EVERY SPOKEN WORD
60 min read · 12,080 words- 0:00 – 4:21
Introduction
- DGDavid George
It was, like, a very simple premise when we started. It was like tech markets are bigger than ever, companies are staying private longer than ever. Uh, and as a result of that, uh, you know, the opportunity set for us is huge. Uh, I was looking at it last night, and I think, I mean, it kinda oscillates a little bit, but I think six of the most valuable ... I think the six most valuable companies are US-based tech companies. It's definitely five, and then sometimes, you know, it bounces around on number six. Um, and then, you know, it bounces around a little bit, but seven or eight of the top 10 are US-based technology companies. So, um, you know, technology has kind of swallowed the whole market, um, you know, and I think increasingly will take market cap, uh, over time. Um, we've got some slides showing this whole trend, and, you know, I guess Databricks was an appropriate way to kick off, um, you know, talking about the trend of companies staying private longer than ever. That's obviously a double-edged sword, you know, for us. Um, it gives us an opportunity to invest in companies more while they're in the private markets, but we also are very mindful, uh, about, you know, generating returns and, and DPI. And then, you know, the big thing that's changed from when we started the growth fund is just, is AI. Um, you know, we've got some slides on it. It's massively expanding the market. The AI companies are getting bigger, uh, faster than anything we've ever seen. The investment amounts are f- are bigger than anything we've ever seen. Uh, and so it's just ... It, it, it looks to be, you know, a huge tailwind for us over the next 10 or so years, um, you know, as we, as we look to make new investments. So let's jump into the details. So AI, I mentioned this already. You can go to the, uh, next slide. The groundwork is being laid, um, in a way that's very different than previous cycles. Um, and the, the groundwork that's being laid is bigger than anything we've ever seen before. So I, I'm all over the team, I'm like, "This is too conservative. These numbers are gonna end up way bigger." Uh, because I think just the big tech companies in their latest quarter, if you run rate their CapEx from the latest quarter, um, I think it's, like, $400 billion of annual CapEx, and most of that is going into AI, you know, AI infrastructure and data centers. And so, um, you know, what that means is the infrastructure is gonna get built for, uh, all of the training and inference needs that the market is gonna need. And th- and this is great for all the companies that are building on top of this. Uh, the best part about this is it's mostly the large tech companies that are bearing the burden of the build-out. Um, and so you, you've probably all seen the charts of, like, CapEx spend as a percentage of their overall sales. It turns out they're the best companies, you know, probably ever created. Um, you know, companies like Google, Facebook, um, you know, Amazon, and Microsoft, and they can bear, um, you know, potential capacity overbuild and, and things like that. And so if you just put it in a conservative view, which again, I think the number's gonna end up way bigger than this. So, um, so the build-out's massive, and this bodes very, very well, uh, for, uh, for our c- for our portfolio companies that are building on top of it. Next slide. Um, at the same time this is happening, the input cost and input quality is getting remarkably better, like, faster than Moore's law. So on the left-hand side, y- you don't need to look at the details of this. Just trust me when I tell you the cost of the inputs, um, you know, of accessing these models has declined 99%, or a little more than 99%, over the last two years. So, you know, sort of 100x declines, greater than Moore's law decrease. At the same time, the models have been improving in sort of frontier capabilities by a double factor every seven months. So massive decline in the input cost, um, at the same time that the quality is going way up, and this bodes really well for building new stuff and new capabilities, uh, on top of AI. Uh, you know, I think our house view now is that AI is gonna end up like, you know, electricity or Wi-Fi. Like, if you're, if you're, if you're getting ... If you're accessing, you know, electricity at somebody's house, you're not like, "Hey, let me chip in, you know, a few pennies for, you know, sitting in, you know, a room with light, uh, in your house." And I think it'll end up being the same thing in the fullness of time with elec- or with, uh, with, uh, with AI. Um,
- 4:21 – 26:48
The Market Opportunity for AI
- DGDavid George
next slide. The market opportunity for AI is so much greater than, um, than the software market, and I think that's really exciting. Um, if you look at the previous cycle that we went through of mobile phones plus cloud computing, the big story behind that was basically creating 10 trillion or so of new market value, uh, across software companies, internet companies, mega cap tech companies, and I think AI is gonna be much larger 'cause I think the impact on the economy is gonna be much larger. Um, and so, you know, if you look at the simple math that we have on the page, uh, US software spend is, like, 1% of GDP. US white-collar payroll is, like, 20% of GDP. And so, um, you know, there's a lot of areas where I think we'll see, you know, augmentation or, or potential, you know, cost savings or efficiencies or, or replacements, um, you know, using technology. There's always a question when these things happen of how much, uh, the new companies are able to capture versus the end customers. My rule of thumb is, like, 90% of the value goes to the end customers, um, and, you know, 10% of the value goes to the companies serving them, and it, and it turns out that that's just a massive amount of market cap, you know, if you're the 10% that you're capturing. You know, the examples I always give are like, you know, what, what does your iPhone cost? Like, uh, I don't know, the latest, like, give or take 1,000 bucks. Like, if, you know, g- gun to your head, what would you pay for an iPhone? Uh, like, you know, if you're, if you're on the higher end income spectrum, like, probably far greater than $1,000. Um, and, you know, the difference between that and the $1,000 you pay is, is the surplus. And it turns out Apple's still a really great business. Um, or, you know, if you take the Google properties, like, you know, Search and, and Gmail, and now I guess increasingly AI stuff, you-They monetize you per year. First of all, you get it for free, which is massive surplus, but they're only monetizing you per year probably, like, 200 bucks or something like that. Um, and there's a tremendous amount more value delivered, um, you know, I would argue, than that per user. So I think the big story is gonna be massive new surplus created. A ton of it gets captured by, you know, end customers, end users, whether it's businesses or consumers, and a massive amount of new market cap goes to companies, um, that are, uh, that are capturing that opportunity. If you, um-
- ETErik Torenberg
There was a great, there was a great meme the other day [laughs] that someone said, you know, imagine if Google had known that users were willing to pay, you know, 100, 200 bucks like they are with ChatGPT, if they had known that people would pay for that with a similarly magically delightful product like Google, like, God only knows what the market cap was [laughs] of Google would be today. Um, m- obviously maybe even more of a dominant, you know, market share contributor than, than it already is. But it is, um, it is amazing when you think about sort of the reconfiguration of how we will view monetization in this new world. And, and before we go even further, David, uh, I think these last couple slides actually set up probably 95% of a lot of the questions around the market. Because if you go back to slide, uh, seven, Monique, like, I think that slide is daunting for a lot of folks, in part because obviously you see it bundled with, with the headlines, but also because it just feels like we've all, many of us, had lived through the early 2000s where it just ended in a very, uh, less than desirable picture. So maybe if we can, you know, summarize, what, what's your case for why it's different? And then particularly talk about the timing cycles, 'cause incidentally enough, you know, despite the massive broadband build out and then the glut, we actually of course get the beneficiary of that today, but-
