The AI Frontier and How to Spot Billion-Dollar Companies Before Everyone Else — Elad Gil

There are moments in time where it's very smart to be contrarian. And there are moments in time where being consensus is the smartest possible thing you can do. And I think right now we're in a moment in time where being consensus is very right. You know, you can really overthink it. And what's a contrarian thing? We should go do a bunch of hardware stuff cuz blah blah blah. You like maybe buy more AI. You know what I mean? I think people make these things way too complicated.

Yeah, as always.

Oh yeah, we were just talking about some of the acquisitions that are happening in the AI world. We saw that xAI just got an option to effectively purchase Cursor. It looks like obviously Scale was sort of partially taken by Meta. There have been a variety of these sort of deals that have been happening over the last year or two. And separate from that, we're just talking about what does that mean for the AI research community and the AI community in general. And I think the most interesting or one of the interesting things that's happened over the last year or so is Meta really started aggressively bidding on AI talent, which was a very rational strategy, right? They're going to spend dollars on compute. So, it made sense to have a real budget to go after people. And normally what happens in tech is a single company will go public and a bunch of people from that company will be enriched and then a subset of them will continue to be heads down and working really hard and focused on their original mission and a subset of people start to get distracted. They may go and work on passion projects for society. They may get involved with politics. They may go start a company. They may just kind of check out and hang out or go to the beach kind of thing. And what happened recently is because of the Meta offers and then all the other major tech companies having to match offers for their best researchers, somewhere between 50 and a few hundred people effectively had an IPO but as a class of people. It wasn't like they were at one company. They were spread across Silicon Valley but all of their pay packages suddenly went up dramatically and they experienced the equivalent of an IPO. And that's really unusual. It's kind of the personal IPO. And the only time in history I can think of where I've seen it happen before is in crypto where a bunch of the really early crypto holders or founders suddenly as a class all went effectively public in 2017ish. And then again more recently. But this is really interesting. It's kind of under discussed. It may not have huge long-term implications, but it does mean a subset of people will change what they're focused on. Try and do big science projects to help humanity, work on AI for science. Maybe some people will go off and do personal quests or things like that.

Definitely not. In that case, you look around say Austin, you've got the Dellionaires, which refers to Dell post-IPO early employees and so on. But as a class of people, when that happens, I suppose we don't know how large or how long-term the implications are, but there seem to be implications.

Very dramatic.

I don't have exact knowledge of the full range and everything else. The rumors and the things that have kind of made it into the press, the claims are that these things are between tens of millions and hundreds of millions of dollars per person. And again, it's a very small number of people who would get anything that's quite that upsized. But I think the basic idea is we're in one of the most important technology races of all times. And the faster that we get to sort of better and better AI, the more economic value will effectively show up. And therefore people were really willing to pay in an outside way for the handful of people who are the world's best at this thing. And 5, 10 years ago these people were like well compensated but it was a completely different ballgame. It just wasn't the core of everything that's happening in technology but also honestly society and politically and for education and health, like it's going to have all these really broad and I think largely positive implications for the world. But it is the moment of transformation and so suddenly these pay packages are going way up.

So basically all the different people, call them labs now. That's OpenAI, that's Anthropic, that's Google, that's xAI, etc. All the labs are basically training these giant models and effectively what you do is you buy a bunch of chips from Nvidia and you're actually building out a system. So you have chips from Nvidia, you have memory from SK Hynix and Samsung and other places and you're building a data center. There's all these things that go into building these big systems and data centers and everything else. And you basically have clusters of hundreds of thousands or millions, or the scale keeps going up, of systems that you're buying from Nvidia and from others. Google has their TPU. There's other systems as well. And you're using that to basically train an AI model. And what that means is you're running huge amounts of data against these big clouds. And eventually the crazy thing is your output or your model is literally like a flat file. It's like almost like outputting a text doc or something. And that text back is what you then load to run AI, which is insane if you think about it. You use a giant cloud for months and months and months and your output is like a small file. And that small file is a mix of representing all of humanity's knowledge that's available on the internet plus logic and reasoning and other things built into it. And you can kind of think about that in the context of your brain, right? You have three or four billion base pairs of DNA and that's more than enough to specify everything about your physical being but also your brain and your mind and how it works and how you can see things and talk and taste things and all your senses and everything's just encapsulated in these very small number of genes actually. And so similarly you can encapsulate all of human knowledge into like the flat file effectively.

You need lots and lots of chips from Nvidia to do this or TPUs or others but then you also need other things. You need packaging to actually be able to package the chips and so there's a whole supply chain around building out these systems and different parts of that supply chain have constraints of them at different times. And so right now the major constraint is memory or a specific type of memory that's largely made by Korean companies although there's some broader providers of it and people think that that memory constraint will exist for about two years maybe plus or minus because ultimately the capacity of those companies has been lower than the capacity for everything else in the system. People think other constraints in the future may literally be building out the data centers or power and energy to run these things, right? But for today, it's this memory. And so everybody in the industry is constrained in terms of how much compute they can buy to throw at these things. And so what that does is it creates a ceiling on top of how big you can scale these models up in the short run because every lab is buying as much as it can. A bunch of startups are buying as much of this computer as they can and everybody's constrained. What that means though is you have an artificial ceiling on how big a model can get in the short run and how much inference can run or how many things you can actually do with AI right now. And that also means that you're effectively enforcing a situation where no one lab can pull so far ahead of everybody else because they can't buy 10 times as much compute as everybody else.

