Bill Gurley——保持私有的馈赠与诅咒——[Invest Like the Best, EP.427]

So, Bill, this marks you retaking the crown as the most frequent Invest Like the Best guest, beating our good friend Michael Mauboussin out for the crown. Welcome back.

I can't think of anyone I'd rather be neck and neck with on the entire planet.

This is also interesting: this is the first one that you and I have done, just the 2 of us, since 2019, if you can believe that. Time flies.

Wow.

Since it's just us, I would love to go very broad and start the conversation by just talking about the state of things as you see them. I know this is something you used to do in your Benchmark days: give a state-of-the-union, markets edition. I would love for you to do that for us as we enter the summer of 2025, with everything that you're seeing out there in the world.

I'm excited to do it. I used to kick off our LP meeting with a state of VC. It's a process and a presentation I'm used to giving. I've been noticing a lot of things recently that are different in the world, and maybe permanently different about the venture capital world. A lot of the talks that I would give were based on what appeared to be inherent cyclicality in the venture business, and that has been upset, kicked over, or made a little chaotic recently, which we'll get into.

I offer 2 qualifications as we dive into this. The first would be—and Michael would appreciate this—I'm a huge fan of system-level thinking. There's a book out there called Thinking in Systems that's pretty cool. All of Michael's and my time at the Santa Fe Institute is basically tied to the theory that systems behave differently than their individual components.

Seeing across systems isn't easy. It's a difficult thing to do. But as we dive in, I think a lot of the components of the industry are bouncing into one another, and it's the aggregate effect of all those things that's super interesting. You have to step back and look at it from far away.

I also want to qualify up front that I offer no judgment on any of the participants. There are people and firms taking actions that change the state of the field, and I think they're all acting reasonably and in their best interest. The aggregate effect may not be positive for the world, but I'm not ascribing malintent or anything to anybody, and I want to clarify that up front.

If you'll allow me, I'll start to dive in. I want to walk through a handful of market realities as I see them. In the first part, I don't really want to think too much about analysis, but just highlight a bunch of things that, if you're in the VC market, are important. By the way, I think what we're going to talk about is important to VCs, founders, LPs—anyone who touches the ecosystem.

This is super high-level stuff. Let me walk through the realities, and then you and I can chat back and forth about some of the interpretations. The first thing I would bring up, which people have talked about, so I'm just putting it on the table as one of the key variables and not trying to overanalyze it, is the continued rise of the mega-VC fund.

When I first started, everything was bespoke. Most of the well-branded funds were focused on early stage. They didn't participate in late stage, and the funds were modest compared to today. Today, many of the branded firms, I think, have moved from maybe $500 million committed every 3 or 4 years to $5 billion. That's 10x.

They're participating very actively in what we would call late stage, although I've always thought late stage was a euphemism for big check. There are people willing to put $300 million in an AI company that's 12 months old. That's not late stage; it's just big checks.

There are a whole bunch of firms that have moved upmarket, and then they've also created different industry-specific funds and things like that, all leading to much more capital under management from many of the brands. There's also a ton of parties that have entered the late-stage market with different approaches, and some of those have always been there.

Fidelity and Capital Group have always done a deal or 2 every once in a while. But I think Petrides, Coatue, Altimeter, and Thrive—which I think is really doing some interesting and differentiated things in the market—are all super active. And then, oh yeah, Masa's back. We hadn't heard from him in a few years, but he's back out there in the market as well.

He's an indicator all his own.

Yeah, I agree. So there's a lot more money out there. The second reality that people talk about that's staggering—I think the phrase that's used most, which I don't love, is “zombie unicorn.”

If you look at the number of companies that were—I always like to think pre-LLM and post-LLM, because it really is a dividing moment as everyone's gotten excited about this new platform shift—there are somewhere around 1,000 private companies that have raised more than $1 billion. ChatGPT told me it was 1,250; NVCA says 900. Let's just say it's near 1,000.

Rough justice: 1,000, yeah.

It seems like they've raised somewhere between $200 million and $300 million each, and so you roll all that up and it's $300 billion. NVCA estimates that it's $3 trillion of assets on the books of the LPs.

I did some one-on-one calls with LPs. They've slowly increased their participation in venture from anywhere from 5% to 7% up to 10% to 15%, and it can be as big as half of their private equity commitment—venture capital alongside private equity. Some have private equity that's a lot bigger. But it's gotten bigger and bigger on their balance sheet, so it's important.

There are, I think, a lot of questions about this group of companies. One is: what's their correct value? A lot of the marks for their last round were set back in 2021.

2021 or something, yeah.

Yeah, which is when you had a real market peak in the second year of COVID. If you remember, all the tech stocks blew up, and Zoom blew up at that moment, and everyone did really well in that window.

So there's a question as to what they're worth. The investment world doesn't seem excited about this group of companies, just writ large. They don't have super-high growth rates, and I want to talk about why I think that is.

The thing that most people may not believe, but I guarantee you is true, is that no one has an incentive to get the marks right. For those who don't know this world, private investing, both on the PE side and the VC side, is this weird world where the GPs—the people responsible for the investments—report the price to the LPs. They get to pick it.

There are auditors in the background messing around, and you'll hear frustrated LPs because some firms will be conservative and price them low, and some will price them high, so they're getting mixed signals from different people.

Different prices on the same asset from different GPs.

Yeah. But the thing people may not realize is that the managers of the VC groups at the large endowments have no incentive to try to right-size this number. In fact, many of them are bonused on paper marks. If anything, they have the reverse incentives to get them right.

Don't the founders have the incentive to get these things right, though? Isn't it ultimately better for long-term company building that you not operate in some farcical way?

It's a great question. I think there are 2 things that fight against that. First, every founder I know has multiplied their percentage ownership by the highest price their company was ever valued at and thought about that number as their net worth.

Right. But who cares? It doesn't mean anything.

I say that without judgment. I think it's natural that you would do that. But then taking that number down by 70% or whatever is tough to do.

And then the other issue is, quite frankly, the liquidation preference. This is another technicality, so I'll explain it for the listeners. The amount of money you raise in aggregate, just the raw number, becomes your liquidation preference. In M&A outcomes, the investor can choose to take the liquidation preference and not convert to common, so they can get their money back. If a company has raised $300 million and it's worth $2 billion, the liquidation preference doesn't matter that much. If the valuation is now $400 million, then the liquidation preference could take 75% of the company in a sale. And it's a real issue out there for people.

If we go back to the list of 1,000 zombie unicorns, as you drill into that universe, how often do you find companies that are profitable, and therefore this problem can just go on forever until they choose to take another mark, versus those that are going to die at some point and would require them to raise and reset the price?