- DGDavid George
Well, we grew into it. I, I mean, we, we ended up growing into it, it was just a time lag in that case, and it was not the strongest companies in the world building that out. And, you know, the thing to watch will be, like, what role does leverage play? And so, you know, I read an article, uh, this weekend that was like, "Is there systemic risk in data center build-out?" And first of all, most of the, you know, people with their neck on the line, again, but, you know, there's a role that, like, as of right now, private capital is playing, there's a role, and, and, you know, the, the biggest funder of private capital is actually banks, um, or private debt. Um, so, um, you know, banks are the ones that are funding the private debt companies, and increasingly they all have insurance companies, so maybe there's insurance companies that are kind of backdoor, you know, funding this build-out. You know, that's a really good sign for the stability of the build-out. Um, so, you know, that's, like, the nature of the supply side, which again, it feels different this time given, you know, who's, who's actually doing the build-out, um, and, and, and who the tenants are. Um, the, the demand side is the bigger, more interesting thing. Uh, so I read a stat yesterday, um, that the time to get to 365 billion searches on ChatGPT was two years. The time for Google to get to 365 billion searches was 11 years, so it's five and a half times longer. So the big story on the demand side for me this time around is AI is built on the back of the internet and cloud computing, and because of that w- it sort of allows for immediate global distribution. Like, if you look at the way, you know, Google and Facebook started, for example, like, it started, you know, mu- mu- they started much more, you know, like, small build. Both had, like, a network effect dynamic, which just takes longer. Um, and, you know, we didn't have full sort of internet proliferation across five, you know, 5.5 billion people in the world, and, you know, smartphones, uh, in the hands of everybody able to access the internet. And so what that means is because of the nature of this technology, you don't have to deliver a new hardware product, and because we have global, you know, internet build-out, and because we have cloud computing, the whole world can access this. And so if you just take ChatGPT, again, they got to the scale that they're at five and a half times faster than Google, which is staggering. Um, but, you know, there's probably, I don't know, a billion, you know, I think the latest they said, there's more than a billion monthly active users. There's probably another billion or so people who have tried it. So if you add up all the different platforms, and a bunch of people have probably tried, you know, Google products and Facebook products, it's probably well over half of the global internet population has used AI tools already. Um, and, you know, we know that there's probably in some shape or form somewhere between one and a half and two billion user- active users of these products. So just the speed at which they got to distribution is unlike anything we've seen before. And so that, that is heartening to me that the supply build-out will be utilized, um, maybe in a more predictable way, uh, than, you know, broadband, uh, in the early internet build-out days. Uh, just 'cause it, you know, it's, it's, it's built on the back of the previous inter- infrastructure stuff. Um, you know, the, I, I, I also wanted to comment on the Google thing you said, which is, like, imagine if they, you know, could get 200 bucks or something from users. The beauty of, as consumers, the beauty of, like, Google, Facebook, Apple for us is there's not, like, a clean way to price discriminate for those companies. Like, uh, like if, if, if, you know, like, they, they don't know that I'm willing to pay more, and they can't charge me more than, you know, the other user, uh, of an iPhone. Um, but in the case of AI, uh, because of the way the business model is structured, and we've already seen some proof points of this, I think there will be greater success. So just today, um, or I guess yesterday-Uh, technically, uh, OpenAI released their India subscription product, and I think it's something like three or four bucks a month. That makes total sense. In the US, there are high-end subscription products that are 200 to 300 bucks a month that are, like, flying off the shelves, like, consumers can't buy enough of. Um, and so to me, the real story of the growth in this market is there's gonna be an evolution of the business model that allows these companies to address the user base and actually price discriminate, I think, in an effective way. So they can do a combination of subscriptions, you know, for higher-end users, um, and sort of get to the point, you know, where they're, they're willing to pay more. Um, and then also, you know, probably end up with freemium products where they monetize, it's through some form of advertising. It's, it's hard to speculate now on what that would look like. I think it's probably some form of, like, an affiliate-type thing, um, that has, like, a dirty connotation 'cause that's kind of a backwater industry in today's internet. But I think it will end up looking like that. Um, and I think, um, I think the way you see... Like, one way to see it in the product is, if you haven't done this already, um, go into, like, one of the deep research products and, uh... Sorry, this is Katherine Boyle calling me about a deal, I think, and I don't know how to turn this call onto silent. Um, so one way to try this out in the product is, like, go into, um, one of the deep research products, either, you know, whatever, OpenAI or, or Grok or, or whatever it may be, um, and have it do, like, a really sophisticated shopping research project for you, and it comes out with incredible stuff. Like, I had it, um, I had to do this for my son's literally baseball bat because it required a bunch of different specifications and things, and I wanted to look at year over year, like, you know, what, what's the, what's the better value and all these things. Um, and it came back with, like, extraordinary answers, and it-- this is a far superior, you know, experience than research- like, typing into Google and then clicking around and, you know, having seven sponsored links above, you know, seeing anything organic. So I think there's gonna be an opportunity for them to monetize free users. And so the big story, you know, in the case of OpenAI is, like, there's, you know, whatever, probably thirty to forty million paying users today. The other platforms are kind of a rounding error relative to that, so maybe add another ten. Um, so there's, like, forty million people paying for this stuff today, um, at some level. And, you know, there's probably two billion using it. So, like, I, the, you know... And, and again, Facebook, Facebook and Google monetize their properties at, like, you know, for US users, call it between 150 and 200 bucks a year per user. So there's just a massive amount of opportunity, uh, to monetize. And on the consumer side, there's probably gonna continue to be tremendous surplus. I find tremendous surplus in it today. Um, but, you know, active, daily active users of ChatGPT already today spend, like, 30 minutes, uh, 20, 28, 29 minutes a day on the product. And, you know, to put that into context, I think Instagram's, like, 50 minutes a day and, you know, TikTok, like, sadly is, like, 70 minutes a day. But, like, th- this is, like, real time spent, um, and, and, and real sort of consumer value already. So I, I just think... I know the question was about, like, risk on the supply side and the infrastructure build-out, but the usage, the actual usage and distribution that we've seen this time around makes me think that it's kinda different. Like, we, we have a really good, um, view of demand signals that is far, that, that took many, many years to get, um, you know, in the case of, you know, the internet, uh, or even in the case of mobile phones, just 'cause, you know, you had to manufacture phones and convince people to buy them. Uh, so, you know, i- in that way, it's built on the back of the previous technology cycles, and it-- and that's great, but it also, in my mind, de-risks, um, the sort of forward growth potential of the, of the AI, uh, companies that we're investors in.
