The a16z Show - When Giants Don’t Go Public: Inside the $5 Trillion Private Tech Market

Episode Date: February 26, 2026

Bloomberg's Odd Lots hosts Joe Weisenthal and Tracy Alloway speak with David George, general partner at a16z and head of the firm's growth fund, about why $5 trillion in tech market cap now sits in the private markets, how that figure has grown 10x in a decade, and what it means for founders, employees, and investors. They also cover SPVs, tender offers, the collapse of legacy software valuations, and why AI companies may be speed-running the path to public markets. This episode originally aired on Bloomberg's Odd Lots podcast.   Resources: Follow Joe Weisenthal: https://twitter.com/TheStalwart Follow Tracy Alloway: https://twitter.com/tracyalloway Follow David George: https://twitter.com/DavidGeorge83 Listen to Odd Lots: https://www.bloomberg.com/oddlots Stay Updated:Find a16z on YouTube: YouTubeFind a16z on XFind a16z on LinkedInListen to the a16z Show on SpotifyListen to the a16z Show on Apple PodcastsFollow our host: https://twitter.com/eriktorenberg Please note that the content here is for informational purposes only; should NOT be taken as legal, business, tax, or investment advice or be used to evaluate any investment or security; and is not directed at any investors or potential investors in any a16z fund. a16z and its affiliates may maintain investments in the companies discussed. For more details please see a16z.com/disclosures. Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

If you actually want to invest in the highest growth, most promising companies that could be that next mag seven, chances are they're in the private markets. You know, AI, first of all, I think these have the potential to be some of the best businesses ever created. They're run by exceptional founders. They're building products that have grown at rates that we've never ever seen before. So, you know, they're kind of speed running the process of company growth. models are improving at a like eye-popping rate you know they can basically double their ability to complete long-form tasks over six to seven months and so you know if you were to just arrest model development today I think we would have the chance to build 10 to 20 years of really interesting applications on top of it highly valued private tech companies now represent about five trillion dollars in market cap almost a quarter of the s and p 500 10 years ago, 88% of market cap creation for the best tech companies happened after they went public.

For the recent crop of IPOs, 55% happened while they were still private. David George runs the growth fund at A16Z. This conversation previously aired on Bloomberg's Oblotz podcast covers why companies stay private longer, the SPV problem, why legacy software is getting crushed, and why outcome-based pricing could be the business model shift that finishes off incumbents. Bloomberg's Oddlots hosts, Joe Wisenthall and Tracy Allaway. Speak with David George, General Partner at A16Z. Hello and welcome to another episode of the Odd Lots podcast.

I'm Joe Wisenthal. And I'm Tracy Alley. Tracy, it feels like 2026 could be a big year for some mega IPOs that have been private for a while. There's talk about a SpaceX IPO, possibly, maybe. some of the big AI labs, like some pretty massive companies that might be hitting the market soon. Someone recently gave me a Facebook IPO hat from J.P. Morgan when they worked on it. Like, I'm very proud. I need to start wearing it around the office. Yeah. But that was like, that was a mega IPO at the time. And there was so much hype about it and then like technical difficulties.

So many people eager to get in on that one. It's so many funny. People called that a flop, I guess, because of the technical difficulties, and it didn't do that great for a little bit. That would have been a great time to buy it. Yeah. Seriously. The interesting thing about the market or one of the interesting things about the market is you have these companies that are going to IPO when they're already gigantic. So, like, people point out that in earlier eras, they might have IPOed when they're

like billion-dollar companies, and now they're like octacorns or whatever. And then you have other companies that are also enormous. And there's no, it's not clear that they're going to IPO at all. At all. You know, I saw a headline. about Stripe, perhaps raising more money. People, they could have probably IPOed years ago. And one of the questions I have is, are companies choosing not to IPO or delaying IPO

because the public market is not that fun? Or because the private market has gotten so much richer, so much more liquid, et cetera, that that impulse to go public just isn't the same way as it might have been in a different generation. Yeah, this has kind of been a long-running question in the market for a while now. But one thing I would just point out on the last point, it feels to me like companies in the private market, even though they're in the private market where presumably the pool of capital is smaller. Yeah. It feels like they're always fundraising. This is the other thing, too.

That like it used to be when I started covering tech companies is like you're a series A round and series B round and C. And now it just seems like this permanent round, especially with some of the AI companies. Always be raising. Always raising. Anyway, we need to learn more about how giant companies are thinking about. capital markets, both public and private. And I'm really excited to say, we do, in fact, absolutely have the perfect guest. We're going to be speaking with David George.

He has the head of growth at Andrescent Horowitz, A16Z, someone perfectly situated to explain of all these things that's going on. So, David, thank you so much for coming on odd lots. Hey, great to be with you all. What is growth? I thought all VC was growth. What does it mean when we talk about growth or growth round when we're talking about private markets? Yeah, so first of all, it's great to be with you all.