Well, I have to admit, I haven't done a statistically significant survey, which might be interesting for someone to do. Maybe there's someone at a fund of funds, someone at PitchBook, or someone at Carta who might be able to come up with that.

I'll tell you—it's a lead-in to what I think happened. We were in the middle of a very long zero-interest-rate period, which I think the acronym ZIRP is used for now, and that was unprecedented in 100 years. Zero interest rates for—what was it? It ended up being 5, 6, 7 years, something like that.

Yeah, a long time.

One, it postponed any VC correction, but it just created a ton of money and speculation. As a funny aside, I've only been to see Mr. Buffett once in my entire life. It was a group of 20 people, and it was tied to a fundraiser. We got 1 question each, and I said to Warren, “You know, your DCF doesn't work if interest rates are zero. All it does is create a lot of speculation.” And he said, “You betcha,” as you would expect. So that was it—a brush with greatness.

But anyway, there was a lot of speculation. That amount of money that I mentioned—$275 million, anywhere from $200 million to $300 million—was unprecedented. Before that window, it was unprecedented for companies to raise that much money.

When they raise that much money, a couple of things happen. I think you end up with too many participants in a single field where you would have had whittling earlier, so that makes market expansion more difficult, because there's 3 to 5 survivors instead of 1 or 2. When you overfund, you do everything. There's tons of great articles and research about how constraints lead to creativity, and you're better off choosing 1 or 2 primary product initiatives. But when you have that much money, you do 7.

You do them all.

You do them all.

Yeah.

And I think we had a mini-correction in 2022 and 2023. This was before AI blew up, and most of them ran toward breakeven, to your exact point. So when you run toward breakeven, you stop doing those 7 things and go down to the 2 things.

But those 7 things, and overfunding your sales force, led to revenue. It was not very sustainable revenue. So when you cut it back and go toward breakeven, your growth rate gets hit. It would just be natural.

So yeah, that's what I think led to low growth. I do think what you said is true: many of them had enough capital to get toward breakeven or near it, and you would think that would be a positive based on all my previous talking about the wonderful nature of traditional company building. Of course, I'm supportive of that. But there is an underlying reality, and they could perhaps exist forever, which I think leads to the zombie tag.

What's the so what of it all? No one has the incentive to take the marks. Isn't it just going to stay this way? What's going to change?

Let's come back to that. Let me move on. I want to get these market realities out there, and then we'll dive into what's possible.

Okay, great. The next one is exits, where these things are going to price in a real way.

Yeah. So we've got mega-fund zombie unicorns, and then the capital markets. There are reasons that aren't well articulated or well understood. Both the IPO and M&A markets have stalled over the past couple of years. 2021 was actually pretty good on both fronts, but things have stalled.

I think it's really important: If you look at last year, 2024, the Nasdaq was up 30%, and the window was closed. That seems to be the general belief of everyone out there. Never in my history of paying attention to the capital markets or being in venture capital do you have a successful Nasdaq market and a closed window.

And no IPOs, yeah.

Yeah. It makes no sense. That was what was correlated, and so something else is happening.

I believe a part of it—which I've been very passionate about—is the IPO discount that the banks force upon the market, especially the well-known, high-branded ones. But other people point to the cost of going public being too high. Other people point to the cost of being public being too high. And then, of course, we know money is everywhere.

So we'll come back to this in the second part, but people don't have to go public, or at least the very successful companies don't have to go public. M&A is a bit harder to unpack. Everyone blames it on Lina Khan, but she's gone, and we didn't have record M&A in the first 5 months of the year.

It's likely Magnificent 7-related. Those 7 companies are sitting on an ungodly amount of cash, and in any natural universe, that would lead to massive M&A, and I'm sure they would love to use it for that. I think they buy back stock because they can't.

But Washington's not excited about it. The EU's really not excited about them being active, and it's a stuck situation. People don't want to enter into an M&A agreement with a low certainty of close. Even the Wiz deal, which is the big outlier this year, the minute they announced it, they said it would take over a year to close. That's very difficult for a board and a management team to take on. Waiting a year is hard.

Do you think we'll see a $1 trillion private company soon?

How far is SpaceX from being there already?

It's a third of the way there. OpenAI's a third of the way there. Stripe's a tenth of the way there. There's a good handful that, if they keep their success trains going, will get there. I'm trying to make the point about the need to go public. If you can be a trillion-dollar private company, I mean, it seems crazy.

We will get to that. And then the last thing that could be hurting M&A is the overpricing. We did it in 2021, and we continue today to fund the most exciting companies to perfection. That can have an impact on M&A as well.

Maybe say one more click about why that can keep happening, in your estimation. Is it just because the feedback loop is the first set of things we talked about?

I think it was ZIRP pre-LLM. I think post-LLM, the world believes—and I think this is my fifth point or something—but the world believes AI is the biggest platform shift in anyone's lifetime.

Yeah.

And so if you believe that . . .

The other thing is, I go back to when Mauboussin and I were back at First Boston 30-something years ago. The notions of network effects and compounding effects weren't well understood, recognized, or believed. I think that's a wholesale belief right now.

People who have watched Google or Meta go from looking expensive at $12 billion to $3 trillion, if they assume something might be that, can't overpay in their mind, which I think is rational for the independent player. If everyone does it, the market is starting to price it in, but we'll see.

The next big thing we're talking about is that many LPs—not all LPs, but many LPs—have a liquidity problem, and that's a new reality. It's tied to the lack of IPOs and the lack of M&A. It's also fairly new and unique.

I found this stat: In the first quarter of 2025, U.S. colleges and universities issued $12 billion of debt, which was the third-highest quarter ever. That's an interesting reality if you're using debt to fund capital commitments, because your endowment doesn't have the liquidity it needs to pay out the 3% or 5%, or whatever it is that they traditionally paid out.

And then, very recently—you probably saw this—Harvard announced that it's in the market selling a secondary for $1 billion. They have a lot of unique things that would make them be out there. But even more interesting, Yale has announced that it's in the market looking to sell $6 billion of private equity. The fact that Yale is the one doing it is super important and super interesting from a historical perspective. I would argue no single institution has had a bigger impact on the strategy of endowment management than Yale.

Certainly not.

David Swensen is the historic—

The godfather of this model. Yeah.

No doubt. And so Yale, I think they say, had a 13% compounding return over 35 years under David Swensen. He is known for the Yale model, and the Yale model is to put a lot more money in illiquid assets than liquid assets.

The reason no one did that originally is that there's a lack of transparency, there's a lack of liquidity, they're hard to manage—all this stuff. But he did it, and it worked.

I would suggest that what we might be seeing is the exact result of everyone copying the Yale model. Howard Marks famously said, “You make a lot of money when you do something non-consensus and right.” But what if everyone copies David Swensen? What if everyone goes to 50% illiquid? Will it still work?