- ETErik Torenberg
Totally. Pulling up the website of ChatGPT and, and doing it, being able to do it accessibly and almost, like, the party trick of, like, "Hey, let me show you the cool thing that I just did," like, that usability obviously is very different than we, than the prior cycle where we literally had to wait for the infrastructure to be built out and the device and hardware to catch up as well. Our, our early stage team also did, um, a great post on this topic if you're curious about the future of commerce, what that looks like with AI. And even the recent, if you, uh, uh, have any folks on, on this call who spend time on the public side, you know, one of the things, uh, that has been significantly observable is, uh, the number of public companies who've reported a decline in, um, referral traffic and engagement, um, largely because, you know, with Google Search now, they're just doing the summary, the AI summary version of the result below. And we could talk about the implications for what that means in terms of the downstream effects as a part of that, but that is, um, certainly something top of mind we're hearing a lot about from, uh, folks in the Fortune 500 just given, you know, obviously this reorient your business. Like, how do they engage with the consumer when you could actually do a really detailed search, as David described with, with his son's baseball bat, without ever actually having to go on any website at all?
- DGDavid George
I told my son, I'm like, "You have no idea how much research I did, man."
- ETErik Torenberg
[laughs]
- DGDavid George
You didn't-- I, I, I, I scoured the end of the internet for you, buddy. Uh, and it turns out it was, uh, it was all-
- ETErik Torenberg
All, all, all dads unite in, in their desire f- to, to have endless amounts of research on, on, uh, you know, kind of arbitrary things to say the least. Uh-
- DGDavid George
Here, I'll s- I'll send that. Uh, there, there was a good little, uh, snippet on X about some of the Google Search traffic stuff. It was, like, IAC-
- ETErik Torenberg
Oh, the parking one?
- DGDavid George
... and parking ruse.
- ETErik Torenberg
Yeah, the-
- DGDavid George
Oh, yeah. I was gonna say-
- ETErik Torenberg
I just saw the, yeah, the, the earnings call that, that Martine dropped the other day, uh, just so folks could see, like, again, it's the folks you usually would expect, right? Like, so folks like Groupon, for example, um, IAC, uh, who, who are seeing the impact in, in real time. Um, there's one question from Chris, uh, we should take while we're just on this slide here. Um, and I'm gonna throw in another one that's related to it around the, the shifts in bottlenecks, right? You know, right now there's obviously this massive bottleneck as it relates to compute, but Chris's question was, is there enough energy to actually power this build out as well? And, and, you know, we should layer into there what we think the next bottleneck actually is beyond energy as well.
- DGDavid George
Um, I mean, yeah, like, as of right now, through our current means of energy production, yeah, I mean, energy's a bottleneck. And so, you know, we've made investments on the nuclear side. I'm quite optimistic that, you know, we now have, um, you know, like, an embrace, I would say, of, um, you know, nuclear power. I think Three Mile Island's gonna get powered back up. The big tech companies are building data centers near, you know, nuclear power plants. You know, we have figured out there, there is a lot of natural gas in, you know, places like, uh, West Texas that, um, you know, can be... You can, you can, you can build training, large training clusters, like, very near to them, um, and, and pretty efficiently, uh, power those data centers. Um, but yeah, we're gonna find... W- we're gonna need, you know, different, uh, different sources, I think is the short answer, and, you know, we're probably most optimistic about nuclear.
- ETErik Torenberg
For sure. Yeah, definitely.
- DGDavid George
And then, and then the ability to just build these things. I mean, they're like massive scale operations. And so, um, you know, like, one of the most remarkable things, and we're, we're large investors in xAI, and, and one of the most remarkable things about what xAI did is they stood up the biggest data center at the time in, you know, like, a quarter of the time that anyone else had done the same thing, and they had to do crazy unnatural things like, you know, get, um, you know, every backup generator in the, you know, multi-state region, uh, bought out and, and, you know, buy, buy labor off of different projects, but, uh, but they did it. And so just, you know, the actual construction and build, um, you know, is massive. My, my view on, like, chips and infrastructure is production capacity of those will typically scale to meet demand. There's, there's always a dislocation and, you know, you've seen it. Um, and so, you know, I think energy ultimately in the next, call it, five years will probably be the bottleneck, and that's why we're so excited about nuclear and making investments in that area.
- ETErik Torenberg
Absolutely. Yeah. And, and just to, uh, extrapolate out, you know, once we figure out that piece, which inevitably, you know, with any technology, you, we always do, uh, it then the bottleneck just shifts to another. The big component that I, I think most folks have not yet, uh, realized or zoned in on is, is the cooling piece. Uh, and so you'll see a whole wave of innovation around that part as well once you figure out how to generate all this energy, how to actually cool all this stuff down without boiling our oceans and, and, uh, making the world melt down. Um, so-
- DGDavid George
And making the chips melt down.
- ETErik Torenberg
And making the chips melt down. Yes, that's right. There's one question here, David, if I can interject, uh, 'cause y- you are the business model snob, so, uh, this is a perfect question for you, which is there's a lot of debate on whether the gross margins of a lot of AI companies should be more scrutinized, i.e., particularly, you know, kind of, um... There's a lot of, of, uh, turmoil around, for example, the relationship between Cursor and Anthropic, and whether, you know, the growth for a lot of companies might be actually masked by, um, a reliance on some of these models, and also what is, um, the, the actual unit economics that is considered best in class. Maybe help us distill how you and the team think about that, and particularly, you know, what is this, uh, uh, topic around gross margins where you're willing to make, you know, some, um, short-term exceptions for versus long-term, you know, hopefully benefit and output on the other side.
- DGDavid George
Yeah. I, I love this topic. Um, and I, I, I would just say that, like, the reason that this industry and this job is so fun right now is because the range of outcomes is so much greater than before. Like, the variance is so high. You know, I, I was talking to the team at the offsite yesterday, and I was like, "You remember in 2000..." Some of them weren't in the industry, but, you know, the ones that were in the industry, like-
- ETErik Torenberg
But I weren't born yet too. [laughs]
- DGDavid George
I know. There's some that weren't bo- born yet. Yeah.