Thanks for having me. Excited to have this conversation. So our early stage invest in companies that are growing fast, too. So if that's not clear, that is what we seek to do across all Bulls Capital. For us, growth investing companies at the later stage of their life cycle. So typically once they've found product market fit, and our early stage investing companies early stage when they're kind of trying to find product market fit, we invest in companies once they have found product market fit.

So that's the delineation. Well, what percentage of growth is sort of, directly obtained from the early stage. Because I imagine that's a good pipeline for you guys, right? Yeah, it's a great pipeline for us. We do both. So we have examples of investments out of the early stage and in companies that, for whatever reason,

we were not early stage investors. So some of our largest investments in growth are companies like Databricks and Stripe that we were early stage investors in. And then some of the largest investments are companies like SpaceX and OpenAI and Waymo and Roblox and Figma that we were not early stage investors in. and then the first time that we invested was in growth. So it's about 50-50.

We love to invest in companies where we know the founders really well. One of the things we talk about all the time is game film. This is something that public market investors are very familiar with. They look for game film. They want to back CEOs, management teams that have demonstrated continued success. And we seek to do the same thing in the private markets. If you just name a bunch of random big companies that are all sort of like the most important, like growing tech companies, they're actually just going to all.

be, you know, like when I said SpaceX, data bricks, stripe, etc. I didn't purposely say, they try to name a bunch of A16Z family companies. They just incidentally happened to me. What are you like, let's start big picture with this question of companies staying private

for longer, especially thriving companies, companies that are doing very well, IPOing, if at all, much later in their life cycle. There are a lot of specific questions, but big picture, what's the story? What's the main cause of this, the evolution of this trend. Yeah, absolutely. Well, first, I would love to just set the stage with what the

data actually is. Like, what's actually happened? If I'm totally wrong, that's also interesting. No, no, no, you're not totally wrong. I have a bunch of supporting facts to your statement, but the numbers are pretty eye-popping. So if you look at the private markets, highly valued technology companies represent about $5 trillion of market cap in the private markets, that's almost a quarter of the S&P 500. Wow. It's 15% of the NASDAQ. It's 40% if you exclude the Mag 7. So it's just staggering numbers. To your point about some of the biggest and best companies in the world being in the private markets, the 10 largest private companies represent 40% of that $5 trillion of market cap. So this is a massive power law game and the best of the best

companies at that stage of their life cycle just happen to be in the private markets and not in the public markets right now. That private market, kind of $5 trillion of market cap, that sector of the economy has grown 10x in 10 years. At the same time, the number of public companies, you guys have probably covered this before, the number of public companies had been cut in half over the last 20 years. So this is just a massive shift in the composition of the public markets and the private markets. And the best of the best companies, as you said, largely are sitting in the private markets today.

The other thing that we look for, you know, we're growth with growth means a lot of things. We look for companies that are growing, you know, sort of hypergrowth. We call it hypergrowth, growing very, very fast. If you look at the public markets today, there's only three companies in our universe that are growing over 30%. So if you actually want to invest in the highest growth, most promising companies that could be that next mag seven, chances are they're in the private market. So this is just a big shift in the dynamics, you know, certainly in the last 20 years,

but even the last 10 or 15. So that's the numbers behind the trend. The question is why. And there's lots of reasons why this is the case. I mean, you mentioned it right at the outset. There are structural reasons in the public markets that make it harder to be a public company now than it was 20 years ago. I think the biggest thing is the private capital markets are deeper and more liquid than before. So there's less of a need for a company to go public until they need much, much, much, much more capital. In the private markets, for the best, the best companies, they kind of have the access to. the capital that they need. It's a pretty compelling pitch for the founder to be able to stay in the private markets, have liquidity over time. If you're one of a select few that can access capital in the private markets cheaply,

you can kind of steadily control the stock price movements. You can regularly run tender offers. And for a company, probably the hardest thing aside from access to capital is employee management. And if you can get pretty close to the dynamics that you get in the public markets without volatility, it's pretty compelling to remain in the private markets. So I have a bunch of questions on just the growth of private capital. But before we get to that,

you said there are structural challenges in going public. And I guess the one we always hear is like, oh, you have to file quarterly paperwork, like not the paperwork, which I actually sympathize with because I hate filling out forms and things like that. But like, what are the reasons that the public market is perceived to be so much more difficult or so much more of a operational headache, I guess, than private. Well, there is a cost. So, you know, for a company the size of SpaceX or data bricks or open air or stripe, you know, it's not hugely meaningful. But for a smaller company, you know, it could be 10 to 20 million bucks. And, you know, if you're a company that would have gone public 15 years ago, you know, five years into your life doing 100 million bucks of revenue,

that's pretty meaningful. Like, that's a major change. And so you'll want to wait longer. Public markets, investors, investment banks, research organizations are tilted much more toward large-cap companies or at least mid-sized companies today and less so towards small-cap. So if you're a small-cap company, it's just very hard to get the attention of investors. You know, they can't write very large checks and build huge physicians in you. You know, it may not be worth the investment banks time or the research analyst time to cover you. And so, you know, if that's the case, it's hard to get the attention of good investors and make your stock price grow over time. So those are a couple examples of the structural challenges.