I think that's a provocative question, but I think that's what happened for sure.

And the fact that Yale, which led us into this strategy, is trying to get out, I find super interesting.

If you think about these LP liquidity issues, is it the thing that ultimately breaks up this big logjam that you've described?

Could be.

Okay, sorry. Can't help myself.

The AI wave came at a really interesting time. This is my fifth point, I think, out of 6 realities. We were headed to this mini-correction. You have to remember, Patrick, people were tightening their belts. These companies were laying people off. They were trying to get to break-even, and there was worry about them being able to raise money.

I would argue, over the past 30 years that I have had a window into VC, that the corrections were healthy. Every time the VC community got out over its skis, there would be a correction, and things would settle back down. I would watch Morgan and Goldman open offices on Sand Hill Road and then close offices and go away, and Fortune and Forbes pay attention to Silicon Valley and then go away. I saw that multiple times.

You never had a full correction here because AI came along and everyone got so excited. I'm not saying they shouldn't be excited. If it is the biggest platform shift in our lifetime, then you have to get excited about it. It has implications for the zombie unicorn group and everything else, if that's true.

But we've all of a sudden seen a massive amount of interest in investing. The AI companies—what do you think their valuations relative to revenue are?

Oh my God.

10X, 20X a normal company? Is that fair?

Yeah, something like that, or in some cases more.

Right. And despite the fact that the traditional LPs are tapped, they were able to go find money elsewhere. The Middle East is the area where most of that money was coming from, and over the past 12 months, how many times has a friend of yours been in the Middle East? A lot, and they're talking to fundraisers.

So they were able to find the money. The money has come in, and people are out there spending it against this opportunity. It's something no one wants to miss, and it's just an important component of everything that's going on here.

The last thing I wanted to talk about, which you already hinted at, is there's a new motion in the late-stage market. I think Josh and the team at Thrive—they're not the only ones, but they've been leading the way here—go after companies that were already on the list to go public, that The Wall Street Journal was talking about, that were going to go public next year, and present them with an offer that, I don't know if you'd want to call it too good to refuse, but something of that nature.

Founder liquidity encourages employee liquidity. It might encourage angel liquidity, and you basically encourage the company to stay private. We've seen that now a couple of times, most recently with Databricks. But Stripe's Patrick and John have gone on different podcasts and talked about it. At one point, it felt like they were saying, “Yeah, we might go public, but not now.” More recently, it sounds like what you said: “We might never go public.”

And some of the LPs that I talk to—this is fairly unusual—they've traded in and out of Stripe, and the company's somewhat comfortable with it. That's very new and unique in our world.

I think the ability for those companies to get the capital they need, either primary capital, for their employees to sell some of their stock, or for early investors to sell to later investors, functions like a by-appointment public market or something like that.

Correct. More like the old pink-sheet, trade-by-appointment thing.

Stripe is undoubtedly a great company run by incredible founders. Why take on the additional burden of extra work and scrutiny and data disclosure, and show your competitors what you're doing, and all this kind of stuff, if functionally you have your own captive private market? It seems to make sense for everyone involved, which is why I wonder if it might just keep going.

Well, it might.

And if LPs can create liquidity by selling some of their Stripe without a problem, then they don't care either. It doesn't even come up against the liquidity problem.

We're about to dive into all that. I want to mention one thing that I think is fairly interesting and intuitive once you hear it. Those investors that are encouraging that behavior, I think there's another element that's going on here.

If you think about a traditional IPO, the thing I hate maybe more than anything, as you already know, is that the banks are going to be very deliberate on their allocations. So if a company were going public, let's just say a large public and private investor, Firm X, puts in for an allocation, what do they typically do? They oversubscribe by 100X, hoping—and they may get 1% or 2% of the offering. They're not going to get 30%.

When these firms go to a company like Databricks or Stripe or whoever and encourage this round, they can get 30%. They can get a bigger ownership percentage than they would get through a traditional IPO process. In fact, they share a lot of these deals, so it's kind of an oligopolistic opportunity to hoard the public IPO growth years and take it away from the public markets.

We all know Amazon went public at less than $1 billion and then traded over $1 trillion, and the public market had access to all those great years of compounding. If you step in and delay that and get decent ownership you wouldn't get otherwise, maybe those firms are better off than they would have been participating in those companies while they were public.

Here's a super-important piece that goes on top of that. They then turn around and go to the LP community and say, “Companies are no longer going public when they used to. If you want exposure to that growth in these important high-tech companies, you have to invest in me.” That resonates.

So we've gotten through your market realities. Now I would like to poke and prod at all of them. The interesting premise to me is that I come from a place of wanting very healthy capital markets. The U.S. capital markets are an incredible thing to have happened in world history and have driven so much innovation, so my perspective is that good, healthy, functioning capital markets that price risk well are something I'm in favor of.

I'm curious where you think the system, given these market realities, is most broken from that perspective, and where you hope it changes.

I agree with you on the wish. I think that we are way better off if there are more companies. One thing I didn't bring up in the realities, which I know you know and most people know, is that the total number of public companies in the U.S. is way down from its peak, and so there are fewer companies going public.

I think a big part of that is this IPO process and the brand-name banks. I had my friend Jay Ritter rerun the data. There's upwards of 25% or 26% underpricing. You add in the 7% fee, and you're at something like a 33% cost of capital.

I know one CEO who's on the phone and talking to their bankers, and the banker said, “We think you should price at X.” The founder said, “I can raise $1 billion tomorrow at 20% above that.” To your point, why go public if the private markets are this fluid and liquid and optimized?

So I don't know what it would take. I would think deals with a capital raise are just going to get rid of that piece. There's a really interesting post by Hester Peirce. Maybe you can put it in the show notes. It's only 8 pages long. She's the longest-serving commissioner of the SEC. There are only 4 right now, and she's the one that's been the most crypto-friendly.

This post, called “A Creative and Cooperative Balancing Act,” argues that maybe blockchain is the path to fix the IPO market, which is provocative.

In what sense? To tokenize the private assets and let them trade?

Tokenize the security. No one would go back to doing a crypto allocation the way an IPO works. It would immediately be DEX-like.

Right.

And I mean, that's how ICOs already work. Yeah, so that's interesting. I'm going to watch it.

M&A is tough. The regulatory pressure is so high. We had those weird pseudo-acquisitions in the AI space with the license agreement and the hiring of the people, but we haven't seen one in a while. That was kind of a way around it.

And look, when you price things so high, you look at some of these AI rounds. I could see it making sense for Apple maybe to be in the market for something like Perplexity, but they just raised at $15 billion or whatever. It makes it hard for things to close when the prices are that high.

So I don't know. On the capital markets, it's funny you mention that statement. I think a lot of people say out loud, “We have the best-functioning capital markets in the world. They're the envy of the entire globe.” I'm less convinced, personally.