- ETErik Torenberg
[laughs]
- DGDavid George
So some young folks who are, uh, very AI native and very smart. Um, but um, you know, like, late cycle, there's questions that you are trying to answer that are interesting, but they're far less interesting. So you're like, "Oh, how, just how big can Datadog get?" Or, "Just how big can," you know, whatever. Pick, pick your, pick your SaaS app. Like, how big can it get? And, um, now we have all these questions around business quality, um, market power, who the winners are gonna be. Even if you are a winner, is there gonna be value that accrues to you and where you are in the stack? Um, and so just the range of outcomes is, is so much higher. And I think if we do a good job in that period of time, what that means is, uh, you know, hopefully we can get, you know, greater degree of variance i- in, in, in our own outcome, uh, you know, as investors. And so I'm very excited about that. Um, so yeah. So what, what are, what are we thinking about in business models? So one, uh, sort of value proposition to the customers is the number one thing that we care about. Like, is there customer love of your product, and is that love enduring? So, um, you know, if you, if you made me pick two top line stats to look at, uh, to assess the business model, um, it would be, um, gross retention rate. So gr- gross because, y- you know, I like looking at... We have- we obviously get to look at net retention rates too, but for gross retention rates, it's sort of like, are people getting value out of your product? And what gross retention is, is basically if you have 100 customers, um, you know, in absolute terms, you know, dollar weighted, how many of them are sticking around? And so we look for things where, like, 90% plus-Customers are sticking around. Um, and hopefully they're expanding their usage, and so that would be expressed in, in net retention. Um, but just sort of core value proposition is, is, is shown in gross retention. Um, and then ease of customer acquisition, and so the way you see that is, like, organic customer demand, you know, high value of dollars that they're willing to pay relative to how much it costs to acquire them either via marketing or sales. And so if you have things that are sort of like being pulled off the shelves and you have, uh, high endurance of the customer relationship, that's typically, like, the best things that you can get, um, for, you know, business model quality on the top line. Um, you mentioned gross margins, so we care a lot about gross margins, and there's a bunch of debate right now around some of the AI native application companies and their gross margins. Um, I think our hypothesis and hope in the market is if there are multiple model providers that are somewhat close to parity, you're going to see input costs go down significantly over time. But if you recall, you know, a hundred X decline in the input cost over the course of two years, the hope would be that that continues. And I-- all, all indications suggest on our side that that will continue, and maybe it abates a little bit, but it will continue pretty significantly down, uh, over time, as long as there's competition at the model layer. Um, and so, you know, coding is one of the areas that people have spent a lot of time scrutinizing in this area and, and, and sort of assessing the business, you know, business models and business quality. Um, I'd, I'd say relative to, like, mature SaaS apps, we probably are a little bit more lenient on assessing a company's gross margin today because we strongly believe that their input costs are gonna go down over time. And because of the model improvements, they'll be able to, uh, harness better models and deliver better products to consumers over time. So they won't need to increase price, but they'll deliver a lot more value and stickiness, um, while their input costs go down. That's the hypothesis. Um, I think that's subject to there being multiple players in the market that serve, you know, models. Now, we're thrilled that GPT-5
- 26:48 – 43:15
Pricing, Monetization, and Cash Burn
- DGDavid George
is out, and it's a very credible alternative, and it will put pricing pressure, uh, on Anthropic. Google is also very focused with their Gemini models on coding, and we've seen a bunch of really good improvements and promising progress out of them. So as long as there's multiple players in the market, I think you'll continue to see costs go down. And again, i-in light of that, relative to, you know, sort of mature industry type SaaS stuff or infrastructure stuff, we're a little bit more lenient, um, you know, on assessing a company's gross margins today. We, we, we don't wanna go invest in a bunch of companies with zero gross margins, and we don't do that. Um, but you know, if, if you, if you sort of took like gross margins, retention rates, and sort of like ease of customer acquisition, like I'd, I'd place far more emphasis on making sure that we feel like there's greatness in those latter two, um, and, you know, give them a little bit more benefit of the doubt that they can improve, uh, on, on the first.
- ETErik Torenberg
For sure. Yeah. I flip back to this because I, I, I've seen this personally myself like a dozen plus times, but, but something you said there really just clicked for me, which is to say like, you know, a lot of people try to make comparisons to the dot-com era, and they're like, "Oh, well, remember we measured eyeballs as well, right? Like, isn't that the akin to in terms of retention or usage?" And it's like, yes and no. It-- this is reaching such an en masse so quickly because of all the reasons you alluded to earlier, David, in terms of ability to just pull up a website and actually try it. But also people are paying for this. And it's an interesting juxtaposition when you think about the last cycle of things we paid for, so to speak, like a Spotify, you know, subscription or a Netflix subscription, where as soon as, you know, obviously pressures come from, from a budget standpoint, those are first to go. But this, you'd probably sacrifice a few things just given how much of an impact it has made either professionally for you personally and accelerated your productivity and, and hopefully time in terms of, of acceleration as a part of that. Um, there, there-
- DGDavid George
Thomas's question. I'm happy to take Thomas's question.
- ETErik Torenberg
Yeah. It's from-- this is from Thomas. So it seems like there's downward pressure on the likes of OpenAI on consumer pricing, yet the cash burn of OpenAI is ramping up to levels not seen before previously, you know, reports of a billion dollars plus per month. How does the cash burn moderate in the future relative to what you think in your opinion?
- DGDavid George
Yeah. So, um, so yeah. So, so what I was saying is like, I think effectively like there's greater consumer stickiness than you would think. And there have been a tremendous amount of free alternatives thrown at consumers over the last twelve months, and it hasn't had any impact on their business. Now, that could change over time, uh, but so far what we've seen is, is no effect that creates price pressure. And if you think about what I had described earlier, which is like call it a billion people using it and only thirty million people paying for it, I think there's way more upside to monetize the base than there is risk of price pressure, uh, on, you know, today's thirty million people paying for it. Um, and so, uh, so I think there will end up being-- Like, if it's a p, a p times q, and this is-- You know, we're talking about ChatGPT. I, I would say it probably applies to, you know, most of the consumer-facing stuff in the industry. If there's a p times q, which is like price times quantity, which is the really simple way to think about these consumer internet businesses, q is like, at this point they've gotten so big on the, on the monthly active users, like over the course of the next five years, like maybe it gets to two billion or something. I don't know. But like the Google and Facebook properties are in the twos billion. So there's-- It only gets so large. Um, but I think there's a tremendous amount of room to run actually on the upside on the p, uh, on the price. So, so again, if you think about, you know, how are they monetizing today, you know, it's thirty million people out of a billion, um, you know, at a, a modest subscription that is not really reflective of like real price discrimination yet. So I suspect the p is probably like a thing that we get surprised on the upside by. Um, uh-You know, I, I think about, like, lessons learned from previous internet era companies, and I remember looking at internet companies 10 years ago, and we would, we would always look at, like, Facebook and Google, and we were like, "Okay, Facebook and Google, they're monetizing their users at X." Like, that's the max we could get to. Um, and the big story about what's happened over the last 10 years is they've, like, 8X'd their own monetization of their users. And so I, I, I suspect that if they, you know, have some, uh, thoughtful ways of monetizing free usage while, while still maintaining trust, there's probably more upside than downside, uh, on, on pricing, uh, to To- to Thomas's question. And then on the burn, um, the, the actual... Like, most of the burn comes from res- like, research, like R&D, you know, and so fu- future investments. Um, and, you know, I- we could apply this to the whole industry of all the model companies. Um, but we could, you know, we could talk... If, if it's specifically OpenAI, I'm happy to address that one. But o- on the OpenAI side, um, I think they're in an advantaged position because they have the consumer base, and that's stickier. Like, my family, like my parents in Kentucky u- use ChatGPT. Like, if there's some better, slightly better model that comes out, like, they're not gonna switch. Um, and so I think that's pretty durable. Um, and so it's probably better, a better position to be in, um, to fund those research efforts for continued model development. Um, you know, it, there used to be, like, a thing which was, you know, enter- enterprise companies are stickier than consumer companies. Um, in this case, their developer com- like, these are s- like, developers buying these things, uh, on the B2B side for the most part today. Like, buying kinda r- you know, raw access to the models, and that's not very sticky, uh, yet. I think it, it's possible that it does get sticky over time. Um, but as of right now, it's not very sticky. So to the point about coding models, you know, if there's a new coding model that comes along that's better than the latest version from Anthropic, like, our coding companies will just switch, and it's pretty easy to do because it's an API call. And so, um, so, you know, I think interestingly enough, it's a little bit different this time, where the consumer's a little stickier. I think that gives you an advantage. And I think the companies will not irrationally spend on new model development if there's not a financial return. Uh, I would say one of the things that we've observed over the last now almost five years of spending time with these companies is a lot of the founders started as, like, research brain AI people that were like, "We're gonna... There's gonna be no economy, and, like, everything's gonna end 'cause we're gonna have AGI." And it's like, and then you know what, like, has happened? Like, competitive forces have kicked in, and they've become, like, hard capitalists. And so, you know, my, my expectation and from conversations with them, you know, on an ongoing basis is, you know, they're not gonna do totally irrational things on the research side if there's not gonna be a financial payback for them.