You know, the other one is just volatility. You know, this generation of founders has seen the 2021 run up and then the certainly in the technology market and in the falloff in 22 and 23. And some of them have reset their stock prices in the private markets. But that was a volatile time. Like if you were issuing stock to an employee at the peak and then, you know, that employer was looking at their stock grants and saying, oh, my gosh, these are down 70%. Like, my comp just got cut by 70%. That's a hard dynamic for a founder to have to deal with. And if you can minimize some of that volatility in the private markets as a founder,

I understand why you would do it. Talk to us about employee retention, employee management. Employees, you know, they love an IPO. They love liquidity. And, okay, now you can have tender offers where they sell some of their stock to new investors. And then there's other dynamics, too. because there's all the SPVs out there and maybe like tokens where someone can hedge their holdings in a private market on hyperliquid or something like that. Talk to us a little bit about the sort of the plethora of options that employees have now in private companies and getting that liquidity.

And so they actually get some real money. Let me start by just explaining the way it works in the public markets. And, you know, we can contrast the private markets with that. If you're a public market employee and you have RSU grants, this is basically. basically the way it works. You get a quarterly deposit of net stock, net of taxes in your account every quarter. And so if you're a highly paid employee at meta or Alphabet or Apple, it's like clockwork. You get these grants. They hit your account. You have for some of the highest paid people, you know, hundreds of thousands or millions of dollars that hit your,

hit your account every quarter and that's liquid. It's already netted of tax. And so you can turn around and sell it and it's like cash top. Or in the case of those companies, they've appreciated tremendously in value over the last 10 years. And so the employees, you know, who help but stock have been handsomely rewarded. Then they get RSU grants stacked on top of those. So you kind of have this like waterfall of RSU grants over time. And increasingly, you know, over time, your quarterly net deposit can go up. And so if you're a private CEO competing for talent, that's what you're competing with. Right. And those companies are flush with cash. And so they're able to pay a lot in stock-based comp to their employees.

And so that's a difficult dynamic to compete with. That's one of the strongest forms of the argument for telling a private company CEO that they should go public. Because that is a very compelling financial reward for your employees in that public market situation. In the private markets, the way it typically works is you get, you know, you get these RSU grants. Sometimes they stack up over time, but generally they're illiquid until you go public. Now, what's happened over the last, I'd say, six or eight years is companies will do tender offers in the private markets where they offer to buy a certain percentage of employees vested stock in the private markets. So they'll set the price, they'll work with somebody like us, and they'll say something like, hey, you can sell 25% of your vested stock that you have in this tender offer and we'll do it once a year. And so that's a decent substitute for what I described in the public markets.

It's not a perfect substitute, but for, you know, the employees or potential new hires for true believers. I think it's enough to combat that RSU public market dynamic and compensation scheme. And, you know, it's worked pretty well. SpaceX has, you know, famously done a really good job running twice a year tender offers for their employees. And, you know, they've had tremendous employee satisfaction, ability to hire retention, et cetera. and some of the biggest tech companies have followed suit. And I would argue that there's a tradeoff that the founders and the private markets are making where often if they were public, they probably would have a higher stock price.

So maybe they're taking a little bit more dilution. And so maybe employees would benefit a little bit more in valuation in the public markets. But it's not a massive gap. And it's a pretty compelling alternative. Just culturally out in San Francisco, do people still expect like the default exit to be an IPO or has that mindset kind of gone away? The best, the best companies want to IPO. And I think you mentioned some of the big name companies out there who may stay private

for a really long time. You know, I think the ambition for most founders is still, you know, to have an IPO and to be a large, established, important public company. You know, we talked about the pools of private capital and how they've grown over the last 10 years. The pools of capital in the public markets are still much deeper. So if you need to raise $50, $100, $200 billion, like some of the big, big companies may want to do to pursue some of their goals, chances are they'll end up in the public markets. Part of the reason we wanted to have this conversation, a few months ago, Tracy and I interviewed the perplexity CEO, and he was talking about the rise of SPVs and wanting to say no to money.