And so that may mean that over the next two years-ish all these labs should be roughly close to each other because nobody has the capacity to pull ahead. And when the constraint comes off there is some world where you could make an argument that suddenly somebody can pull far ahead of everybody else. So right now OpenAI, Anthropic, Google, you know, they're reasonably close in terms of capabilities although some will pull ahead on one thing versus another that should roughly continue everybody thinks for the next at least two years because of this. Google is also constrained by the memory from Samsung, Micron, etc. They're similarly constrained as the other players. Right now, everybody is similarly constrained and a subset of these labs either are already making their own chips or systems like Google has TPUs and other things. Amazon has actually built its own chips called Trainiums. And so there's basically different systems for different companies, but fundamentally all of them are limited in terms of how much they can either manufacture themselves, purchase themselves. And a year or two ago, the main constraint was packaging. Now it's memory. Two years from now, who knows? Maybe it's something else. We constantly are hitting bottlenecks as we're trying to do this build out.

Part of that is because the way that some of these things are built and it's basically the capacity that you need for example for memory is basically a type of fab and so you need time to build out the fab and to get the equipment and put the lines in place. So it's a traditional sort of capex and infrastructure cycle. And these companies basically underinvested in that because they didn't quite believe the demand forecasts that other people had around this stuff. And so now they're trying to catch up. And so it's one of these things where everybody keeps saying, 'Well, AI is growing so fast. How can it possibly keep growing at this rate?' But it keeps growing at this rate. It just keeps going. And that's because its capabilities are so impactful and so important. And so you look at the revenue of these companies. And it's interesting. I can send you the chart later, but Jared on my team pulled together a graph of how long did it take for companies to get to a billion dollars in revenue and then from a billion to 10 billion and then from 10 to like a hundred. And there's only a small number of companies that have ever done that. And you can literally look by generation of company how long it took. And so for example, I can't remember, it's ADP or somebody, it took them 30 years to get to a billion in revenue or whatever it is. Anthropic, OpenAI did that in like a year. For Google it took four years or whatever. I don't remember exactly what the numbers are, but it was kind of like as you go through these subsequent generations, it gets faster and faster to get to scale. Right now, OpenAI and Anthropic are each rumored to be roughly around $30 billion run rate, which is insane. That's crazy.

It's bananas. Yeah, it's bananas. And it's really actually important and useful. That doesn't include like the cloud revenue for Azure for doing AI stuff or Google GCP or like, it's just those two companies. It's insane.

In terms of the precedent, and that doesn't mean it's going to happen here, but if you look at every technology cycle, 90, 95, 99% of the companies in that cycle go bust. And that dates way back even to what was high-tech 100 years ago which was the automotive industry in Detroit, dozens of car companies and hundreds of suppliers and it collapsed into a small number of auto companies virtually. And so this is not a new story. During the internet cycle or bubble of the '90s, 450 companies went public in '99. 450 or so companies went public in the first few months of 2000. And so that was 900 companies. And say another 500,000 went public in the couple years before that. So you had somewhere between 1,500 and 2,000 companies go public. So that means they kind of made it. And of those, how many have survived? A dozen, maybe two dozen. And so out of 2,000 companies, 1,980 or so went under in one form or another. Or maybe they got bought for a little bit. And so there's no reason to think the AI cycle will be any different. And every cycle is like that. SaaS was like that and mobile was like that and crypto was like that. So most companies are not going to make it. A handful will. And we can talk about those. And so if you're running an AI company right now, you should ask yourself, what is the nature of the durability of your company? And are you one of that dozen or two that are going to be really important 10 years from now? Or is now a good moment for you to sell because what you're doing will start to get commoditized or will be competed by a lab or will be something that the market will shift or the technology will shift and you'll become obsolete. And there's a handful of companies that will continue to be great. They should never sell. They should never exit. They should keep going. But there's probably a lot of companies that now or the next 12 to 18 months is the best moment for them possible in terms of the value that they'll get for what they're doing. And for every company, there's a value maximizing moment where they hit their peak. And it's usually a window. There usually, you know, 6, 12 months where what you're doing is important enough, you're scaling enough, everything's working before some headwind hits you. And sometimes it's very predictable that that headwind is coming and you can see it. And often you see it in the second derivative of growth, like how fast you're growing starts to plateau a little bit and you're either going to keep going up or you should sell. And so that's really what that's meant to be. I'm incredibly bullish around AI as you can tell from the rest of the conversation. And so it's less about the transformation that's happening overall because of this technology and more that only a handful of companies are going to continue to be really important. And so are you one of them or not? If you're one of them, you should never ever ever sell.