- ETErik Torenberg
So can DG also comment on durability of revenue for many of the AI applications built on LLMs, i.e. outside of OpenAI, Anthropic, xAI?
- DGDavid George
Yeah. Um, it depends on the nature of the use. So I think some of them are really sticky. Um, so, you know, companies, you know, like medical scribe stuff, I think is pretty sticky, uh, 'cause there's a bunch of doctor workflow built around it. Um, uh, you know, customer support I think is pretty sticky. Some of, like, the high-end financial analysis type stuff I think is pretty sticky. Um-
- ETErik Torenberg
Can you explain why those spec- specific areas you think are more stickier than others?
- DGDavid George
Yeah. I think the more stuff that gets integrated and the more company-specific, uh, like, kind of rules built around the model stuff you have, the stickier it's gonna be. Um, and so, you know, I think stickiness comes in the form, like in software, in, in applications, like the same way it's kind of always come with software. Uh, I'm sure one of my early stage partners has written a blog post about it, 'cause we talk about it all the time. Uh, but, you know, it's stuff like integrations, uh, you know, rules engines, workflows, um, you know, and stuff like that. Inter- you know, sort of enterprise capabilities. So, like, I think that-
- ETErik Torenberg
You need to have, like, a customer, for example, the rules to go down the sequencing of how to troubleshoot are so embedded that you probably wouldn't experiment a ton once you've got a workflow, you know, across multiple different scenarios that-
- DGDavid George
Yeah. Yeah. Or, and even, like, a style with which something engages. Like, th- these co- like a lot of the companies that are end customers of these things are, like, brands, and they care about-
- ETErik Torenberg
Yeah
- DGDavid George
... the way that, you know, like, the, um, the customer support agent interacts. And so, um, so I think that stuff's probably pretty sticky. I think there's a bunch of stuff that's not sticky at all. Um, so some of the emergent behavior that's like, um, you know... I, I don't wanna talk negative about anything. Um, but, like, some of the stuff that's, like, not as sticky is, like, experimental usage of tools to build, um, you know, some of the internal tooling, software replacements, or, like, very low-end prototyping of websites and things like that. I just think it's, like, TBD, you know, who the players are and what the use cases are, and I think, I think the market for that will probably segment out. Like, we're already seeing it some, where some of the tools are just being kinda used for prototyping, and then some of the cool tools are being used to, like, actually build and deploy apps. Um, and I think you'll see, you know, kinda continued bifurcation of those things. Um, but, you know, it's so early. Um, I don't think that, like-Companies are gonna vibe code up like their salesforce.com.
- ETErik Torenberg
[laughs]
- DGDavid George
It's just not worth it. Like, it's not, like, core competencies. I hope they would. I, I wish we could vibe code away our salesforce.com.
- ETErik Torenberg
Okay, I'm gonna speed run through some of these questions 'cause, uh, I, I recognize we've, uh, opened the flood gates on them, and I wanna make sure we get through as many as possible. So Umberto, um, asked, "Given the speed of go-to-market for AI companies, do you think 100 million ARR is the right milestone to measure against, or are you starting to move the goalpost on, on what success is there?"
- DGDavid George
Yeah, it's so funny. Like, you know, the- it used to be, uh, I can't even remember. One of the, one of the VCs coined the term, you know, like, triple, triple, double, double or something like that.
- ETErik Torenberg
Yeah.
- DGDavid George
And it's like, that looks, like, very modest.
- ETErik Torenberg
Yeah.
- DGDavid George
You know, something like that, yeah. And then that, that time compression has, like, massively gone down. Um, so I, I, I gotta look up the stat. We, we, uh, we ran the analysis on it recently, but it was like, you know, the top companies that we've seen, um, you know, have gotten to 10 million then 100 million, like, four times faster or something like that. Um, for us, you know, it's really important that we have market context and, like, real-time market context to make judgements about what great looks like. And so I don't have an absolute answer for you, but if you were to sit there and look at, a- and, like, and listen to one of our investment discussions, um, we are, we are comparing, like, the growth of, you know, XYZ new app to, like, Cursor, uh, and, you know, and nothing looks like Cursor, so it's unfair, but, like, Cursor, and Decagon, and Abridge, and Eleven Labs. Uh, you know, and not, like, Shopify and DocuSign and, you know, those companies. So, um, you know, if you're not, like, fully in the market and seeing everything and meeting all the companies, you're not gonna be able to have that context to assess what great is at any given time.
- ETErik Torenberg
For sure. Um, for AI companies, what do you think is a good question from Adika. Um, "What do you think about seat-based pricing versus usage-based, and, and what are the experiments that, uh, some of the startups you've seen do well around configuring pricing in this age of AI?"