What do you think are other interesting pockets? You mentioned the Middle East and how they have been incredibly front-footed about this entire wave of technology, really pushing hard to be majorly involved in the most interesting companies, technologies, and infrastructure. Are there other innovations in capital markets that you've seen that interest you?

I don't know if you'd call it innovation, but Coatue had an announcement recently that was different. I haven't talked to Philippe about it; I'm just mentioning what I read.

They used to have a $5 million minimum for a commitment. They're taking it down to $25,000 or something, and they're going to work with an investment bank to place it. It's a similar reflection of the point I made about how they would pitch their LPs, but it's tapping into a capital pool that people sometimes refer to as dentists and doctors, who might not otherwise have access to a manager like Coatue, bringing more capital to bear.

The same thing I hear is happening in the PE world. I think one of the big PE firms is in Washington, begging to let 401(k)s invest in private companies and trying to unlock different sources of capital. It's interesting: someone pushed back on me when I was testing this theory and said, “Oh, but the U.S. institutional LPs are tapped. We'll find capital elsewhere. Look, we're finding it elsewhere.”

But that's just putting more money in the top, and I couldn't quite think of the best metaphor. You have a pipe that has input and output, and the output is stuck. I guess the human digestive system might be the best way to think about it. Just eating more food doesn't help with the constipation problem.

When you talk to LPs, obviously you don't have to name the specific people, but what are they saying to you? Are there things that they're not saying out loud—things that they're talking about more in private—that you think are important?

I think there's a heightened awareness of all the market realities that I discussed, and I think, in their place, they have to make a decision. Talk about long-term decision-making: if you're working in an endowment, you don't get much time to make a decision, and your feedback cycles are 10 or 15 years, so it's tough. But you have to start thinking about whether these things we're discussing are temporary or permanent.

If they're permanent, you have to change the way you do things. As I mentioned, one of the LPs I talked to had moved in and out of Stripe, knew the person to call who was running capital markets at the firm, and was starting to consider that it could be permanent and thinking about how they need to be positioned for a world like that.

Apollo came out, I think maybe today or yesterday, with this interesting report that, of the firms that have more than $100 million of revenue, 87%, just by count, are now private. Obviously, if you did that by market cap, it would be more skewed toward public companies, just because of the huge technology companies. But that's pretty crazy. I mean, even with that minimum of $100 million, that's a lot of revenue. So we're just living in a world that is a very private-markets-heavy world. It just seems undeniable.

Yeah. Maybe I'll change some of the ordering around here, because I had 5 analysis points, but I think that's a messier world. The comment you made about the best world being one with highly functioning capital markets, where it's efficient to go public and where things trade in liquid markets and you trade daily with low transaction costs—I do think that's the better world.

If we move to a world where the answer to getting the everyday consumer into high-growth tech is to put their endowments, 401(k)s, and IRAs into these venture funds that are charging 2 and 20, I don't think—what was that famous investing book, One Up on Wall Street? Peter Lynch—

Yeah, I know the one you're talking about.

He would never have wanted that world to be the world that emerged, but it looks like we might be headed toward that world. I just think there's more obfuscation, less transparency. There will be more fraud and higher transaction costs. It would be the nature of the beast.

When we use the example of a Stripe or whatever, that's 1 company. We might go up to 5 companies in the example, but we're worried about 1,500 companies, and they can't all do that. They can't all be Stripe.

One of the things that you taught me many, many years ago is that you have to play the game on the field and also think about where the game is going, and play for that future reality as well. But if we take the game-on-the-field approach of this messier, private-markets-heavy, liquidity reality, I'm curious what you think it means a couple of different groups should do, starting with founders, actually, going all the way down to the entrepreneurs who are actually driving all this value creation funded by these capital markets.

What is the logical thing for them to do, given the reality, especially in the AI world? If they can raise at $15 billion, maybe they should. So I'm curious how you would advise them as a group to respect this game on the field and do the optimal thing for their own success.

They are forced to play the game on the field, and this is, I think, the worst part of this whole world. There's a word that I found called a gavage tube. Do you know what a gavage tube is?

I do not.

A gavage tube is what the French use to force-feed geese so that they can create foie gras. Here's a picture of one, which is the funnel into the mouth.

What ends up happening in this world, because it's the same thing that happened in 2021, is that the minute there's a company that has any amount of excitement about it whatsoever, someone's knocking on the door trying to give them $100, $200, or $300 million. I think for founders who have struggled their whole life to raise money, this must sound like the most ridiculous comment ever, but it's a reality, and I think you know it. You know this is a reality.

What that does is it forces everyone to go all or nothing, to swing for the fences. And I lived it in the Uber–Lyft situation. But we're going to have that type of capital battle in every category under the sun.

You mentioned the notion of traditional company building. Traditional company building isn't spending $100 million or $150 million a year in cash burn, but all the big AI companies are doing that, maybe more. I think OpenAI said they're going to be at $7 billion in a year. That's not your grandfather's startup business or your grandfather's venture capital. That's a radically different world.

If you're a founder, you'd like to think the advice is, “Well, ignore all that and build your company the way you want to build it.” But if your competitor raises $300 million and is going to 10X the size of their sales force—or 50X it—you will be dead before you know it. You won't be around. So you are forced to play the game on the field.

I guess the good news is that because these investors are so eager to throw money at you, you can probably take founder liquidity. I think that's bad for the company's potential long-term success, but because it fits with their strategy, they're all encouraging it. So I guess there'd be no reason not to. If someone's going to pay 30X revenue and force you to play a game that you're not comfortable playing by burning hundreds of millions a year, you should probably take a little off the table.

I think it's bad for the ecosystem that we're going to remove all the small and middle outcomes and just play grand-slam home-run ball all day long. But that's what it feels like to me, and it feels like we didn't learn anything from the ZIRP days. All of the problems we talked about that created the zombie unicorns—we're just rerunning that. That's maybe part of not living through a correction, but we're funding these AI companies the exact way we funded those companies.

I'd love to play with 1 piece of this that I think is really important before I get back to asking the game-on-the-field question for GPs and LPs, which is really just to get your take on AI as a new general-purpose enabling technology. That is the key difference between now and 2021. Just like we've never seen rounds like this, we've also never seen revenue ramps like this in companies.

I know you're like me because you love technology. I use this stuff all day, every day. It's the most amazing technology I've ever encountered. So I would love you to riff on the—I'll call it the bull case in all of this—where no one's acting that irrationally because we really do get 5% GDP growth or whatever crazy numbers, because this just is a different class of technology, even versus, say, the internet, which was probably the last one this big.

First of all, I agree with you. I would never take the opposite side of the argument that it's not a legitimate platform shift. And if it's a platform shift, as were mobile, the internet, or the PC, that's big enough. It doesn't have to be better than those.