I think the core labs will be around for a while. So that's OpenAI, Anthropic, Google, barring some accident or disaster, some blow up, but it seems like they're in a really durable spot. And to your point on like market structure, I wrote a Substack post, I don't know, three years ago or something predicting that that would probably be an oligopoly market and there'd be a handful and be aligned with the cloud. That's roughly kind of what happened. I mean, there's Meta and there's xAI and there's other players that may change this. It didn't exist when I wrote that post, but it feels to me like in the short run that's an oligopoly. Like there's no reason for that to be a monopoly market unless one of them pulls ahead so much in capabilities that it just becomes the default for everyone. And that could happen, but so far it hasn't. And again, this compute constraint may prevent that in the short run or at least provide an asymptote on it. As you move up the stack and you see, well, there's different application companies. You know, there's Harvey for legal, there's Abridge for health, there's Decagon and Sierra for customer success. You know, there's these different companies per application. There's three or four lenses that you can look at. One is if the underlying model gets better, does your product or service get dramatically better for your customers in a way that they still want to keep using you? Second, how deep and broad are you going from a product perspective? Are you building out multiple products? Are they all integrated in a cohesive whole? Is it really being built directly into the processes in a company in a way that it's hard to pull out? Often the issue for companies in adoption of AI isn't how good is the AI, it's how much do I have to change the workflows and the ways that my people do things in order to adopt it. It's about change management usually. It's not about technology. And so if you've been able to embed yourself enough into workflows and how people do business and how they work and how everything else kind of ties together, that tends to be quite durable. Are you capturing and storing and using proprietary data? Sometimes it's useful. I think data moats in general are overstated, but I think sometimes it can be actually quite useful and that's usually the system of record view of the world. So, you know, there's a handful of criteria around like will this thing be long-term defensible or not and the application level that's often one potential lens on it.

I think there's a lot of exit options. And the thing that's crazy right now is if you go back 10 or 15 years, the biggest market cap in the world was like 300 billion. The biggest tech market cap was, I don't know, 200ish or something. I think the biggest one at the time was Exxon or somebody like 15 years ago. And over the last 10 or 15 years, what happens is we suddenly ended up with these multi-trillion dollar market caps, which everybody thought was nuts at the time, but things will probably only get bigger. There'll probably be more aggregation versus less into the biggest winners. And there's more and more companies who have these market caps between say 100 billion and a few trillion in a way that's just unprecedented. And that means there's enormous buying power because 1% of 3 trillion is 30 billion, right? You can get 1% and pay $30 billion for something which is insane, right? That's pretty unprecedented and that means that these really big acquisitions can happen.

Depends on what it is. And it may be a lab. It may be one of the big tech incumbents, Apple, Amazon, right? Google's kind of both things. There's Oracle, there's Samsung, there's Tesla, there's SpaceX now in the market doing things. There's a bunch of different buyers of different types. There's Snowflake and Databricks. There's Stripe. Coinbase if you're doing financial services, there's just a ton of companies that actually are quite large, that's kind of the point. And so often you end up selling to one of four things. You can sell to one of the big labs or hyperscalers or giant tech companies. You can sell to somebody who cares a lot about your vertical, so for example a Thomson Reuters if you're doing legal or accounting or things that are kind of related to that. I mean I think actually one thing that doesn't happen enough is merger of competitors, particularly private companies where you can do that because ultimately if your primary vector is winning and you're neck and neck with somebody and you're competing on every deal and you're destroying pricing for each other. Like maybe it's better to just merge. It actually was X.com and PayPal in the '90s, right? Elon Musk were running different companies and they merged because they said we're people doing this. Why fight?

Yeah. And the rumor is that that almost happened and then the Uber side walked away from it. But all the money that Uber spent on fighting Lyft for all those years maybe would have been better spent just buying them. Maybe not. I don't know the exact math. But often it actually does make sense to say, you know what, like we'll just stop fighting it out and we'll just combine and just go win. Cuz if the primary purpose is to win the market, you're already fighting all these big incumbents that already exist anyhow. So why make it even harder?

The math stuff helped me I think in two ways. One is it's helped me with certain aspects of like technical or algorithmic CS and understanding. And sometimes that's useful in the context of how certain things work in AI or things like that or just fluency of numbers and data and I call it nerd language or something. And I did the math degree honestly just for fun. And I think that's actually the thing that was helpful. We did an undergrad degree in math so I didn't go that far with it. I did the very sort of abstract pure math stuff and I think that was a good forcing function of how to really think logically step by step about things because roughly the way that at least I learned how to do proofs was you do the logical sequence but then sometimes you do these intuitive leaps and then go back and try and prove it to yourself or flesh out the reasoning behind that intuitive leap. And I think sometimes investing is a little bit like that.

Not really. I'm really hard on myself, so even now I second guess myself a lot. Somebody was telling me that the two people that always beat themselves up the most in hindsight is me and this one other person who's another well-known founder/investor. And so I don't think there's a single moment where I'm like, 'Wow, this makes sense for me to do.' I think it just kind of organically kept going because I was getting into some very strong companies and then, you know, that allowed me to sort of continue what I'm doing. But yeah. Wish I hadn't done it like that.

Yeah. I mean, ever since I was seven, I've been thinking about investing in technology.