- DGDavid George
Yeah, this is, like, the, the big question. Uh, I, I love this question 'cause, like, I, I feel like on X and, like, other blogs, people have written all these, like, really long posts about, you know, the, the new prices that are gonna happen and with AI and, like, it's a, uh, to me, it's, like, a little hand-wavy. I, like, one, it's subject to, um, you know, innovations in the technology side and the products getting better. Um, and then two, we've really only seen a true business model innovation, like, early, early days in one area. Um, uh, and that's customer support. It, it, 'cause you can kind of, like, definitively resolve a task. So even, even with that, like, it's hard to do. And so, you know, the, the, the question, maybe to reframe it, would be, like, we had... Actually, we had, like, licenses, right? Like, you know, perpetual licenses with maintenance, and then we switched to SaaS, and that was mostly seat-based, and that was huge innovation, very disruptive. That was a huge enabler of the startups actually going to beat the incumbent software companies, uh, because it was so disruptive at the time. Then we got to usage-based, and so, you know, ob- obviously, all the cloud companies, uh, run that way. Um, you know, Databricks runs that way, Snowflake, et cetera. Um, and then the hope would be y- that I've seen, you know, and, and, and heard about is, like, with AI, can you just monetize the replacement of tasks humans do? Um, and so when I talked about the customer support piece, like, that's the furthest along in the monetizing the tasks that humans do. Um, but that's super, super early. In all the other areas, I think it's really early. Like, end customers want to buy stuff on seat and consumption-based pricing. Um, and, you know, you kinda gotta meet the market where it is. So, um, for now, we're not seeing hugely disruptive things. I think as the capabilities get much better, um, and sort of measurability of the task completions gets more objective perhaps, maybe we get there. But I would say it's, like, super early days, and I'm low conviction that we end up, you know, five years from now with all the software companies monetizing in a completely different way.
- ETErik Torenberg
Absolutely. Um, I've, I-
- DGDavid George
And, and, and by the way, sorry, there's one more thing.
- ETErik Torenberg
Mm-hmm.
- DGDavid George
This also goes to my point that I made earlier about surplus. Like, it's gonna be really hard to capture that surplus unless you can monetize based on the value of the completed task. But because that's gonna be really hard to measure and there's gonna be competition in the market, I, I suspect that a lot of the surplus or savings may end up in the hands of the customer. Again, you could probably still build really interesting, good companies that have really high market cap, um, but that is, like, you know, sort of directly related to who captures, uh, the value and the surplus. Like, like, uh, like the, uh, I always say, like, the steam engine got invented, and, like, you didn't price the... Like, the steam engine didn't get priced based on, like, the calculation of how many, you know, humans it replaced in doing a specific task. Uh, competitive forces kicked in and, you know, it was priced at some, like, appropriate, you know, competitive level with a return on capital, um, while still, like, capturing a lot of value but delivering much more value to the end customers or the users of, of the steam engine. And I think the same will happen here.
- ETErik Torenberg
For sure. Actually, it's a good, um, transition for us to actually, uh, talk about what this all means for growth, um, given this setup.
- DGDavid George
I, I, I try and sit in on as many pitches in our early stage-... practice as I can, just because I feel like I learn a lot about, you know, what's starting, w-w-what are interesting founders starting to hone in on and, you know, and what areas are they building. Um, and our early stage team is doing a great job of, uh, you know, doing the most exciting early stage deals. So that to me just bodes really well for the next 12 to 24 months of, of deal activity for us, um, you know, looking out.
- 43:15 – 51:30
Companies Staying Private Longer
- DGDavid George
All right. Growth partner. So [clears throat] this is a, this is a crazy looking chart. I mean, it's not surprising, 'cause it, it was, like, the first thing we talked about, which is, you know, tech markets are big, and of course, companies are staying private, uh, longer. But, um, but you can see it in the chart here. It used to be that companies, uh, would go public, you know, within a, call it, like, five to 10-year window of, of inception. And despite the fact that the companies are growing much faster and getting better quicker, they're staying private way longer. So, um, you know, all, all is that, all that is to say, you know, it's like 14 years, and I think that's getting, getting even longer. If you take the market cap of, you know, the, um, the private markets valued above a billion dollars, and, you know, we could argue are some of them overvalued, undervalued. Um, but, um, that whole value in aggregate is, like, three and a half trillion dollars, and that's, like, 11%. Three and a half trillion is like 11, 12% of the NASDAQ, or 10% of the NASDAQ, something like that. Um, if you go back 10 years ago, that whole private market cap of three and a half trillion was like 500 billion. So over the last 10 years, the market cap of these private companies has, like, 7X'd. So there's huge growth, like, again, that is related to this point, which is some of the best companies taking longer, deciding to stay private longer. Um, but, you know, it's, it's pretty, it's pretty stark. Um, so the other thing that's going on, and I mentioned this earlier with Snowflake, is, like, the public markets are no longer the place of extremely high growth. Like, it, because of this, and it's sort of logic, it, it's logical, it follows, right? Um, but something like 5% of software and internet public companies are forecasting 25%-plus next 12 months growth. So 95% of the public market universe in software and internet is growing less than 25%. So if you want access to, like, the, and, and this is, it's a little bit too bad, because, you know, there's obviously implications for retail and stuff like that. Um, [clears throat] but the reality of what's happened is the high growth segment of new technology companies is all living in the, in the private markets now, um, you know, for the most part, and so I don't see that changing. Um, I do think, you know, Figma and, and a bunch of good companies, like, I think a bunch of g- of good companies will go public. Um, but this is a trend that, you know, is, is, is pretty long-standing, and I don't see it reversing, uh, anytime soon. So, um, we have a bunch of really good AI companies. Some people have asked, they're like, "Oh, is- isn't there, like, crowding that's happened in some of the best, you know, companies?" And, um, what I would say is, like, one, we're getting in a lot of these companies much earlier than o- than others are able to, and two, even for the ones that are later, where, you know, there are bigger rounds, like, it's really important that we have that early relationship because it gives us ball control and access into the rounds, and X AI, uh, first outside money and beyond Elon. So not only, you know, and, and a lot of these are growth fund, kind of, first money in. Um, so, you know, it's important to get in earlier to generate returns, but also to position us, you know, with management and to have access, uh, to shape later rounds. If I were just to spend a second on, you know, how we have, like, our, shaped our AI strategy, there's basically two parts to what we're doing. Um, one is the companies where they're, like, flying off the page, you know, uh, like undeniable momentum. Uh, so, you know, co- companies like Cursor and Decagon and Eleven and Abridge and, and, you know, others. Um, the second bucket is, um, very early deals in the very, very, very best teams in the market. And I say very, very best teams in the market, like, with emphasis, um, because I think we're talking about, like, the top five teams in the world, and anything beyond that we're not doing. These companies, they're, they're sort of a different shape. Um, they're growth dollars earlier than normal, and there's a higher degree of variance in the business outcomes. But because the teams are so special and so great, we feel like business risk looks very different from capital risk. And so w- w- we feel like there's, it's kind of, they're kind of asymmetric bets, where the asymmetry lies in the fact that we're probably downside protected, um, because the team quality is so high, and there's so much, you know, demand, uh, for talent. Like, you know, we would be downside protected even if it doesn't work out. And so high business, you know, sort of variance, but, you know, sort of asymmetric capital or returns profile that looks a little bit different, uh, than a typical early stage investment.