Yeah. It could just be another one.

So it's certainly 1 of those and might be bigger, which leads to everything that we've talked about. And as I started, I offer no judgment on any of the individual players. I think it is what it is.

There is some chance in my brain—and I haven't fully thought through all the implications of this—that some of the revenue growth is resale of compute. Many of the players in the market are reselling a wrapper on top of a foundation model on top of a hosting service, and many of them, I think people believe, are at negative gross margin.

So you might, in buying something from a wrapper company, be getting compute cheaper than you would have gotten it from the model company, who's getting it cheaper from the hosting company, and that revenue is being counted 3 or 4 times with negative gross margin. Until we get to a point where unit economics matter—and they can't matter in an all-out-war market that the gavage-tube funding creates—you have to go for market share. You have to.

I think that window is in front of us in terms of how that settles out. But I have no doubt, even if you want to step away from the foundational models, that the work Bret Taylor's doing at Sierra is real and will impact every 1 of those companies that he touches, or they touch, and will change those companies materially. I just don't have any doubt of that whatsoever.

I guess that's a long answer of saying I think so many of these things are a rational reaction to what's happening.

What about this class of technology? You've lived through and invested through lots of these technology paradigm shifts. What about this 1 gets you the most excited, especially relative to the other ones that you've lived through?

My answer to that question's decidedly personal, and it relates to what you just said.

I'm probably doing 40 or 50 searches a day on AI platforms, which is more than I ever did Google searches. It's almost all a form of very quick learning—super-quick learning—about either particulars I forgot or things I don't know about, and it's every day. And I think to myself that, for those people who are inherently self-learners, the speed at which they'll be able to get things accomplished and move up the ladder is breathtaking.

Then I think, outside of LLMs, from Tesla FSD to other types of problems that are being solved with traditional AI, those are super interesting to me as well. Maybe more profound. I do worry that LLMs have a limitation. It's potentially solvable, but they were created around language, and they're not great with numbers. And when people say, “Oh, generalized AI is just going to replace all compute,” I don't see that. They're going to have to fix some things or merge it. The way that, when you ask an AI to do math now, it goes off and writes Python, you're going to have to do more of that type of work to get to that place. If you're espousing the “But isn't it real?” argument, I can't push back on that.

Let's jump one more degree ahead to the GPs. Again, the same question about this is the game on the field: What's the rational approach? There are 2 versions of this question. One is the Spock answer, and one's the Kirk answer.

The Spock answer is just, “Yeah, this is what's available. I want to build a platform. I'm building my own company, and I want to do the rational thing to build the biggest company as an investor.” And the Kirk answer would be, if you were to restart a venture firm today, how you would approach it. Would you have a small fund like you had at Benchmark? Would you have a more go-anywhere fund with different fees so that you can play the game on the field? I'm curious about both perspectives on the GP side.

As I answer it, I want to highlight one of the last 2 things I wanted to get on the table, and that is that time is a massive problem. We're moving the time to liquidity of these companies from 5 to 7 years to 10 to 15 years. I don't know the exact number, and I think every LP is aware of that. I think I forwarded you this NVCA graph, and in there it has the percentage of committed funds that's paid back in the 5- to 10-year window of a venture fund. It used to average 20%. It's been as high as 30%, and it got down to 5% last year. It's in the 5% to 7% range, which hints at the big LP liquidity problem.

But it's a problem for GPs, too. The reason time is such a massive problem is you have the cost of capital—the IRR—that just eats away. And everyone loved to say, “Oh, it's not IRR, it's DPI.” But if time doubles, it is IRR. That's what really matters. In addition to time and the cost of capital, you have dilution. So every one of these zombie unicorns is diluting 3%, 4%, 5%, 6% a year through equity issuance to the employee base.

And when you combine those 2, it's a real problem. Let's say you were expecting to get $100 back from an investment in year 10, and you want to delay it to year 15. If you just take that 10% compounding, it now needs to be worth $160 in year 15. If you make the argument that these people invested in venture to get a big return, then your cost of capital is not 5%—that's the risk-free rate—it's 15%, and then it's 20% a year: 15% plus the 5% from the equity dilution. And now, if you wait 5 more years, guess how much money you need instead of $100? $250 for a 5-year delay, just to meet the same return expectation people had of the asset class. So that's a real problem.

It's also not clear to me—I think there were a certain number of companies that were either acquired or went public in a stage, and then entropy exists. All companies have trouble growing over the very long term. And once again, I think you're taking that window out. People love to talk about, “If you take out the big winner, what's the return of the fund?” But I haven't asked anyone the question, “Well, what if you keep the big winner but get rid of everything else?” Because that feels like where we're headed.

That's a long way of saying I really don't know the answer to your question. I spent my whole career in early stage, and I still love that time period. I think it's the time window where you can make the biggest bet and have the biggest outcome. I really hate to think about every one of the next generation of general partners having every company live through what was on the field in the Uber-Lyft situation.

Because you go into a board meeting and the other company raises another $1 billion, the thinking that you're forced to do at the table is, “Well, should we go negative gross margin for 2 more years and take market share?” You're not going to find it in a Harvard case study, I'll tell you that. It is a unique set of cards to be playing, and it's super-high-stakes poker, with strategies that you're not going to read about in Good to Great. This isn't how people who traditionally ran companies and made them great operated. All the stuff you read in every Buffett letter will not apply in that world of capital competition.

I want to talk about LPs now and the tendency, or lack thereof, for capital to seek the highest risk-adjusted return in general. It should be the case, in a rational sense, that over time the pools of capital would shift around and seek the highest risk-adjusted return. That's the whole point. So I'm curious what you think the impediments are to that just happening, and it's another way of asking: What should LPs do now?

They're the capital owners, or they represent the capital owners. Their job, ostensibly, is to get the best risk-adjusted returns relative to their own personal needs. What should they do, and what might stop them from doing it?

Well, this would probably be the last point that I'd love to drive home, and then we can just talk broadly about the situation. But you asked a provocative question at the very beginning of the podcast. Early on, you said, “Could the LP liquidity issue be a catalyst of some kind that causes this world to change?”

And there are a lot of things pressing at that. Time is a problem, as we've already talked about, and they've been putting debt in place. There's broad talk in Washington about endowment taxes, which drive more liquidity requirements for these endowments, and it's something they've never had before. You have the research cuts, not just the aggressive stuff at Harvard, but even the research cuts on the normal NIH and NSF grants. What was that extra part where they cut it from 60 to 10, the overhead or whatever?

Even that is going to cause the universities to tell their endowment, “Instead of 3%, we need 5%,” or, “We need 6% a year.” So those things are the kind of things that could push the LPs into a more difficult situation. Yale maybe being first into the secondary market looks exciting. You're a small endowment. You've never had access to Sequoia. You're going to get to buy a slice here through Yale.