I think there's what happened early and what happens now. And I think those two things are different. I think to your point, the single most important thing for anybody wanting to break into any industry is go to the headquarters or cluster of that industry. Like move to wherever that thing is. And all the advice of you can do anything from anywhere and everything's remote is all BS. And you see that for every industry, not just tech. You know, if you wanted to get into the movie business, people wouldn't say, 'Hey, you can write a film script from anywhere. You can digitally score from anywhere. You can edit it from anywhere. You can film it anywhere.' They'd say, 'Go to Hollywood.' And if you want to do something in finance and you're like, 'Well, you could raise money from anywhere and come up with trading strategies and a hedge fund strategy from anywhere and you could do it from anywhere.' People wouldn't say, 'Hey, go to Seattle.' They'd be like, 'Go to New York or go to XYZ financial center.' So, the same is true for tech. And Shan on my team has been performing this sort of unicorn analysis of where is all the private market cap aggregating for technology. And traditionally about half of it's been the US and then half of that has been the Bay Area. But with AI, 91% of private technology market cap is the Bay Area. 91% of the entire global set of AI market cap is all in one 10 by 10 area. Right? So if you want to do stuff in AI, you should probably be in the Bay Area. Probably the secondary place is New York and then after that it just drops off a cliff. And really it's the Bay Area. If you want to do defense tech, you probably should be in Southern California close to where SpaceX and Anduril are and sort of Irvine, Orange County, etc. or El Segundo. There's a lot of startups there. If you want to do fintech and crypto, maybe it's New York. But the reality is these are very strong clusters. So to your point, number one is I was just in the right location. I was in the right networks and I default was I was running a startup myself. I was at Google for many years and then I left to start a company and people just started coming to me for advice and the way I ended up investing in Airbnb is I was helping them when they were eight people or something raise their Series A and I introduced them to a bunch of people and help with some of the strategy there in very light ways, right, they would have done it without me, but and they said hey at the end of it do you want to invest a little bit? I said great that sounds wonderful. So it's very organic. Or the way I invested in Stripe is I'd sold sort of infrastructure early API company to Twitter and when Twitter was say 90 people or so, and I sent an email to Patrick, the CEO of Stripe, just saying, 'Hey, I've heard great things about you and I really like what Stripe is doing and I want to use it for my own startup.' And I sold this API company myself. Do you want to just talk about this stuff? And so I went on a couple walks and then a week or two later, he texted me and he's like, 'Hey, we're doing a round. Do you want to invest?' So the first few things that I did were very organic where the founders were like, 'Want you on board.' I didn't think, oh, I should be an investor and I'm going to chase things and I just really liked talking to smart people and I liked working on certain business problems and I love technology and his translation, and so it was very like, you know, I was just a nerd and I met other nerds and we hit it off. It's kind of the early like story for me.

Oh. Yeah, that's a good point.

It's a while ago now. It's probably like seven-ish years ago. Something like that.

90% yes. Every once in a while you meet somebody exceptional and you just back them or something maybe so early like when I led the first round of Perplexity, like the very very first round and the way that came about was Arvind the CEO just I think he like pinged me on LinkedIn literally and this was when nobody was doing anything in AI and he was like an OpenAI engineer or researcher and he's like hey I'm at OpenAI which nobody cares about at the time and I'm thinking of doing something in AI and I heard that you're talking about this stuff and nobody else is talking about it and can we meet up? And so we just started meeting every two weeks and brainstorming, right? And then that led to like investing in that. And that was kind of a people first thing where he was just so good and every time we talked, he'd show up a week later with a thing that we discussed built. Like who does that?

So good. Or the way I ended up investing in Anduril was Google shuts down Maven, which was their sort of defense project. And so I think well if the incumbents aren't going to do it what a great place for startups to play because there's been a long history of Silicon Valley and the defense industry, that's HP and that's a lot of the early brands and so I was just looking for something there or somebody to work on this area and it was very unpopular at the time and I ran into I think it was Trae Stevens who's one of the co-founders of Anduril who's also at Founders Fund at lunch or something else again, right city to be in, and he said oh I'm working on this new defense thing and I said amazing let's talk about it. Sometimes it's just looking for these things too in a market and sometimes it's people. So Anduril was looking for a market and then finding amazing people. Perplexity was kind of in between where it was like I was looking at everything in AI cuz I thought it was going to be incredibly important but not very many people were. And then I just ran across an exceptional individual and that's when I funded OpenAI. That's when I funded Harvey which is the early legal. I funded a lot of really early stuff because they were the only people doing anything in this market that I thought would be really important.

I mean, I think he pinged me in part because I was involved with a bunch of the prior wave of technology companies. Airbnb, Stripe, Coinbase, Instacart, Square, a bunch of stuff like that. And so I think at that point I was already known as founder and investor. But then on top of that I was just trolling AI researchers and just asking them about what's going on because it was so interesting. There's a bunch of art that was being done with these things called GANs at the time, these generative adversarial networks. And so I was playing around with that. I tried to hire engineers to build me effectively wasn't Midjourney because I just thought it'd be really cool to make it easy to make AI art.

I think there was two or three things. AI was one of those things that people have always talked about. So, when I was doing my math degree, I took a lot of kind of theoretical CS classes. There were the early neural

Network classes and things like that, and the math behind it. And so there's always this promise of building these artificial intelligences of different forms. And one could argue Google was a first AI-first company, and back then it was called machine learning and it was different technology basis in some sense. And I think 2012 was when AlexNet came out and there's this proof that you can start scaling things and have really interesting characteristics in terms of how AI systems work. And then 2017 is when the team at Google invented the transformer architecture which everything is based on now, or roughly everything. And so for example, if you look at GPT for ChatGPT, the T stands for transformer. And around 2020ish, I think, was when GPT-3 came out and that was such a big step from GPT-2. And it still wasn't good enough to really do stuff with, but you're like, 'Oh shit, the scaling wallpapers are out. The step function and capabilities was huge.' You suddenly have a generalizable model available via an API that anybody can ping. And so just extrapolate that out to the next step. And this is going to be really important. So it's basically looking at that capability step and playing around with the technology and then reading the scaling law papers, or just in general the scaling laws seem to work for everything, and you're like, 'Wow, this is going to be really, really important,' so let me start getting involved with it.