You know, people come to him with investment opportunities. And then it's clear that they just want to package up that equity in some way and create a secondary. instrument. And I know, you know what, someone messaged me recently and they're like, I could get you some pre-IPO Anthropic shares. I'm like, no, I don't do that. But like, I know that there's a lot of interest in there. And you mentioned that if you really want massive growth at this level, that it's the only ones that growing super big are the private companies. Talk to us about the emergence of SPVs and these third party entities and you don't even know who's on your cap table and how founders are thinking about this phenomenon. I think founders for the most part really don't like

it. Okay. Because they want to know who is on their cap table. And there are certain types of investors who will come to the founders and misrepresent, I think, what vehicle or where the capital is actually going to come from. In our case, we don't do those. What we do is we invest directly out of our funds. You know, we make that a point of pride with the founders to say, hey, you know exactly what you're getting. We're going to shoot you straight. For the SBB industry, you know, look, it works when, you know, times are good and it can be horrendous and bad, you know, when times are bad. You know, Andrel is one of our companies very close with. We're one of the largest investors. And they have famously, you know, gone to war with some of the SBB hucksters who are

trying to assemble capital to do a deal through some obfuscated way around them. And they say, look, we want to know exactly who's on our cap table. You know, we want to make sure that we're doing right by our shareholders. And so it's a risky maneuver. You know, it's not to say that there's not some that are good. But we try to avoid it and we try to counsel our founders to stay away from it as much as possible. So you mentioned Anderall, and can founders completely lock them out or like how does it work such that the founder doesn't want to have these vehicles get access to their equity and yet somehow they do anyway? What is the path into the company that these entities are taking?

The example would be a fund shows up to the founder and says, yeah, we're going to invest out. of this fund. And they have some legal entity name that they have as the fund. And then it's not fully clear to the founder that that legal entity actually will just be an assembly of a bunch of new investors that are only investing in that single vehicle. I would say the SBV interest is one other sort of additional risk that comes with it. You know, are large investors, RLP's. They're some of the largest institutions in the world. They trust us to make investments. And they actually like the fact that they're investing in a fund that has some degree of diversification. SPVs are inherently risky because they're single company. And so, you know, if that company

happens to not go well and you invest in A16Z, you can live to fight another day because we have a number of other companies. You know, it could be devastating if you put a large amount of capital into it. Talk to us a little bit about pricing because something you tend to hear is that, well, if you go to the private market, there are some benefits to that. But on the other hand, you're probably going to pay, you know, a little bit more in terms of cost of capital. And then the other thing you hear is that, well, if everything is private, it's not necessarily being marked to market as often as the public market. So maybe there's some concern around pricing. You're not getting that wisdom of the crowds effect. You're just getting a bunch of tech bros investing in tech bros. And that can

kind of be a self-reinforcing cycle. Yeah, look, I can only speak to our own business that we're in. Interesting data point, if you look at the returns generated by dollars in the private markets, historically over the last, call it, like, seven years of good IPOs, about 50% of the dollars of gain in an IPO come from the seed through series B. And then 50% of the dollars of gain come from the series C and later. as companies have stayed private longer, that will massively shift to the series C and later, and the C through B will be a smaller proportion of the dollars of gain.

Similarly, if you go back 10 years and you look at all the companies that have gone public, there's a little bit of like time lag in this, so just bear with me. The best companies that were going public 10 years ago generated 88% of their overall dollars of return in the public markets. So if you just took total market cap creation, only 12% of it was happening in the private markets. If you look at the recent crop of IPOs in the last five years, 55% of their market cap creation

happened in the private markets. Wow. Forty-five percent happened in the public markets. So there is a massive shift that's taken place in terms of where value creation happens. That's benefited us. In terms of pricing, we've invested in a portfolio of great companies. I've listed all the, some of our investments, some of our biggest investments are companies like Databricks and SpaceX and Waymo and Open AI and Anderl, you know, and strike and flock

safety and companies like that. Now, I recognize there are flaws with revenue multiples and all that. But I would say if you could let me have a career, an entire career of investing in market leading great technology companies where we could buy them at 21 times revenue and they're growing 100%. That would be an incredible trade or a incredible investment. opportunity. And so in terms of valuations, I feel like there is a discount to being in the private markets. And founders, I think, understand that for the most part and make that trade off. But it's definitely a dynamic that we see. There's one more thing that I would call out. You know, I mentioned earlier that only three companies in our universe in the public markets are

actually growing greater than 30%. There is a dynamic where I do think it's hard for public market investors even to rock really, really high growth rates. So if you're growing like 60%, I think automatically a public market investor is going to build a financial model that says 60, 50, 40, 30, 20%, and then they're going to value you as such. And then they'll pick an exit multiple at year five, and they'll apply it to 20% growth. And they'll probably say at that point, you'll be 20% margins. and they'll be happy and they'll call it a day and they'll have your stock price be that. And so I do think people like us

with a longer time horizon in the private markets may have an easier time of actually grocking that very high growth rate. But for the most part, I think it's pretty hard for them to grok that. And so again, we have conversations with the founders of these companies in the private markets. You know, I think we understand it.