- ETErik Torenberg
Talk about timelines to, um, exits, because on one dimension, you know, folks may, may take the reaction of, gosh, you know, private for longer means just extension of hold periods for LSB companies. Maybe the reframe on it is why is that great for, for us, but frankly as private investors, and, and why do we think it's actually advantageous in, um, some respects for some companies to actually delay time to IPO, and how are you counseling folks as a part of that?
- DGDavid George
Yeah, I mean, this is a, it's a tri- it's a tricky issue. I mean, I said it right up front, which is, like, I, we get paid on, on DPI too. Like, uh, [chuckles] you know, that's, that's, that's how we, that's how we compensate ourselves and our teams. And so, you know, we care, we care a lot about it. I would say we're in a fortunate position, uh-Where our portfolio is, is pretty good. Um, and we've had a number of companies that have acc- have exited to the public markets or sold themselves. Um, and so, you know, we're not in a position where we haven't been able to deliver some liquidity. Um, for those sorta champion companies that have been private for a long time and, and are getting a lot of coverage, uh, I think each kind of has idiosyncratic reasons why they've wanted to stay private and, you know, I think, I'm confident that they will go public. So, um, so I think most of them will probably end up going public, it just will take a longer amount of time. Um, you know, one of the big, uh, things that we've observed is the private markets have adapted to some of what you get from the public markets. Um, and, you know, we've fortunately been very active in, in some of those situations. So things like, um, you know, ten- tender offers in the private markets. A big thing that is hard for private companies is competing for talent. Um, and, you know, part of the reason why it's so hard is because public companies grant RSUs that typically vest on a quarterly basis, and so they hit your account, they're already tax withheld, and, and it's basically like getting paid a lot of cash. And it's hard for private market companies, in some cases, to compete with that. Um, so, you know, some of the bigger, better ones have done more regular tenders, and we've been very active in shaping those, uh, with the companies. So to me, it's a balance. We kinda want what's best for the company. If there's strategic reasons why it benefits them to be private, that's great, and we'll help enable that. Um, but we do recognize that obviously, you know, our goal is to, to monetize, uh, great investments. We're not there yet where we feel like there's a need to do unnatural things. Uh, I, I just don't see that being the case.
- ETErik Torenberg
For sure. And I, I do wanna call out as well, um, I'm gonna brag on behalf of you, David, because I, I do think, uh, DPI generally is the issue and challenge across the industry. I don't think it is our issue necessarily. Um, okay, I'm gonna also cover, uh, a question here. Uh, thank you for letting me, uh, do our commercial, David, on DPI. You know it's my favorite topic. Uh, on the topic of, of, um, when you think about some of the names that we have underwritten, you know, I think there is this broader question of, gosh, like, there's a lot of, of names, for example, like, um, you know, the chat, uh, OpenAI's of the world and also Databricks, that feel like, you know, they're doing more of these tender offers, they're doing more of these, you know, kind of SPVs. How do you counter that portfolio
- 51:30 – 57:18
Portfolio Composition and Construction
- ETErik Torenberg
composition of names like that versus names like an Anduril or Flock, where it's probably very, very difficult to get any access, and how do you think about the configuration of the portfolio as a part of that?
- DGDavid George
Yeah. I mean, look, um, I mean, the OpenAI investment that we made, um, you know, we, uh, we were able to write exactly the check that we wanted to write, um, you know, in terms of sort of portfolio construction. So to me, it doesn't seem... You know, I'm sure, I'm sure some folks have had access because other managers are, are trying to do SPVs and stuff, but I think that's increasingly hard. Um, uh, you know, it's also sort of our decision on how much to weight these opportunities and build a portfolio. Um, so, you know, I happen to think that the return profile of, of that recent one in Databricks is, is very attractive. Um, but, you know, we don't want an entire portfolio of things that we think are, like, pretty safe, two Xs, you know, where I think the question is, can we make five X on them? But, like, we feel pretty confident we can make three to four X, and two X is pretty safe.
- ETErik Torenberg
Awesome. Speaking of, um, public companies, we had a question from, from Paige here on, um... well, actually, it's, it's two related questions. Um, what publicly traded software companies do you think won't be disrupted by A- AI and have real moats, and any obvious ones that you think it will be? And related to that, I think we had some questions earlier on around, uh, how much, if at all, any public companies you'd do in the fund.
- DGDavid George
Yeah. Ver- probably very few public companies in the fund. I mean, we'd have to have a really, really strong thesis and relationship with the management team to wanna do that. Um, again, the, the public universe is just way slower growing, and there's really, really exciting stuff that we can do in the private universe. And so the bar is extraordinarily high, um, you know, for things in the, in the public markets. And we've done a couple which have worked out, I would say. Um, but, you know, the bar is very high. And the, the opportunity cost is probably something that may be even higher growth, uh, in, in the private markets. Um, yeah, the, what, what companies are gonna get disrupted in the public markets, like, I love this question. I don't know. All, all I, I would offer this framework, which is, which is, m- maybe there's three interesting ingredients to consider when you think about the safety or durability of the, of the public software companies. One would be, you know, UI/UX. So if to the extent that we get a complete reimagination of UI/UX, um, you know, that'll be really exciting. Like, what is salesforce.com? Salesforce.com is, like, a set of really uninteresting, um, you know, checklists and forms that people fill out, for the most part, on the front end. Um, the promise of this stuff, you know, agents is, like, a way overused term, but, like, like, the new technology should be able to, like, do things for you, as opposed to, like, keep your records for you. And so you could imagine a completely reimagined UI/UX which is proactive, which is like, "Hey, instead of you going and inputting things, I'm, I know what you're doing. I'm just gonna tell you what to do." Um, and so a total reimagination of the workflow, um, would be one ingredient for a startup to win versus an incumbent. The second would be access to data. Um, so, you know, an entirely new form of data that gets sucked in to take the actions on your behalf. Like Salesforce, you know, I mentioned the form thing on the front end. The reason it's really, really sticky is the database on the back end. So to the extent that instead of using that-Sort of structured database on the back end. You take all of your unstructured data and dump it into, uh, Databricks, for example, um, you know, and, and query it or access it through that, um, that would be another interesting sort of opportunity for a startup to attack. Um, and then the third is a business model innovation, and I mentioned this already, it's super, super early days, but to the extent that startups can come up with an, you know, sort of like novel way of shifting the business model in a disruptive form, uh, against, like, seat-based pricing salesforce.com, uh, you know, it would be, uh, it, it would have a chance to win. So, I think for startups to win, I think you need all three, um, and that's just in the head-to-head stuff. I think startups are already finding interesting windows of opportunity in and around these systems of record, um, which we've made a bunch of investments in. We haven't found the startup that's, like, got the killer idea for, like, dethroning salesforce.com yet. I hope we find it. So I know that's an unsatisfactory answer, but at least it's like a, a fr- that's the framework I would use to, to think about, those, those three buckets.