I guess it's three things. I mean, I've always ingested a lot of information from a lot of different fields just because I like learning about stuff, and I was always this mix of like math and biology and anime and art and other things. So, you know, it was always kind of a mix. And then it was something that my friends were talking about, but it was a bit more like toy-like, 'Oh, this is cool and look at what came out,' but most people didn't then extrapolate. It's kind of like early crypto or Bitcoin. Like, everybody was talking about it, but very few people bought it. And so, I think that was part of it. And then third, honestly, I just thought it was really neat stuff that I kept playing around with. This is back to the GAN stuff and the art where these different models would come out and you could mess around with them. And one of the things that's really under-discussed in terms of the importance of it relative to this wave of foundation models and AI and everything else is the way AI or machine learning used to work is your team at a company or wherever else would go and there'd be what's known as an MLOps team, operations team, whose whole thing was like helping you set up all the data and the pipelines and everything to train a model, and you train a model that was custom to your use case and what you were trying to accomplish. And then it was you had to build a bunch of internal services to interact with that model. So it's a huge pain to get to the point where you had a working ML system up and running in production, and then suddenly you have a thing where you just do an API call. So with a line of code or a few lines of code, anybody anywhere in the world can ping it. But not just that, it's generalizable. So it's not just specialized to one use case like spell correction or whatever. You can use it for anything. And it has all of the internet embedded in it in some sense in terms of the knowledge base, and it can start having these advanced reasoning capabilities. But one of the most important things is hey, you can get it with a couple lines of code. You don't have to go and build an MLOps team. You have to host it. You have to interact with it. You don't have to do all this extra stuff. It just works. That's really important.

Yeah. Are you finding any signal there? Yeah, it works pretty well. Wow. So, I've been doing the weird shit, right? Like practice smiling people. Yeah. Yeah. I think it's interesting, right? Because we do this all the time where we read people, right? And that's part of the prompt. It's like you're a very good cold reader of people based on micro features and etc., etc., kind of spell it out and then based on that, not only give me your interpretation of this person, but explain the specific micro features for each thing that you're stating about the person, and it'll break it down for you. It's amazing. Like, imagine what this technology is. It's crazy. And again, I'm not saying it's fully accurate and I'm not saying it'll be predictive, but it's done pretty well in terms of nailing people. It's even done things like, 'Oh, this person probably has this type of sense of humor,' or, 'This person probably holds themselves back in most social settings and then chimes in with a witty right thing that nobody expects,' or what. I mean, it's very specific. It's very specific. Wow, that's amazing. Right. And so, I've been doing stuff like that, which may not be your question, but I've been finding it really fun. It's related, right, in the sense that, and I'm sure I'm missing some steps, but I love angel investing and the dose makes the poison. So, there's usually a case to be made when I get to a certain threshold, I'm like, 'Okay, this isn't fun anymore.' Like, I love dark chocolate, too. But I don't want just to be force-fed dark chocolate all day. But, and you and I have talked about this, right? But I really do enjoy the learning and the sport of it frankly and interacting with some very, very smart people. Not all of them work out as far as founders of companies, but ultimately I'm trying to figure out how to separate signal from noise. And also it's fun to try to use anything, but in this case investing, to sharpen your own thinking, right? And to stress test your own beliefs and the assumptions that undergird some of your predictions, things like that. I'm just wondering if you've ever done like sort of a retrospective analysis of your startup investing or if you're like no more market reason style, only forward.

Yeah. Early on when I was first starting to invest, I would have this long grid of things by which I would score each company, and then I'd go back and see if it was correct. It was roughly correct. I think the hard part is there's a lot of like randomness in outcomes. There's the company that sells for a few billion dollars that you thought was dead or whatever it is. Sure. And so how do you score things like that? Right now we're in this really weird market moment where trillions of dollars of market cap are all chasing the same prize and so they're going to do all sorts of stuff that wouldn't happen normally. So it's really hard to account for that kind of thing, right? Relative to all this, I'm much more in the Mark and recent camp of like I think very little about the past. I think close to zero about my own past. I just am like, let's keep going. And maybe that's bad and there should be dramatically more self-reflection. And I try to self-reflect in the moment, but I don't try to re-extrapolate and examine my entire life and decisions. And if anything, most of the decisions have been ones where I'm really upset with myself for not being more aggressive on something. In other words, I invested in the company, but I should have tried even harder to invest more even if I tried really, really hard because there's a handful of companies that really matter and that's all that kind of matters as an investor. Obviously, as a person, I enjoy getting involved with different companies and different founders and helping them whether the thing works or not or I think the technology is interesting or whatever. But the reality is from a returns perspective, there's a very clear power law that people talk about and it's true. And I remember a friend of mine did this analysis. I think it may have been Drew Milner or someone where it's like look at all the companies from like, I don't remember the exact dates, 2000 or 2004 until today in technology. And it was something like a hundred companies drove like 90-something percent of all the returns. And 10 companies total drove like 80% of all returns over a two-decade period in technology. If you weren't 10 companies, you were a bad investor. And once you start dealing with these power laws and these outcomes, how can you rate that? Right. It's basically, did you hit one of 10 things or not? That's really the rating. That's probably the correct rating for investment.