I think we understand multi-product and you're less likely to find, you know, a full public market of folks who will give you credit for that hyper-price. growth. I find this to be a very interesting observation. I mean, just generally public or private, it does seem like several generations ago, you assume, okay, here's a really big company. It's not going to be one of the fastest growing companies in the world. And you apply that sort of model mindset where you mark down their growth expectations. I mean, even Alphabet, I think in its latest

quarter actually had faster top line growth than it had in the quarter before, like on a year-over-year basis. I have a question for you, something I've always wondered, okay, let's say one of your big portfolio companies goes public eventually, you know, a stripe or something like that. How does a GP, or how does the firm think about selling at that point? Is it, okay, you're a private market investors, so you get out fairly soon, you don't want to hold that. Do they get distributed? And then it's up to the LPs to think about it. What is the sell decision like once a company is no longer private? Yeah, it's a great question. And by the, way, to your point on these large cap high growth companies, it's a really good learning for all of us

that the very best companies can, you know, sort of defy your expectations and still grow fast, even though they're big. Like it's sort of the breaking of the law of the absolute highs of a company, like meta-accelerated revenue growth to 30% last border. Like that's shocking. A company that's worth, you know, almost $2 trillion can accelerate to 30% growth. Same with Google, nor if the 20% growth. So if you can find those opportunities, I would say strike. is a good example of this. Previous generations of this would be like Visa and MasterCard. Like if you had a strong thesis that those would grow at north of 20% for 15 years,

you would value it a very different way than if you thought that that growth would tail off over time. Sure. So, yeah, I think it's great that you brought up Google and, you know, we think about the large cap companies to try and inform what could go right with the best of the best, the companies and the private markets for us to, not just over five years, but over 10 plus years. So to your question about distribution, or selling when our companies go public, our LPs, our investors tend to like to have distributions

as opposed to us selling. Okay. You know, many of them have public desks or public operations on their own, and they would prefer to take the stock and, you know, manage tax consequences on their own or maybe they have, you know, a long book that wants to hold it. And so, you know, we tend to distribute companies. I would say, for the most part, when our companies go public, I'll speak from the growth where we're coming in at a more mature stage of the company, just because a company goes

public doesn't mean that we will exit. Now, we always seek to return capital to our LPs and we want to make sure that we're doing that on an appropriate time horizon. But often, you know, we'll find situations in the public markets where we think our companies are massively undervalued. You know, I'll give you an example. One of our best companies was a company called Samara, which does fleet management, video tracking for drivers. They're the market leader, second time founders. They were the founders of Morocco, which sold the Cisco. So they're an exceptional team. They went public right at the tail end of 2021. And the IPO market was starting to freeze and they were the last IPO. And so we ended up being the largest buyer in their IPO. So all of the public funds,

Fidelity, T.Row, some of them invested. But we ended up actually being the biggest buyer and then we held that for a long period of time, even though it was in the public market. So we'd like to think about, you know, what is the future of the company, even if they're in the public markets. We place a tremendous amount of value in founders running companies, which I'm happy to talk about. And if we think the growth prospects are really bright. What's the actual catalyst for going public then? Because if we think there's plenty of capital in the private markets, employees are generally pretty happy with their compensation, people can live with the capital cost. Why would you go public? Or like, what's the most common reason in terms of timing? Yeah, the biggest thing would be access.

to larger pools of capital. And, you know, you finally feel like you're making that tradeoff where the cost of capital in the public markets would just be much more attractive. And so you want to go do it there. So in the case of building, for example, data centers in space, that will require a tremendous amount of capital over time, you know, you can get a lot of that capital in the private markets. And, you know, maybe at some point it makes sense to tap the public markets to get that, you know, if you have huge ambitions. Sometimes it's easier to get debt. And it'll alternative forms of financing in the public markets as well. And then lastly, you know, if you feel like the competition is really fierce for your employees. And, you know, I described that dynamic of,

you know, quarterly RSUs and, you know, stock currency that might be a little bit easier to manage in the public markets. That would be another reason. It's not just employees. It's also if you want to do meaningful M&A, you know, there's probably a little bit of a benefit to being in the public markets and having a public currency that you can use where, you know, you don't have to debate with the sellers what the value of your equity is because it's in the public markets, you know, there's a daily stock price. So those are a few of the dynamics. Does the private market trend, does that hold as we see, you know, more and more tech is just about AI, right? And if there's one thing we know about AI, it's that it requires quite a lot

of capital investment. So does that start to change the dynamics or the balance of power between private and public in your mind? It goes back to the same point, which is at some point, the amount of capital required, it probably makes sense, you know, to be in the public markets. You know, I do think for large-scale consumer businesses, I think there's some value in being in the public markets, you know, letting retail take part in the ownership of your stock, you know, having greater brand recognition. You know, we saw this recently on the B2B side as well with one of our public companies. And after going public, you know, sometimes there's sort of a brand benefit that you get, Brand Halo, you're better known.

