00:00I'm so glad you're here with the thesis. I would just point out there's a lot going on in public
00:04markets right now. But actually, you know, private markets are quite active in physical AI. You know,
00:14I was tracking the data pitch book like 2019 to last year. It's going up. What are you saying?
00:20Yeah, well, thanks for having me. And it's great to be here. And of course, index and other folks
00:24in the valley have been spending a lot of time investing in the fundamentals of AI, which have
00:29been really based on LLM. So at index, we've done everything from foundation model of anthropic to
00:34inference layer of fireworks and then a bunch of application companies. But if you take a step back,
00:39actually, most of the disruption in software has really been around the knowledge worker. And that's
00:45where a lot of the investment has gone. But there's a big white space, which is what does AI do
00:49in the
00:49physical world? And that's where we've been spending a lot of time. Yeah. You know, for a really long
00:53time in many cycles, software is eating the world. And now people say, well, hardware is eating the
00:58world. You're basically saying both, you know, they are joint. Yeah. Well, I think if you look at it,
01:06there's all of these functions, electrical engineers, mechanical engineers, aerospace engineers. And for
01:12decades, they have been using the same software stack. And a lot of these incumbents were built in the 1980s.
01:18And there's this huge race, not just the race to get to space, but the race to put better things
01:23in the
01:24physical world. And unlike some of the things in the digital world, if you get one character wrong in
01:29these code, something could actually blow up like a nuclear reactor or a rocket. And so we're really
01:35interested in companies that are trying to solve this deeply technical, very domain specific problem.
01:39Most multimodal that trying to ground the model in real world physics, you know, in the case of
01:45robotics, you know, in my study of software, the VCs that come on the shows would say, we have
01:55conviction about this because there's a public market proxy, an example we can point to on how we value
02:01this company or the market opportunity for it in humanoid robotics and other physical AI domains, much
02:07harder to do. Is SpaceX going to be a good proxy for that? Well, I think that's why all eyes
02:13are on the
02:13SpaceX IPO. And I think that it's really opened the eyes to folks of just how compelling this market is
02:19and how
02:20lucrative it could be. And that's drawing a lot of dollars into the private sector in companies that are
02:25doing similar things. Yesterday I was on stage with Trey Stevens from Founders Fund. He made a very simple
02:32outline of a thesis on why SpaceX is going to work. Just we'll play the soundbite. We have a SpaceX
02:38investment in almost
02:39every fund, both across our venture funds and our growth funds. So there's a ton of exposure for LPs that
02:45were with us since the very beginning and even the ones that joined us very recently. I think, you know,
02:50the core lesson that everyone has learned from this experience is never bet against Elon. It's just a bad idea.
02:58Pretty simple. Don't bet against Elon Musk. It's a bad idea. I don't know, like, how deeply you read the
03:05S1 and
03:05you were and the world that is envisioned for the future or out of this world and how much that
03:12gave you
03:12conviction in your own thesis. Yeah, well, I think the way that we have played it is that Elon is
03:18a huge
03:18talent attractor. And so actually two of the most recent investments I've made are both in former SpaceX
03:24founders. So one is Scott Morden from Revel. He spent 10 years at SpaceX. That's right. Doing launch control.
03:30And his role there really exposed him to how big of a need there is for better software for hardware,
03:36which is why he's now built Revel. And then Sergei Nestorenko, who's the founder of Quilter,
03:40spent six years at SpaceX where he was really frustrated that he had to wait for a human to
03:45manually lay out a printed circuit board. Of course, printed circuit boards are really important to
03:50everything that's happening in ships. So he started Quilter to use AI to autonomously lay out PCBs.
03:55We just have 30 seconds. You know, I think that we've made the point that whatever happens with
04:01the IPO, it will cause some volatility in public markets one way or another. But from your venture
04:06capital industry standpoint, I won't ask you to speak specifically on Anthropic. How much do you
04:11guys need some big IPOs to get the world moving? Well, I think VC is always rooting for the IPO
04:16window
04:16to be open. It creates more options for our founders for capital and liquidity. And it's really
04:22exciting to see the ones that are lined up today. And so I think folks will be watching them very
04:26closely.
Comments