I think to your point, I've always been terrified of losing other people's money. Like I'm fine if I lose my own money. It's my decision. I'm an adult. It's okay. But I've always been, and people giving me money are adults or institutions etc. to invest on their behalf. But similarly there I was just terrified of ever losing money for people. And so I've tried over time to be judicious behind the SPVs that I did early on. And the focus was on things that I thought would really be outsized companies. And so that was to your point Instacart. It was early Stripe. It was Coinbase. It was a couple things like that that were amongst my very first SPVs. And the emphasis was very much on do I think this can be a massive thing and also do I think there's enough downside protection in some sense that even if it didn't work as well as I thought it would, it would still be a good outcome for people. So yeah, I try to do that very diligently. It's interesting because a lot of people ping me for help as they think about becoming investors or they're scouts for a fund which means basically they're given a small amount of money by a venture capital fund. Sequoia famously has this program. They give people money and then those people invest money on their behalf. And some of the scouts that I've talked to basically treat it like free money or an option. They're just kind of like, I'll just write a bunch of stuff. Maybe something works. And I pointed out to them, hey, if you actually want to become a professional investor at some point, this is kind of your track record. A, you're a fiduciary in some sense, so maybe I'll be more careful from that perspective. But B, you know, this will establish like your track record. And do you want to have a good one or bad one? And how do you think about that? And again, sometimes people just get lucky and they hit the one thing out of a hundred, but that more than returns everything and they look great. But it's hard to be consistently good at this stuff or consistently hit great companies.

I think there's a big difference between early and late things. On the early side, there's a point earlier I tend to spend a lot more time in the market than most early stage investors. Most early stage investors say, 'I just care about the team and how good are they?' But I've seen amazing teams crushed by terrible markets and I've seen reasonably crappy teams do very well. And so, you know, at this point, I think the market is more important. Although I think obviously great teams can find their way if they decide to shift around a bit. I index a lot on market early and that may be customer calls. That maybe is trying to understand, do I think something could be big? It could just be some intuition around, hey, you know, defense is really important. Nobody's doing defense. Let me find a defense company. Right? I tend to index a lot on that. And relatedly, I've tended to avoid science projects. And there's some people who get really distracted by, 'Wow, this is really cool. It's quantum and it's this and it's that.' And I've largely avoided those things. And, you know, sometimes I miss things that were really good. But often that was the right call. I actually think SPACs saved sort of hard tech and science-based investing industry because if you look at what happened basically at the market peak, a bunch of SPACs took a bunch of companies public that would not have been able to raise money in private markets later and they gave them enough money to keep going but more importantly they returned a bunch of money to these hard tech funds and that saved them from going under. It gave them all the returns was basically the SPAC era. So, Chamath basically saved hard tech. I mean that seriously, not cheek. And I largely avoided that kind of class of companies. And I'm not saying it was smart. I would have made money off of it. I just thought there was all sorts of capitalization issues and science risk and market risk and other things to them. For later stage stuff, the hard part often is everything on paper gets modeled out for a late stage company as a 2 to 3x from that investment point, right? Because all the funds that are driving the rounds underwrite against some IRR, a clock 25% IRR or whatever it is, and so they all come up with these models and the models all say all these companies are basically going to 2 to 3x and the art there or the science there, whatever you want to call it, is is that a 0.5x company, is it going to drop in value, or is that a 10x and how do you know it's a 10x versus a 2 to 3x versus a 0.5x? And that's the harder part of growth investing and there's a subset of things that you're like this thing will just keep going and here's why but often it's not mathematical often that's just like some market dynamic or some core insight or some market share question and people tend to make that stuff really complicated and they have these really complicated multi-page models and 50-page memos and all the rest and often these things boil down to one single question. What is the one thing I need to believe about this company that makes me think it's going to continue to be really big? If it's three things, it's too complicated. It's probably not going to work. If it's no things, then it doesn't make much sense. So usually there's one or two things that are really the core insights you need to understand like the outcome for something.

I'll give you two or three of them. I mean, Coinbase part of it was just hey, this is an index on crypto and crypto will keep growing because if Coinbase trades every main cryptocurrency and they take a cut of every transaction and have enough volume to effectively bought a basket of every cryptocurrency by investing in Coinbase. That was the premise there. Stripe it was they're an index on e-commerce and e-commerce will keep growing back then. Now it's much more complex and there's all sorts of great drivers of its performance. Anduril was hey, machine vision and drones are going to be important. AI and drones are going to be important for defense. That's it. I mean it was more complicated than that. I'm just saying like that was the fundamental, well that was it for the belief, the core belief. There was like cost-plus model versus hardware margin. You know, Andrew actually had four or five things that were important there that were kind of like a checklist for a defense tech company, but for a lot of the other ones, it was like e-commerce is good.

Well, I first invested in Stripe when it was like eight people and then I kept following on and I ran out of my own money, frankly, and that's when I started doing SPVs. So, I think I did my first SPV in Stripe around the Series C-ish. We're in there. Something like that. Got it. And were the others more or less similar-ish? Instacart, etc. It was probably roughly in that ballpark, Series C, D, kind of that range. I didn't have funds and everything else. And, you know, I was putting as much as I could personally into these things both earlier, but honestly, I just kept going when I could.

Oh, I mean I do enormous due diligence. So meet with the CFO multiple times, walk through all the financials, walk through the financial model, walk through customers, call customers, look at executive team, you know, it's a bunch of stuff. My fund is the only one I know that actually does like cash reconciliations where we'll go through and do a cash audit to look at cash flows for later stage things. So I do enormous diligence because I want to make sure I'm not doing something inappropriate. But the flip side of it is most of it just collapses into like what's the one thing? So when I work with a company, I actually try to be very fast and straightforward on the diligence in terms of saying let's just talk about, we need to just make sure financials are correct and you know like there's the basics but like let's collapse it down into one or two core questions right that help us understand if this thing will keep going, not here's 30 pages of questions that don't matter, right? Which is what a lot of people, they're like hey we need to know the secondary cohort on this fucking thing that's like a tiny product that who cares, they just waste time. They waste the founders' time and the team's time. And I try very, very hard not to do that. As a former entrepreneur myself, I know how precious the time is and I know how annoying those questions are.

