thing in talking to Intel folks recently as I share this hypothesis I always want people to either poo-poo it or sort of validate or to give me some variant and they said something that the sort of Hitler or Mussolini or strongman of the 30s and 40s is already in Europe and they don't know if he's 35 years old or 45 years old or 55 years old, but he's there waiting for the stage. And that's chilling.

31:38Josh Wolfe:So that's my—

31:40Allie Garfinkle:So that's your very rosy prediction for the later this year in 2027.

31:44Josh Wolfe:But I will say, against that backdrop of dark cynicism and preparedness, because I always like to say that failure comes from a failure to imagine failure. So I've imagined a geopolitical and socioeconomic failure, and I've imagined sort of a bubble in defense and imagined a crisis in the NASDAQ. Life will always continue to get better because of technology. And while I am very cynical about people and people are unchanging in their Shakespearean motives and there's daggers and men's smiles and we're full of envy and jealousy and all of that kind of stuff and vainglorious you know this but technology compounds because of science and the scientific pursuit of and not born because people want to you know just discover stuff but because they're ego driven and they're competitive and life keeps getting better the idea that we have glp1s and gpus and rockets it's just it's like when you zoom out it is amazing and the march of progress will continue despite the daily headlines.

32:34Allie Garfinkle:I was going to say, I think your cynicism about people is maybe slightly more complicated because if you believe people are capable of all of that, you can only be so cynical. We'll have to leave it there. Thank you, Josh.

32:45Josh Wolfe:Great to be with you.

32:47Allie Garfinkle:And that was my interview with Josh. You know, I've always found Josh really compelling because he really has a perspective. Whether you agree with it or not, it is interesting. If you're interested in other companies that Josh has backed, we have a feature about Osmo AI, which is Shazam for scent. That's it for Term Sheet. I'm Allie Garfinkel. We'll see you soon.