An IT buyer trusts a little bit more in your future. So there are some of those dynamics that exist. You know, AI, first of all, I think these have the potential to be some of the best businesses ever created. They're run by exceptional founders. They're building products that have grown at rates that we've never, ever seen before. So, you know, they're kind of speed running the process of company growth in a way, getting to be bigger and more consequential much faster than the previous generation of companies. maybe that means they should be in the public markets a little bit faster, too.

Maybe last year at some point, Open AI had a big tender, and then suddenly no one could afford to buy a house again in San Francisco because all that money went into real estate. But there's a serious question, actually, and particularly acute at the AI companies, is there any stigma of being an employee who sells some of their shares? It's like, all right, you can sell 25% your shares, and then it's like, I want to sell. It's like, what? You don't believe in the singularity? you don't believe that we're on the path to AJI. You don't think, you think we're done with our, you think we're almost at the end of our mission

and that our value won't be 10? Is there like, is there any anxiety on the part of employees at a real like mission driven fast growth company is of like, I want to dip my toe in the water. I want to hedge my own company. I want to diversify away from my own company. Suddenly the founder isn't playing ping pong with you anymore.

Yeah, exactly. Yeah, exactly. Yeah. You're not allowed to have lunch in the after you. I would feel like this would be a real phenomenon. I would be anxious about raising my hand. I think everything in moderation. Most of the time there's not like a stigma and most of the time it's not, you know, an opportunity to sell so much stock that it would be a vote of lack of confidence,

if you will, right? You know, if it's 25% of your vested stock and you've been there for two years, that means you probably have a ton of unvested stock. And so you're talking about a small proportion of your overall organs. You know, we never really see a chance for employees to, it's never designed In that way where employees say, oh, my gosh, I'm out. I don't want to sell 100% of my stock. In fact, we don't have data on this, but I would suspect that in the public markets, you know, employees are probably selling out of their stock grants at a higher rate than in the private markets.

And I think, you know, if you're at one of these companies in private markets, you probably have a greater degree of confidence. You know, if you're a good employee, you could always go to Google or Facebook or meta and click coupons. But I think that most of the time, the good ones are true believers. We do have companies where the founders, one of my favorite things, things is when the founders just say, I'm not selling a share. And, you know, that is like the ultimate extreme point of confidence. We spent time with a very high profile internet CEO of the previous

generation. It's now a public company. And one of the meetings we were talking to him about secondary and he was so resolute. He was like, I'm not selling a share. Like, why would I sell a single share? Like, I'm so confident it's going to go up. I don't want to do that at all. And that's a pretty strong signal of confidence in the future of the company. So we look for some of those signals when we invest. You know, my partner Alex Rampel wrote a good piece many years ago, to your point on real estate and buying a house and affordability. Yeah. That showed Bay Area stock prices, you know, of like the mega cap stocks indexed against home values. And it's like perfectly correlated. So, you know, home price appreciation is kind of tethered to the local economy in a way.

And so, you know, it's not totally surprising that you see it move in that direction. Yeah, I think it's fair to say there are a lot of people in San Francisco right now who are excited about making money off of AI and some of them have been doing so already. But on that note, how are you actually differentiating between AI models? I guess like this is a way of asking, how are you cutting through the hype? Because we see all these new companies. companies launching. A lot of them use similar language. We see total available market, Tam, being thrown around quite a lot nowadays. How do you decide this is actually like a good business model versus this is just something that has AI in the name and is getting some attention?

Yeah, so we're, and we're investors in, we're probably the largest investor in the AI industry as a firm. I think we're investors in two-thirds of the aggregate, AI revenue. So if you were to just sum together all of the AI revenue of the companies in the private markets, we're investors in about two-thirds of it. So I think we have a front row seat to it. First of all, I'll just start with the demand side. This is the biggest thing that we look to. There's a billion people plus using this technology, getting extremely large surplus or value out of it. The companies are the fastest growing that we've ever seen. You know, active users spend something like 30 minutes a day on it. And, you know, there's some, you know,

there's a tremendous amount of surplus that comes with that, even if you're paying as a subscriber for those. So on the consumer side, you know, I think the capabilities are extreme. They're very, very good. We're just now starting to scratch the surface on the capabilities of, you know, sort of doing work on your behalf. And so I think we're going to see a ton of progress on that front over the next year. But the demand signals that we see are probably, not probably there, are definitely the best that we've ever seen in my career, you know, certainly much faster than, you know, the internet phase or mobile social, you know, cloud, SaaS, e-commerce, any of those. So tremendous amount of value on the demand side, that's the biggest thing we look for. On the supply