I haven't looked at that post in a really long time. I'm actually writing another book now that is sort of the zero to one startup phase and it gets into some questions like that. I think the reality is venture capital has changed dramatically since I wrote that post. Right. Because in 2011 the venture capital funds were largely doing seeds through series D maybe and then companies would go public. Yeah. This whole 20-year private company thing didn't exist. Do you know why there's a four-year vest on stock?

In the 1970s, they came up with a four-year vest on stock options for employees because companies would go public within four years. And so then you're done. Yeah. Yeah. Literally, right? And so it's like a four-year clock usually. And then when Google took six years to go public, everybody's like, 'Oh my gosh, it took them so long to go public. Six years, like they just sat on their hands.' Do you know what I mean? Literally people would say that, right? And so what happened is venture capital used to be very early stage and then what we now call growth investing was public market investing, right? That was a stop that people in the public markets would do after four or five years of a company's life. And so public markets used to be involved very early. And then as Sarbanes-Oxley came out and companies decided they didn't want to go public and there's more private capital available, the timeline until going public stretched out, right? And so suddenly venture capital firms are doing all the growth investing that used to be public market investing. And in 2011 that really wasn't happening much. It was kind of Yuri Milner from DST and a few other folks, but it wasn't that much of an industry. And so the nature of venture capital has shifted radically over the last 15 years. And that means those questions that I listed there didn't include what I'd consider more growth-centric questions because there wasn't a lot of growth investing in venture.

Honestly, it would overlap with some of the early stages. You know, by the time you hit a very late stage, it's very financially driven. And so often what at least I and my team look at is what is just the core business and how do we extrapolate that going and then what are these ancillary things that the company's doing that are almost like options in the future that may or may not come through. And so usually we base our investment on that core. Can they just keep doing the thing they're doing forever? Because most companies mainly get big off of one thing at least for the first decade, right? There's very few companies that end up with multiple things that all work. Usually with one thing and then 10 years later you maybe come up with the second thing that really works, right? It's like Google Cloud for Google, although obviously there's YouTube and there's a bunch of other stuff and Waymo and all these interesting things now, but it took a while, right? For a long time just search, search and ads. But then sometimes there are these extra things that are potential really interesting drivers on a business. Like SpaceX was launch and then it became satellite, right? It became Starlink.

So, when I read the book, I had an outline for it that was two, three times the length of the actual book in terms of chapters. So there's a lot of stuff I didn't write about sales and marketing and growth and a bunch of other stuff. But the book was basically written as sort of like a tactical guide. It wasn't meant to be read from start to finish. There's a bunch of interviews with different people who I think are amongst the best practitioners in the world at those areas. But fundamentally it was meant to be more like you're suddenly involved with the M&A, jump to the chapter and read that and then put it aside until something else comes up around hiring that you need to look at or whatever. And so it really is meant to be like a handbook or guide or companion to a founder versus, hey, I'm just going to read it start to finish and there'll be some pithy quotes in it or whatever or one concept over 500 pages. You know, I try to avoid stuff like that. It's very tactical. It's very tangible. It's very specific. And this new book that I'm working on is basically the zero to one version of that. It's like how do you hire your first five employees as a startup? How do you, somebody tries to buy you, what do you do? How do you raise your first round of funding? You know, it's that kind of stuff. So, it's kind of like the zero to one technical guide.

So I think when founders pull together boards, often the early boards are investors because the investors ask for a board seat as part of it or as part of the investment and sometimes the founders want somebody on board who's really committed to the company and will help out extra. And to some extent when somebody takes a board seat it really means or it should mean that they're all in to help you versus you can have lots and lots of investors if you have very few board members. Reid Hoffman has this thing which is like a board member at its best is like a co-founder that you wouldn't be able to hire otherwise and so you bring them onto your board. It's somebody that you want to spend more time with on specific issues related to the company. Fundamentally, your board should be able to help with different areas of the company. It could be strategic direction. It could be closing candidates. It could be product areas. It could be customer intros. It could be a variety of things. And usually, you want to kind of think of your board members as a portfolio of people. It's going to change between an early stage company and a late stage and a public one. You only need different types of people over time usually. But most companies are very reactive on their board versus proactive. And so they tend to end up with a couple investors and then they kind of add somebody from an industry seat and they don't really think through like who they want and why. And if your co-founder is kind of like your spouse, your work spouse, your work husband or your work wife, your board members are like your in-laws. You know, you have to see them at Thanksgiving and you have to chat with them all the time. And so hopefully you have somebody you want to see all the time and who's helpful and wonderful. And the bad version is like, ugh, it's the like father-in-law or mother-in-law who's always like berating you or whatever. And so you kind of need to find the right person. And it's for many, many years, right? You end up sometimes with people on your board for a decade. And if they're an investor, you can't get rid of them. You literally can't fire this person because they have a contractual ability to be on your board because of the investment. So that's why it's really important to figure out the right person. And that's back to valuation. Sometimes founders will take a better price from a worse person because it's a better price. And our mutual friend Naval has this great quote that valuation is temporary but control is forever. Yeah. Very novel. Very novel. And I think that's very true. And so if you're choosing a board member and part of that is a control thing. People who control the board can in some cases fire the CEO. You really want to choose the right people and maybe take a worse price for somebody who's really going to be helpful and they're minimally non-destructive and hope you get to have them around for 10 years.