But if you had a secondary pricing fallout as more and more big players come to the table, that could have a reciprocal effect on all of this stuff. And then I think the other big thing to watch is whether the Middle East might change their mind. I sent you a link that you can put in here from the chief investment director in Qatar and Sheikh Saud Salim Al-Sabah. He said, “The head of the world's largest sovereign wealth fund said the clock is ticking for private equity and joined the chorus of investors who've grown worried about the industry's valuation practices.”

That's a different perspective out of the Middle East, and if that were to get infectious and become the universal opinion instead of one of the players' opinions, that would have a big impact out there. So I think that's the area to watch.

If I were an LP, what would I do? I mean, I certainly think you dabble in the late-stage private market on both sides, as a seller and a buyer, to see what the motion looks like so that you can feel it out. I don't want to create a run on the bank, but you might really reevaluate whether the Yale model works if everyone's doing it. I think it definitely worked when Yale was the only one doing it, but I don't know that it works now.

I think it'd be interesting to find a PE firm that was going to very aggressively go through the zombie unicorn group of 1,000 companies and try and extract value. I would think there's some opportunity there to look at it optimistically instead of pessimistically. I might be interested in that as well.

If you were thinking about the strictly returns part of this equation, one of your original partners, Andy Rachleff, was very fond of saying you want to be non-consensus and right to earn the most money. Is part of the answer perhaps looking to invest in private markets away from AI, where the pricing and supply-demand story is extremely different?

If you just go to a more run-of-the-mill company, the capital markets are not super excited to fund them, and they're evaluating them in very strict calculator terms to a degree that's nothing like what they're doing in AI. Is that a place to go spend more time?

I even think some of the names I've already mentioned that are considered to be these late-stage investors are thinking this way. They're thinking, “What if I can find a traditional company that may not understand that AI would enhance it, but where we can go do that ourselves?” And maybe that is a disruptive way of looking at things.

That non-consensus, right quadrant—the first time I read it was Howard Marks, who I read everything I can that Howard writes. There is, I think, an incongruence between that point of view and these platform shifts, because these platform shifts have now become consensus. You'd have to not invest in AI, which sounds outlandish. So I don't know that you can apply those 2 things simultaneously.

One thing that's super interesting about AI, to your point, is that the big companies seem to have moved very quickly.

I mean, if you go on ServiceNow’s website, it just drips with AI. Microsoft’s earnings transcript had 67 occurrences of AI, and Satya just talked for 2 hours about AI. It’s a weird thing. I think a lot of what we read in Crossing the Chasm or The Innovator’s Dilemma—where the big companies are supposed to be slow to mobile, slow to the internet—that’s an opportunity for the startups. This is an interesting one where I think a lot of the big companies paid attention early.

Do you think that’s just happening now, though, in a different form? Maybe the example would be that, in theory, Google should have been in the very best position to dominate every AI use case, and yet basically no one I know is using Gemini or Google to do code generation, their daily-driver LLM work, or frankly much of anything else. They’re using startups—Cursor, Anthropic, OpenAI. Even though they are moving fast, the technology companies themselves are just redemonstrating the same phenomenon again.

I think there are data points on both sides. I think that’s an interesting argument. Apple is an interesting argument. Microsoft, having missed one and survived, puts them in a better position to be alert about the next one.

I saw an interesting interview between David Friedberg and Sundar, where he asked him if he’d ever read The Innovator’s Dilemma, and he admitted that he had not. When your company’s crushing it, those kinds of things are, “Eh, those were for somebody else.” But yeah, maybe now I should read it.

Evaluating an exciting new AI company where the revenue is of a different nature than enterprise SaaS was, how would you go about assessing the quality of revenue in a new AI startup today as an investor?

I think it’s tough for the reason I mentioned before. You might be getting a million-dollar deal, and it has negative gross margin for you. But on the flip side, any AI model that’s 2 generations old sells for 1/100th the price per token as today’s models. I think you could probably have confidence that you’re going to optimize pricing later.

One of the interesting things that partners at Benchmark have been looking at and assessing is when companies move to optimization mode, and how they make decisions differently once they do that than when they’re in experimentation and sandbox mode. With the amount of capital you have, you can run sandbox mode longer before you go to optimization mode.

We saw this on the internet. I like to highlight that the first 2 years, everyone built on Sun and Oracle. All the startups did, and 5 or 6 years in, no one did. That’s why it’s so important to pay attention to that shift.

What do you think about the interesting international competitive dynamics in AI, which existed to a lesser extent in some of these other platform shifts, where it was mostly United States and Western technology at the forefront? China’s the obvious big question here around DeepSeek and things like it, but now also startups coming out of China that are offering what look like incredibly impressive products. How do you process the international, especially China-versus-U.S., component of this race?

I think there’s a super interesting development in the China situation that will be very, very fascinating to watch, and that is when DeepSeek hit and took off. We were all focused on how the U.S. reacted, the U.S. models, Washington, and the fact that AWS hosted DeepSeek or whatever.

What happened in China, however, is Alibaba made Qwen open source. Xiaomi has a model out now. I forget the name of it—MiMo, I think, maybe. It’s open source. And Robin Li of Baidu had his model proprietary, and he said in June it’s going to be open source.

That level of competition, if it leads to 4 deep-pocketed, all-open products, is going to be ultra-powerful. We’ve already learned that these models can train each other and help each other get better. So if you have 4 open ones that can all train on each other, and everybody can get ahold of that, I think that’s going to lead to a massive amount of optionality and experimentation that we’re not going to have here. That’s the most fascinating piece of the international AI narrative that I’ve seen.

How much do you find yourself having allegiances where you want a certain group to win versus another? What are you most rooting for, I guess, in the whole thing? Competition?

It’s funny you bring that up. I was noticing that some of the people who are the biggest China hawks are the ones who have bet on the new VC-backed military companies, and I just hate that you might become a warmonger. But I know that’s possible because when I was an investor in Uber, you defend it at all costs. That’s natural. It’s like your child. You’re looking out for it, and so your allegiances are going to go where your investments are. I still feel that for any company that’s got Benchmark attached to it. I don’t know that I’ll ever not feel that way.

So that is what it is. That’s, I think, a lot of how this world works. In terms of just the technologies, I think some of the non-LLM stuff is super exciting. I can’t wait to see what’s possible with robotic intelligence. I’d love to see us make gains in the healthcare space.

I don’t think all disease will be gone in 10 years like the AI founders are saying. I think that’s a ridiculous thing to say out loud, but it’ll be fun to watch it all. As you said, I’m using this stuff every day. The pace of change is the fastest I’ve ever seen in my entire career. If you miss a week of news, it’s like a different world a week later.