- ETErik Torenberg
Awesome. Okay, I'm getting the hook because I know we only have a few minutes left of this webinar, so maybe in the last couple minutes we'll, uh, fast-forward here to, uh, to talk about the team. I don't know why it took us over, uh, an hour plus to get to this point, so maybe we'll go through that and then, David, also how you work with the broader a16z team, and that'll bring us all back home with how this all fits in in the broader, uh, franchise and firm as well.
- DGDavid George
Yes. I love the team. Uh, they're, uh, they're very, very smart. Uh, they've come together in a, in an awesome way. We have a really strong sort of team subculture. Um, you know, we are very, very, very intertwined with the early stage teams. Um, you know, if you think about, like, where we get alpha in our business, um, you know, there's access, which I've talked about already, um, you know, in terms of, like, you know, being already involved in the companies. And, and 80% of the time when we make a new investment that's not an early stage investment, one of our early stage folks has some pre-existing relationship. So, um, you know, access is a big piece of it. Insights is the other piece. Like, I think the way you get
- 57:18 – 1:03:26
Team Culture and Collaboration
- DGDavid George
really, you know, like, outsized, upsized, you know, returns in our business is market and product insights. I mean, I think you, you have to do all the analysis around business model and financials, 'cause you can make big mistakes if you get that wrong, and we go super deep in each of those. Um, but I think the real outlier opportunities come when you have a market or product insight that maybe the rest of the market hasn't figured out yet. And, you know, the fact that we're attached to the early stage team gives us the opportunity to have those and gives us a tremendous amount of leverage to have a relatively small amount of team, uh, s- small amount of team members given how much, how much ground we cover.
- ETErik Torenberg
I will also, uh, take the opportunity to, uh, brag on your team and my team, 'cause I think both of our teams got a 91% on the employee engagement score. So, not that we're competing across teams, but, uh, you know, this is a good barometer for team subculture, uh, no doubt, and, uh, something we always, we do around the firm, uh, year round, sh- which is very company-like, but hopefully gives you a sense of what we're measuring, uh, you know, ourselves across as well in terms of, of team culture. Um, but I wanna spend a, a minute to illustrate the point that David alluded to around how we work in collaboration with the early stage, and also, um, think about the late stage venture in terms of sitting across all the six different early stage buckets, uh, across the, across the firm. And so, maybe, David, if you wanna give just, like, the high level talk track on if you were to hypothesize what the composition of the portfolio would look like across these six different buckets, that'd be helpful. And then I saw a question around how much we would expect around crypto investments, and particularly around token crypto investments, um, that would be helpful.
- DGDavid George
Yeah, absolutely. The best part about where we are is we've got people at the early stage are figuring out where they think they should be spending their time, and typically, you know, our world is, like, 12 to 24 months downstream of that. So, when you see the way we size our funds at the early stage, that's a reflection of what those early stage teams think is their opportunity set and, you know, ours sort of follows that. So, I think the largest amount of opportunity will continue to be AI infra, AI apps. Next, American dynamism, you know, our early stage team is killing it in that area. You know, beyond just AI, there's obviously a pressing, uh, you know, sort of market and world need for American dynamism companies. So, you had, at the same time, you have market need, you had people that figured out how to do this inside of SpaceX and, a- and, uh, Palantir, who have then left to start companies, so the talent is there to navigate, um, you know, a, a very complex, uh, go-to-market, uh, motion. Um, and then you have advances in technology beyond gen AI, like, uh, autonomy capabilities and vision, uh, advances that enable American dynamism. So we're very excited about that area. We're starting to see a little bit more interesting stuff in AI-enabled health things, so we've done a couple. Um, you'll continue to see us active there, but it's probably a little bit less than, uh, the AI infra and app side. And then crypto, um, the way we're doing that is working, um, you know, hand-in-hand with Chris and the crypto team, uh, for their high conviction bets, where they're at a stage that they fit, you know, the growth fund and, um, and, you know, I, I, I'll brag on them, I think they're the best crypto investors in the world, and, and we sort of are happy to attach ourselves to that. And it certainly depends on the opportunity set. Like, we're seeing really exciting stuff in the enablement of stablecoins. Um, and, and so, you know, to the extent that we see, like, a massive takeoff in that market, it could be more. I always say, it would be great if 100% of our investment activity in the growth fund was follow-ons, uh, 'cause that would mean that our early stage team is absolutely killing it and, and they're getting a lot of market share. Um, but, you know, I think there'll always be a place for us to do really attractive new things. Um, and we don't-Build the portfolio based on some target around new verse follow on. It, it's, it's really, like, best ideas win, you know, where do we have access, right to win, et cetera.
- ETErik Torenberg
Great. Um, and then, uh, less than 60 seconds. Do you wanna take Annie's question? So, on the topic of portfolio construction, how much exposure do you want to top research teams with a wide fan of outcomes, versus opportunities with a more narrow range? How many more researchers do you think are actually out there, um, you know, beyond the ones that you've already backed, that, that you will get more exposure to and-
- DGDavid George
There are extremely high-end researchers at some of the big labs that we ha- we, our early stage team is tracking. And I should have mentioned this, when we do these, we do them with the AI infra fund. So we would not, you know, we, we rely on them heavily, uh, to make sure that we're doing the right assessment of the teams, and, uh, you know, the research ideas that they're pursuing. Um, you know, in terms of portfolio construction, I, I, I quite like the way that we've done it, um, where, you know, we have some, you know, absolute champion companies that we, we think actually have, like, a lot of room still to run. Um, where on the downside we think we would, if things really go poorly, like, we'd still probably make two times our money. And on the upside, you know, there may not be the opportunity for 10X over five years, but we think there will be the opportunity for 5X over five years. The research team with the high variance is a little bit of an output of that great opportunity set. Like, right now, there's not another Ilya floating around in the AI market, and so we would never try and say, "Oh, we want 10% of our portfolio to be in these, so we need to find someone." It's more reactive when we do find those people who are really special. I think the only shift that we'll see over time is, I think you're seeing a bunch of really, really, really exciting AI apps and American dynamism companies at the earlier stage that are poised to become kinda champion companies over the next five years. Um, and so you'll probably see a lot of activity from thus, uh, from us, uh, in those kinds of companies.
- ETErik Torenberg
Awesome. And with that, I'm gonna close out here. Lastly, thank you, David. Appreciate it.
- DGDavid George
Great to hang, as always.
Episode duration: 1:03:40
Install uListen for AI-powered chat & search across the full episode — Get Full Transcript
Transcript of episode 6HxT4nQmtvc
Get more out of YouTube videos.
High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.
Add to Chrome