I don't have anything super useful there. I think the best thing is to call other founders, other people who've added people to their board and see how they approached it. I do think writing up a job spec, you write a job spec for everything else in your company. Why wouldn't you write one for a board member? So, it's good to write that up and say, what am I actually looking for and why and what am I optimizing for? So, there's a common view of that. You can use search firms, you can ask people, you can target people that you know, you know, if you have angel investors, getting to know them is a great way to see if you want to add one of them eventually to your board. That's what we did. We eventually added Sue Wagner, who was a co-founder of BlackRock, onto our board. Her other board seats were Apple, BlackRock, and Bridgewater when she joined our board, but I just got to know her through just like she invested and we just started working together and really enjoyed her feedback and insights and so we added her to the board there. So it's kind of like that, you kind of want to maybe get to know some people.

Yeah. Yeah. I mean actually the origin stories for founders is always like ever since Sarah was three years old she dreamed of starting an accounting software firm, you know, like come on, you know what I mean. And so a lot of the stories that are told about founders are very revisionist and they make it the life's passion of this, you know, and sometimes it really is but you're like no when they were five they did not, you know, collect things and then that turned into Pinterest 30 years later or whatever. They always dreamed of building AGI when they were four and that's why Sam started OpenAI or whatever. So I think a lot of these things are very kind of ridiculous in terms of how they're written later. And I think the product really, really matters and I think sometimes great product just wins and the reason great product just wins is it opens up a form of distribution that didn't exist before or people will buy it despite the lack of distribution or relationships for a company. And the flip side of it is though the companies that are really good have an enormously good product engine and then they have an amazing distribution engine and sometimes that distribution engine is built into the product that's like Cursor or Windsurf just distributing through product-led growth where developers just find it and start using it and it helps them and so they tell other developers and it spreads word of mouth but often there's very aggressive sales, marketing, other components to it. And so for example when I was at Google they were spending hundreds, hundreds of millions of dollars a year, which at the time was real money, on distributing search. And they had this little thing called the toolbar that would like fit

Back then the main browsers were like Netscape and Internet Explorer, and the browser bar thing didn't exist. They had this little client app that you'd install, and they paid basically every company on the internet to cross-download it.

So, very aggressive distribution tactics. And to your point, that was Facebook and Facebook buying ads against people's names in Europe.

They're basically trying to create network liquidity in markets where they were earlier behind. And so, they would basically buy ads of literally a person's name. And one of the most common queries is people searching themselves. And so you'd be like, 'Oh, let me look up Tim Ferriss on Google or whatever.' And there'd be a Facebook ad saying, 'Hey, Tim Ferriss on Facebook.' And you'd click and you land on the signup flow for Facebook. Right? This was years ago. This was TikTok and ByteDance, right? It was basically they spent billions of dollars distributing TikTok so they could build enough of a network to train AI algorithms to start telling people what to do and also to get content creators on.

My guess is ads. Again, you kind of see this over and over again. I mean, for enterprise, Snowflake spent billions of dollars on salespeople and compensation and channel partnerships. So, again, distribution is really important.

Every once in a while, you see a company that actually wins not because of product, but because they're just better at sales and marketing and distribution. And often that's a bummer for technologists such as myself because you're like, you know, the best product should always win.

You know,

Yeah.

It's really hard because you often end up you start off with some set of beliefs. You think something's interesting or maybe you invest in it, maybe you start a company in it, and then it turns out the thing you think is really interesting turns out to be really hard and you get killed and then 5 years later a company comes up that actually does it and wins.

And the question is why? Why did the thing suddenly work when it didn't before? Or there's 10 attempts to do X and then suddenly is it the technology got good enough. It could be a regulatory change. It could be a market shift. It could be whatever. An example that may be Harvey and legal where selling to law firms traditionally has been awful and Harvey is not much broader than that, right? They also had very strong enterprise adoption and lots of different people using them in different ways. But the dogma was always like building stuff for law firms is crappy as a business and you should never do it. But what AI did is it shifted things from selling tools to selling work product or selling units of labor. That's really the shift in generative AI. We're going from seats and we're going from software and SaaS and we're moving into a world where we're selling human labor equivalents. We're selling work hours or labor hours or whatever you want to call it of cognition. And so Harvey is effectively helping really augment lawyers in different ways. And part of that's a knowledge corpus, but a lot of it is this tooling that really helps lawyers achieve the goals that they have in different ways in a collaborative manner in some cases. And so it's just a fundamentally different type of product from what people were selling before. And so it opened up the market in a way that the market wasn't open before.

There's actually a broader conversation around is the world market limited or founder limited in terms of entrepreneurial success. The Y Combinator school of thought is that we just don't have enough founders and if we had 10 times as many founders, we'd have 10 times as many big companies. And there's an alternate school of thought which is how many markets are actually open in any given moment in time. And those are the ones where you can build big companies because if the market isn't open to innovation or change or whatever or hasn't is undergoing a shift, you can't really build anything. So why do it? And the striking thing about AI is it's opened up tons and tons of markets that were closed for a long time. And it's opened it up because of capabilities, but it's also opened it up because every CEO is asking themselves, 'What's my AI story?' And we're way more openness to try things than I've ever seen in my life. And so we have this odd moment in time where things are massively available for founders to do new things.