You brought up the defense startup ecosystem. I would extend that to say the physical-world hard-technology ecosystem, lots of which has nothing to do with war—mining companies or whatever. How do you think about this category of company? They’re undoubtedly technology companies. They’re often operating in very large markets, but they have a very different capital-intensity profile, typically. They require tons and tons of capital to get to revenue over long periods of time. There’s all the nuclear stuff, fusion and fission. How do you think about that kind of private-markets technology investing? I know you didn’t do a lot of it, which makes me think maybe you don’t love it.

As a rule of thumb, if I were a professor, I would say you could study it mathematically, and it hasn’t been a great place for returns. You could look at the massive amount of venture capital put into solar 15 or 20 years ago, and it didn’t work.

There’s 1 exception to this whole rule, and it’s anything Elon Musk touches. So SpaceX and Tesla are data points, but they’re outliers, really, and they’re both attached to Elon. I think we’ll have to see 4 or 5 of those from non-Elons to know if it’s possible.

What I’ve learned and heard and studied about his execution prowess and his speed, that he develops inside of these companies—I don’t know that others can handle that or are capable of it. It’d be great for the world if they are and they’re successful.

And by the way, we’ve seen waves of capital availability lead to more interest in businesses that are less capital-efficient. There’s a correlation there. So the other thing to watch is, if that got tighter, would the appetite still be there?

Many of those businesses involve regulation, and mining gets better if we’re not allowed to use China. I hate that part of this world. But I gave a speech on regulatory capture a few years ago, and no one from Silicon Valley was in Washington at the time. Now they’re all there—Hill and Valley, everything else. So that’s another element. Maybe it should have been on my reality wheel.

It seems like that’s just going to happen—that venture-backed, early-private-markets-backed companies that deal with big regulated industries are just going to get processed. I’m curious if something like Anduril, which is maybe a $30 billion valuation, something like that—not SpaceX level, but big—is another data point in your mind toward, “Okay, we actually can execute in companies that require lots of capital.”

Unquestionably, that’s true from a regulatory standpoint. I think it had historically been very difficult for companies to break through in these industries solely because of regulation.

Prior to Tesla, there were, like, 7 Motors. There were other attempts at building cars, and none of them worked. I think a lot of them got stuck along the way from a regulatory standpoint more than anything.

So Anduril being cleared by the DoD and actively selling into our military is certainly a new data point and a very impressive one for a startup to have achieved, no doubt. I don’t know that that means every VC under the sun should jump into this stuff. It’s hard.

If you can make a software company—or, as people like to say, a social-network company—spring to life and generate revenue growth at high margins, boy, that’s a much easier path to riches than what we’re talking about now.

Are there any other pockets of the ecosystem right now that we haven’t talked about that you are especially interested in—company types, investing strategies, or dynamics?

If I were still an active GP—and I think a few people are doing this—but if I were an active GP, I would be thinking about verticals in AI and thinking about where AI is exceptionally great. It’s exceptionally great in language. Coding is a tighter form of language, so it’s even better at coding.

There are areas where that matters. A lot of this has played out: it matters in legal, it matters in customer support. But there are probably other pockets that have yet to be fully explored. That fit is super interesting to me.

Back to our original set of realities that you explored from the LP side and just the capital-markets-systems-level stuff, what do you think is going to happen? You’ve laid out the realities on the field and the various incentives, or lack thereof, for change.

What do you think is going to happen in the next, let’s say, 5 years across the domain?

My gut is that we have a problem. I’ve always been more of an analyst than an optimist, even though I was successful in venture, so I had to be somewhat of an optimist. But I started as a security analyst, and I’m born with more of a critical-thinking hat, so my bias would be that way. Someone could certainly take the other side and say, “Oh, Gurley’s always predicting the next downturn or whatever.” But my gut is that we have a problem.

The system as it exists today promotes less liquidity, less traditional high-quality company building, and way higher burn rates. That’s just not a great combination from my perspective, and it’s all self-reinforcing. So all of the components that I listed, unless something happens at the LP level, I don’t see a corrective mechanism. I think we’re getting sucked more and more into that loop. There’s a great video you may have seen where Josh Kopelman just walks through some simple GP math from his perspective.

Yeah, with Jack Altman. Yeah, I saw it.

It’s—what is it, 3 minutes long? Maybe we’ll put a post in there. But I have a hard time disagreeing with what he did there. Pretty simple math. This isn’t in a place where it’s going to work. From the prices we’re paying, the amount of money we’re spending, and what you would need to have happen for VC returns to match what they’ve done historically, it seems like a tough situation to puzzle out from my perspective.

What would happen on the other side of a reset? Let’s just imagine a simulation where we can bring public-equity pricing scrutiny or mechanisms to every available asset and get a big pricing reset as a result. Then what? What are the pros and cons if we have a bad moment that we have to go through? What are the good and bad things that happen on the other side of that reset, do you think?

I have a hard time thinking about it. I think most people would consider it awful, having lived through a couple of these resets. I did find—this is maybe a bit humorous—that as an acting GP, I was much calmer and happier, and found my job more fulfilling, efficient, and productive in the resets than in the manias. Certain other actors may prefer the manias, like maybe someone with a sales DNA who likes to be out there amongst it all.

I just found the conversations about traditional company building and all this were more efficient and authentic in those winters. The pretenders left town. When the internet bubble burst, there were consumer companies called B2C, and enterprise was B2B. There was a joke that it became “back to consulting” and “back to banking,” because people left Silicon Valley when the money wasn’t easy.

I don’t love the opportunists. I don’t think they’re in it for the right reasons, and they tend to over-promote, over-raise capital, over-participate in secondaries, and leave situations that can crash and burn. I don’t love that. It’s a part of the world when you’re dealing with it at this speed.

So if it corrected, people would look for opportunities. One of the things that causes this situation is, I think, that everyone has studied history, everyone knows about compounding effects, everyone knows about network effects, everyone’s studied cycles, and they’ve seen boom and bust. Do you remember how long the stock market was down during the original COVID correction?

3 weeks or something?

Yeah. And then people started buying the opposite side. So I suspect that the conviction in AI is high enough that even if we were to have a 6-month period where people thought AI was overstated, it would start rebounding very quickly.

If you were starting a brand-new investment firm today, what do you think would be the most important components of brand-building for that firm? We’re now in the era where some of the upstart private-market firms, like Thrive, Greenoaks, Andreessen Horowitz, and Ribbit—the ones that were started around 2010 and a few years later—are incredibly big, well-respected brands, and they had their own way of doing that. We’re in a new era. What advice would you give to upstart investors who are starting their firm this year and want to be those firms 12 years from now?

It’s funny. You just provoked something in my brain that is not the question you asked, but—

That’s fine.

