Invest Like the Best with Patrick O'Shaughnessy
· Patrick O'Shaughnessy, Mitchell Green
[00:00:00] S0: So decompose and expand on that a little bit. So I guess the question is your view on the state of markets and technology markets in general. [00:00:00] S1: Overhyped, overfrothed, and I believe this AI CapEx bubble will end badly. It's like the telecom bubble all over again. It'll be very interesting if Apple may look like the really smart one in all this at the end of the day. We’ll see, but I think people are just going to overspend. I'm convinced that people investing in all these AI companies, all these VCs, have to portray the view that software is going to be dead because they have to justify how much money they're going to spend. If you start to run these assumptions on how much money is going into these companies and what that means for how much earnings you have to drive, and what that means for how much power you need to generate, it just doesn't work. Where are the nuclear power plants coming up? It just doesn't work. That presents the opportunity. That's when you're going to buy it. That's when you're going to buy these companies. [00:00:00] S0: The counterargument would be in telecom, it was all dark fiber. In AI, it's all burning GPUs. And yes, the CapEx is crazy, but everything still feels mega undersupplied. I'm just curious how you think about when the opportunities will present itself for an investor.
· Patrick O'Shaughnessy, Mitchell Green
[00:00:00] S0: The counterargument would be in telecom, it was all dark fiber. In AI, it's all burning GPUs. And yes, the CapEx is crazy, but everything still feels mega undersupplied. I'm just curious how you think about when the opportunities will present itself for an investor. [00:00:00] S1: My fundamental belief is that the models will commoditize, and that companies like Google and Facebook and Amazon and Apple have a competitive cost advantage. Companies like Amazon and Microsoft and Google have more data to train a model than these new model companies will ever have. And then oh, by the way, if you were all these Chinese models or European models, a bunch of these things cost a fraction of the cost to run and you can run them locally. Especially companies outside the US. Like, why would you pay that amount for OpenAI tokens or for Anthropic tokens when you can just run DeepSeek or one of these other 10 models? I think we worry the most about model commoditization. I have no clue when this will stop. It will probably go longer than people think. In '99 and 2000, people also thought we were in a bubble. People think we're in a bubble now, and it will just stop. Is it one of these monster IPOs happening and it just doesn't go like people think it does? I think this Anthropic round was kind of like an IPO. We're trying to hit doubles and triples. A lot of these companies we struggle with, they're either going to be 200x's or 100x's or zeros. That's the struggle for us.
· Mitchell Green, Patrick O'Shaughnessy
[00:00:00] S1: My fundamental belief is that the models will commoditize, and that companies like Google and Facebook and Amazon and Apple have a competitive cost advantage. Companies like Amazon and Microsoft and Google have more data to train a model than these new model companies will ever have. And then oh, by the way, if you were all these Chinese models or European models, a bunch of these things cost a fraction of the cost to run and you can run them locally. Especially companies outside the US. Like, why would you pay that amount for OpenAI tokens or for Anthropic tokens when you can just run DeepSeek or one of these other 10 models? I think we worry the most about model commoditization. I have no clue when this will stop. It will probably go longer than people think. In '99 and 2000, people also thought we were in a bubble. People think we're in a bubble now, and it will just stop. Is it one of these monster IPOs happening and it just doesn't go like people think it does? I think this Anthropic round was kind of like an IPO. We're trying to hit doubles and triples. A lot of these companies we struggle with, they're either going to be 200x's or 100x's or zeros. That's the struggle for us. [00:00:00] S0: What kind of company in the AI center of the heat map – I know you're probably not investing in any of them because of multiples or whatever. What kinds of companies are the most interesting to you just as an enthusiast?
· Patrick O'Shaughnessy, Mitchell Green
[00:00:00] S0: What kind of company in the AI center of the heat map – I know you're probably not investing in any of them because of multiples or whatever. What kinds of companies are the most interesting to you just as an enthusiast? [00:00:00] S1: I think it's fascinating some of the stuff that's being done in infrastructure software. And actually agents appear to consume more resources than actually people. Some of these consumption-based models, the growth of companies. By dumb luck, we were very early investors in ClickHouse, which is a database company. We were early investors in Grafana Labs, an infrastructure company that competes with Datadog. I think Datadog's growing like high 20s, 30% a year at scale. It's those types of companies I think we find super interesting. I find them fascinating. I really struggle with the valuations, but the growth rates are like we've never seen, with very good economics. You see how much money a company like ClickHouse has raised, what they've burned is very little compared to what you might otherwise think. [00:00:00] S0: What do you think is the most surprising thing about you? I have a good sense of you – how you operate, persistence, enthusiasm, energy, process. What do you think if I spent 10 hours with you, I would be most surprised about?
· Patrick O'Shaughnessy, Mitchell Green
[00:00:00] S0: What do you think is the most surprising thing about you? I have a good sense of you – how you operate, persistence, enthusiasm, energy, process. What do you think if I spent 10 hours with you, I would be most surprised about? [00:00:00] S1: Probably how driven I am and how much I truly love what I do. And I just put my heart and soul into everything I do, whether it's like racing cars, which I race cars competitively. I was a nationally ranked ski racer. Or how I run Lead Edge. I probably sleep 5 hours a night, 4 hours a night. It's 'cause I love what I do. I'm insanely competitive. If you spent 10 hours with me, you'd be like, oh my God, this guy is the most persistent, competitive person I've ever met. [00:00:00] S0: Were you born that way? [00:00:00] S1: Yeah, I think I was born that way. [00:00:00] S0: Was it enhanced through formative early experience? [00:00:00] S1: Ski racing, skiing growing up as a kid, 100%. [00:00:00] S0: Can you make that tangible for us? What was it like?
· Patrick O'Shaughnessy, Mitchell Green
[00:00:00] S0: Can you make that tangible for us? What was it like? [00:00:00] S1: Process. Do these things and you'll get better. Do these things on video in a GS course and constantly analyze video and do these things the next run and change this. And you fell, get up and go do it 10 more times. I grew up on a ski hill that was 500 feet. I mean, Lindsey Vonn, one of the best skiers in the world, she grew up skiing on 500 feet Buck Hill in Minnesota and doing laps from 4:00 PM to 10:00 PM at night. Just repetitive. Mikaela Shiffrin is one of the best female skiers in the world. She views it as, her time on snow is limited. When you get off the chairlift, everything is a drill, constantly be trying to improve. I think that's it. At Lead Edge, what you would find in me is constantly trying to improve. What surprised me the most, actually, if you had to say like, huh, you started the firm 15, 20 years ago. I think I've been able to recruit and maintain, motivate, and build a really good team. I mean, very good to pick really good partners that treat other people really well, and that feeds on itself. [00:00:00] S0: Is there anything else from skiing – I'm not a skier – that you find visceral and helpful as an analogy for how to do things elsewhere other than reps and practice?