side, you know, look, the buildout, there's, there's a ton written about the buildout of infrastructure that's required. It's, you know, it's larger than, you know, the U.S. highway system overall, you know, $5 trillion over the next five to seven years. The thing that I say, about the supply side is we don't need to make the decision on investing every single one of those dollars as an industry today. The industry can monitor demand and we can make decisions about capex as we go. Cycle times are not five years long. Cycle times are more like 12 months. And so if we see a weakening of demand or a lack of payback in some of this infrastructure build out, then, you know, that's fine. We can adjust force. The people who are, I think there's

probably a misconception about the people who are in charge of these companies and, you know, at the center of all this like us, like we and they are all sort of RIC or return on capital-minded people. Like, we're not going to, you know, invest a bunch of capital if we think there's not going to be a high return from it. So, you know, to the extent that there is some overbuild or, you know, signs of a lack of demand to meet the supply, I think the industry will adjust. You know, we've written about this before, you know, if you contrast this build out with the fiber buildout. So far, if you put a GPU or a TPU online, it immediately gets used. And this holds for actually very old GPUs and TPUs. Google has disclosed this about their TPUs that are 10 years old.

You can pretty easily find pricing of two generations ago GPUs on the market. And the price of these have all held. So there are no dark GPUs. No one is building data centers that aren't being fully utilized right out of the gate. You know, contrast that with the internet buildout. The internet build out was characterized by dark fiber. And so, you know, you had to lay all this brown work before you actually had any signs of demand. And so obviously there was, there was a mismatch in supply and demand. So that's the sort of overall kind of view on supply demand dynamics. I'm very, very optimistic. I think we're in the early days of figuring out really interesting applications to build on top of this technology. And, you know, in terms of model capabilities,

models are improving at a like eye-popping rate. You know, they can basically double their ability to complete long-form tasks over six to seven months. And so, you know, if you were to just arrest model development today, I think we would have the chance to build 10 to 20 years of really interesting applications on top of it. The big thing that happened in public markets so far this year is the absolute slaughtering of sort of non-AI companies

and legacy software companies of various sorts and so forth, how are you thinking about companies that don't own a model that maybe have to buy intelligence from another company or so forth? Who in your view, when you're thinking about private or public companies, who that's not an AI lab, what types of businesses survive and which ones are not going to survive if they don't own intelligence in the raw? Yeah, so I will cover both of these topics because I have pretty strong views on what's happening in software market. and then we can talk about what's what the sort of future of companies that don't own a model is. On the latter point, we can cover that one first.

Well, companies buy solutions, right? And so, you know, context is still king. In most industries, there will be companies that compound their knowledge of industry-specific workflows, industry-specific data that they attach to. And so oftentimes that won't be what the model companies choose to pursue. You know, customers buy solutions. They don't buy some discrete workflow or just a database or a system to take action. So the most important thing is industry context.

You also need a throat to choke, right? I think that's going to be increasingly important for customers. You know, so support, maintenance, integrations, data partnerships, user community, I think are all important things for companies that are building applications that are not model owners. And I think in most verticals and many of the functions in an organization, there will be independent companies that do that. Model companies are going to be an arms dealer to most industries. For some tasks or work that are highly horizontal or general,

so something like general knowledge management inside of a corporation, the model companies are probably pretty well positioned for that. But, you know, things like legal work, medical work, customer support tasks, you know, a lot of stuff that will happen in sales, accounting, finance. Like those, I think those are going to be independent vendors. and the model companies are going to be arms dealers to those. So that, I think, is the future of applications for companies that do not own models. I think it's a very bright future.

We've invested a lot in some of these leading companies. But I think both the model companies and those companies will be big and successful. On the software side, yeah. I mean, look, the software industry has been crushed in the public markets. Debate whether it's overblown or not. But I'll give you my diagnosis of the situation, you know, not particularly on evaluations of specific companies. But here's what's going to happen with the software companies in the public markets. I think the issue is not that there's going to be a ton of new software in

the future. There is going to be a ton of new software in the future. The whole story of SaaS and cloud was that the market grew 7x in size. And some of that was captured by incumbents. Some of it was captured by startups. The issue is, are the incumbents that are in the public markets going to actually be the ones that capture that? And by the way, I think it'll be much bigger than that. And by the way, I think it'll be much bigger than 7x this time. So why is it, you know, sort of a question of whether those incumbents have the chance to do it?