Another negative element of all the systemic issues that we’re talking about is that the firms that buy their way onto the cap table by writing a $300 million check or whatever differentiate themselves by being the best friend of the founder they possibly can be. This is a lot easier for me to say since I’m not writing checks, and if anyone doesn’t like what I say, they’re never going to compete with me to be a new board member, so it’s fine.

But they don’t take responsibility for being someone who helps you make better decisions. They’re never going to tell you no. A grand example of that is the SBF/FTX situation, where no one took a board seat, everyone believed that he wasn’t commingling, and it ended poorly. It’s helpful to have someone along who will call bullshit when it needs to be done, who will push on unit economics. One fear I have about this world is that there’s less and less of that.

The very best CEOs—and I put people like Barton and Benioff in there—even Mark at Meta has said this: They believe being public makes them run sharper. One other negative of these companies staying private forever is that they’re not getting that feedback.

Now let me try to answer your question. I don’t know, man. What would I do? It’s a hard time for me to imagine starting on that journey because of everything that I just laid out, so I really don’t know. I’m just going to have to take a pass on answering that question.

Well, I’m glad I sparked the other thought with the question. It was productive one way or the other. Maybe in closing, you could leave people with a few thoughts for founders specifically. I always try to come back to them because without them doing their thing, basically none of this matters at all.

That’s absolutely true.

Founders are facing maybe the best setup of all time, with more tools to build companies, the most exciting new enabling technology maybe we’ve ever seen, and lots of capital that’s willing to fund that journey. Any closing thoughts? You’ve seen so many founders build some amazing, big businesses across your career and backed a lot of them directly. Closing thoughts for them on the opportunity in front of them and how they should think about it?

Yeah. If you’re lucky enough to be in one of these hot companies, and you’re in the middle of this whole world that we talked about, I would offer a couple of things. One, unit economics will matter one day, and that doesn’t mean you have to sharpen a pencil right now. Like I said, two-generation-old models cost 1/100th per token. You can plan that we’re going to move to that. It’s fine to have a burn rate, but the unit economics will eventually matter.

You’ll eventually have to scale the company up and operate in an efficient, productive way. When you have these all-out battles, you can get lost. I’ve found a lot of founders think about certain elements of operating prowess as red tape. They think, “Oh, that’s what big companies do. That’s bureaucratic. That’s not why we’re here.” But as you grow and get over $100 million and get to $1 billion in revenue, you just can’t operate that way. This is advice in any cycle, really, but it gets accentuated here because of the amount of capital.

One of my favorite pieces that Reid Hoffman ever wrote was about Uber, using this pirate-navy metaphor. He said all startups are born as pirates, but they eventually have to become a navy, and that’s true. I think it’s an uncomfortable transition for some, but you have to figure out a way to do it.

Another thing that comes to my mind related to that: There was also a great blog post that I think Ben Horowitz wrote about this. He said, “We only want to back founders that go all the way.” It was a genius thing to write because that’s what founders want to hear. It’s actually true of every venture capitalist in the world because you’ve got a 50/50 shot when you change the CEO, and why would you put that risk on your portfolio?

But deep in that blog post, there are 2 or 3 paragraphs that say, of course, the founder has to want to learn how to lead. I think one piece that we miss so much in this industry is there’s nothing about a founder, in the way they’re born, that makes them capable of leading a 1,000-person organization. There are people who have studied what it takes to be good at that their whole lives.

There’s a handful of founders, maybe 30, who got to work with Bill Campbell to help coach them on what it means to do that, but it doesn’t come natively. It’s not for free, and you have to want to do that. There’s a personality type where that’s very hard. I’ve had great conversations with Michael Dell. He did it, thought for a while that he didn’t want to do it, and ended up finding a way where he could do it and be happy. That part’s hard.

I’d say 2 things. One, network effects are real, and they can be more real if you focus on them. I think if you’re in a hot market and growth is everywhere, it’s easy not to think about them. But are there elements of your business where the data exhaust, or whatever—let’s just say you have 1,000 customers and then you get to 2,000 customers—means that the 2,000th customer should have a way better experience than the 1,000th? How do you design that into your system? If you do, if you can figure that out and design it in, it’ll have powerful long-term implications for the success of your company.

What do you think the way to do that is in the AI era? Is that mostly a data question—that you just want your product to naturally produce more data that then improves the product?

Let’s say you’re serving a functional vertical.

If the learning of an individual customer becomes learning for the whole group and everyone benefits, that's pretty powerful, and I think that's very doable. There are some AI companies in the legal space. I'm not involved in any of them, but they're studying all the intake information that you would put into a lawsuit, and they're also studying all the precedent and legal case history. The AI is going to do X.

If you've got a human in the loop, you're going to notice the failure points, and then you're going to improve the model. Those kinds of things could lead to someone that has an early lead having an even bigger lead in the long term if there's constant improvement of the model.

Do you think that AI reopens consumer as an interesting place to invest? It seems like it hasn't been since the mobile era, when a ton of amazing consumer businesses were built, that we really haven't had a lot of VCs focused on that or capital going to it, and that this might reopen that very lucrative part of the world.

There are people that have studied some of the stuff popping up in China that hints at this, and I should probably do more work on that and be more knowledgeable than I am. I would also say we had an early shot on goal with Character, which traded.

But there are 2 elements of the first wave of LLMs that I think make them not perfect for consumer. One is voice being really good, which they're getting better at, and the second one being memory. They're starting to get better at that. I think they're doing it outside of the core LLM, but it doesn't matter, and then they put it in the context window.

As those things improve, a lot of the negative feedback on the early Character.AI was that they didn't really get to know you, and so the efficiency, or the network effect and learning impact and switching costs, weren't there. It's not that difficult for me to imagine exactly what played out in Her, which was an unbelievable movie for its time, by the way. I think that will happen. I would be shocked if you don't see 4 or 5 companies pop up in the next year.

And maybe that's the contrarian thing that we were looking for that we wandered onto, or you wandered onto, which is that most of the work in the U.S. has been on the enterprise side, and there probably are some real opportunities on the consumer side.

Bill, it's so much fun to do this with you. Maybe every couple of years we'll do a state-of-the-markets update, since you're not doing them directly for LPs anymore, and we'll do it for the whole universe. Thanks for doing this with me and for all that you've learned and shared with us all.

Yeah. And I know the LP community pays attention to everything that you do, Patrick, so I would encourage anyone that has feedback for me, wants to correct something that I said, or anything, to reach out. I'd love to engage and learn more. I find the industry fascinating, and I'm hopeful that anything I share is useful. I'm at the give-back part of my career, and I'd love to be helpful.

Careful what you wish for. We'll put the bat signal out there, Bill. Thanks so much for your time.

All right. Bye.