First of all, it's probably going to be much harder for them to grow, right? All the new budget, basically in any buyer organization is going toward AI initiatives right now. Now, it doesn't mean that they're ripping out their software systems. Gross dollar retention remains extremely high

for these incumbent software systems, and I think it will for a while. But if you look since, 2021, net dollar retention of these companies has steadily declined. If you look at the amount of revenue that the entire software industry is adding, so the growth is actually going toward AI initiatives. And so, yes, they may not be getting ripped out those incumbent vendors, but it's going to be much harder for them to find growth.

So, you know, from a product standpoint, they're not going to get torn out, but they really run the risk of value getting built on top of them. So those are going to be systems of record, but you just build new vendors, build new products that can take action on top of those. And I think that's a real risk for the software industry right now. If building software, the process of building software becomes much faster, there's another dynamic where every vendor can basically increase the amount of skews they offer rapidly,

massively. And so if you are a platform vendor of choice and you sell a bunch of software in this domain, chances are everyone in the nearby domains will also have those products and they'll all try to sell those new products and it'll get more competitive. So I think that's a dynamic that's really a risk for them. The most powerful change that I think is going to happen, which we're only seeing early signals of, is a business model shift. So when these technology shifts happen, you know, new user interface, new workflows, new data that you access, that favors, you know, newcomers, you know, over incumbents, when you pair it with a business model shift, that massively favors the newcomers, right? Because it's so hard to react if you're one of the

incumbents. And so the big shift, if you just take a spectrum of business models that's happened, you know, we used to sell license maintenance software. We moved that to, you know, seat-based subscription. then with the clouds, a lot of companies have become consumption-based pricing models. And in the future, with AI and a lot of domains, it's going to be outcome-based pricing. You can see this first in the customer support industry because there's sort of verifiable tasks that those companies have to complete. But to the extent that we actually get to the point where the predominant way that enterprises want to buy is via outcomes, and they can measure those results, I think it's going to be really tough for the incumbent. David George, Andrewson Horowitz. Thank you so much for coming on Adlaws.

That was a fascinating, learned a ton, and really appreciate you taking your time. Thanks so much, David. Great to be with you all. Thanks for having. Tracy, I thought that was a fascinating conversation. So much there. And, you know, it's fun to talk tech, but also have it be such a capital markets heavy conversation. Yeah, I kind of, the idea of software companies, their value resting in the fact that they're, like, convenient scapegoats for management.

Yeah. It's kind of funny. in a dystopian way. I remember Steve Ballmer said that years ago when he was talking about why Linux wouldn't take off? Yeah. Because he's like, no,

what are you going to just like, who are you going to get upset with your Linux goes down? Although then companies did, you know, Redhead grew up to like professionally service. I thought that was very interesting. The idea of like outcome-based pricing,

but just also some of the numbers are staggering. Staggering. Just like how big, you know, this is, I've thought about this with like, when people look at the public markets and they say, well, yeah,

it's not that overvalued because, you know, most of the public companies are making a lot of money. If you were to sort of, like, build one index of big private and public companies, you would have a lot of market cap on companies that aren't making money yet. And that would very much, like, sort of skew your view of this sort of, like, total valuation of the universe of tech companies. Absolutely. Well, the other thing I was thinking is just that question of whether or not that kind of the balance

between private and public starts to reverse because so much of the excitement right now is about AI, which, you know, is spending billions and billions and needs huge poles of capital. I guess the counterpoint to that is like there's a self-reinforcing trend in the private market, which is like if that's where all the growth is, if that's where people are making all the money, it tends to attract even more capital. So I don't know, maybe AI means that private capital grows even more and so AI can keep tapping it. There seems to just be an endless amount of money in the Middle East specifically. You know, it's like, oh,

raising more from the UAE or Saudi or whatever that just keeps these companies private. If I were at like open ad, I would be really anxious about selling it a tender. I would be like, I don't want to, I believe, I believe, I swear. I believe we're going to get there to ASI or whatever. I would be really uncomfortable about it tapping out and hedging my portfolio. Joe, you'd have to be very, very important for the founder to actually care about what you're doing with your shares. But I'm sure you would be. All right.

Shall we leave it there? Let's leave it there. Thanks for listening to this episode of the A16Z podcast. If you like this episode, be sure to like, comment, subscribe, leave us a rating or review, and share it with your friends and family. For more episodes, go to YouTube, Apple Podcast, and Spotify. Follow us on X at A16Z and subscribe to our substack at A16Z.com.

Thanks again for listening, and I'll see you in the next episode. This information is for educational purposes only and is not a recommendation to buy, hold, or sell any investment or financial product. This podcast has been produced by a third party and may include pay promotional advertisements, other company references, and individuals unaffiliated with A16Z. Such advertisements, companies, and individuals are not endorsed by AH Capital Management LLC, A16Z, or any of its affiliates. Information is from sources deemed reliable on the date of publication,

but A16Z does not guarantee its accuracy.