20VC：OpenAI以30亿美元收购Windsurf：拆解 | 捐赠基金是否正在失灵，以及2025年LP向风险投资配置将如何变化 | 为什么收入倍数是胡扯、AI Rollup崛起与多阶段基金正在摧毁种子投资

Guys, thank you so much for joining me again. I was inundated with messages from my LPs, and from LPs that I don't have, asking to invest in funds, so thank you so much for that and for saying how much you loved our prior conversation. Both of you, thank you for joining me again.

I'm not here to help you, dude. I might just crash out now. But okay, we'll go at it for now.

Guys, I'm going to kick off with the topics I think we have to discuss first: Windsurf's $3 billion acquisition. It broke news. How do we think about this? How did you guys read it?

1. OpenAI Makes Its Coding Bet

The first thing you do as a venture investor is go, "Bummer, you're not in it." Let's be honest. It's like an IPO prospectus: the first thing you do is look at the ownership page and go, "Damn," right? So that's the first response. You have to get over that. It takes a little therapy, and then you go, "What does it all mean?"

First of all, it hasn't happened yet, as I understand it. It's been rumored, but TBD. But it makes sense. If you're OpenAI, and you're looking at what you're doing here, there are 2 or 3 mondo-huge use cases for your product, and you probably want to be relevant in those cases. One of them, obviously one of the most visible, is coding, and picking up something in the space.

Just for 1% of your market cap, it's probably a sensible bet. So that was the big-picture response. Now, you can get into whether they should have bought Cursor, whether they tried to buy Cursor, and whether there are a whole bunch of others. Then we can come back to the "Is it defensible?" argument, which I'll come back to in a second.

But, zooming out a million miles, it totally makes sense. When you're running a $300 billion market-cap company that's predicated on a bunch of different end-use cases, getting closer ownership of one of those big use cases makes sense. Jason.

I'm not convinced this deal will happen. I think Windsurf is pretty epic. I think Varun's a great CEO, but we don't know, right? We don't know if it'll happen. Maybe, Harry, you're closer to the pulse of anything on planet Earth, so you may know, but I'm honestly not sure the deal will happen.

I do think the learning that 1% is really interesting. I'll tell you what I learned being a VP at a big tech company. There were different levels of deals. 10% is "bet the farm." That's like Adobe buying Figma. That wasn't the end of Adobe, but it's a big deal. You don't lose your job, but it's betting the farm.

Instagram was 10% of Facebook. WhatsApp—what's the magic number? 10% of Facebook's value. This is not just an SVP deal. This is the CEO saying, "I'm betting the farm." That's 10%.

1% is like an SVP deal. This is an SVP saying, "I'm betting my BU." So my guess is whoever is at OpenAI—and where is OpenAI slightly weak compared to Anthropic? It's in coding. It's the one area where it's weak.

OpenAI—ChatGPT has pulled away. You cannot catch it. No matter what anybody says, you will never catch that revenue, and Anthropic's even given up there, right? But they own this coding—not just the developer, but the coding. So 1% of your market cap to catch up, you won't even notice 1%, right?

Now, the VP may get fired if it doesn't work out. It's possible, but you really have to see: when companies stall, M&A bets are very different. When they're on the growth path, these 10% and 1% thresholds are the bets you have to make. The clock's ticking, right?

My analysis was that you either had to cede the market entirely or make this acquisition. They had to catch up with Anthropic, which is so far ahead in terms of that developer community and advocacy, so they had to.

I think, bluntly, they couldn't buy Cursor. Cursor had just raised a new round; it was too expensive, and I don't think they would have sold.

Sam Altman doesn't have any shares. I don't know whether he's the founder or not. It's a little confusing, but he can spend 10%. He can spend $30 billion. $30 billion, interestingly, is 10%, going to Rory's point. That's 3X the last round.

I guarantee the VCs at the $10 billion round will take a quick 3X.

But would you buy it at $30 billion if it's not as defensible as people think?

I'm saying that he could. I don't believe it's unbuyable. I do not believe that, for $30 billion, it was unpurchasable. It's possible.

We've all seen deals. Rory can share a few crazy stories of folks that turned away $8 billion deals, like the Wiz deal. But even Wiz, it was just a game in the end, wasn't it? It was just a game to hit the number.

The zoom-out thing here is this: we all operate on 1 order of magnitude, and it's very hard to imagine what life is like 2 orders of magnitude further up. But people exist—companies exist—2 orders of magnitude further up.

$3 billion is a home-run venture deal. But these guys have a $300 billion market cap. They're playing for one of the 3 or 4 companies on the planet that have a $2 trillion market cap. So you do what it takes to make that happen, right?

There's a period in every market where it's exploding. You really don't know how things are going to end up.

It's actually very smart and savvy to make some bets just in case it turns out that way. Picking 2 that are very opposite in terms of outcome, no one even remembers to the nearest basis point how much dilution Microsoft took to buy the elements that became Microsoft Office. I remember they bought something that was like PowerPoint. I think they bought a word processor. I was around then, but it was the early '80s.

Tens of millions, I think, for PowerPoint.

Tens of millions in the noise.

Tens of millions, yeah.

Who cares? Got it done. Fast-forward to an example that didn't work: Excite@Home. It was a search company, public, and the big idea was, “Oh my God, you've got to combine search and literally the underlying piping. Let's combine with @Home, the cable infrastructure company.” It turned out to be a deal as dumb as rocks.

But what you recognize at the point in time when everything's happening so fast and nobody really knows is you're probably better off making some bets. If you think about this particular bet, if you just look at the scope of the chatter—and I don't love chatter—the evolution of the story went from “The models are everything. The AI, all these apps are just AI wrappers,” which was conventional wisdom a year ago, to “Oh my God, models are commodities. AI apps are all it's going to be,” to the third iteration now: “Oh my God, the models have such market caps that they can buy the apps.” What you recognize here is that no one knows anything, right?

The thing I admire about Altman is his bias to action. You sit there and go, “There are 2 or 3 massive use cases for AI. One of them is direct chat. Tick, done. The second one is coding. Hmm. The third one is customer success. TBD, later.” You're just moving down the to-do list. So I think doing something makes sense here, and you can't unsee it now, right?

Even if this deal doesn't get done, to your point on Wiz, the mighty corporate intent has been stated: We need to own one of those things. If they don't do Windsurf now, the line of other coding apps outside the OpenAI office is going to go around the freaking block, and at some point someone is going to buy something. Will it work? Again, who knows. Will it be PowerPoint or will it be Excite@Home? That's, as they say, why they play the game. But not making a move is akin to losing.

Another thing that maybe people miss a little bit on M&A is that if OpenAI does buy Windsurf, you don't have to do the brutal Salesforce-Oracle strategy and say, “It's us or nothing.” You can build a platform and let the customers decide. You really can. You can say, “Listen, this is now Windsurf powered by OpenAI. We're going to put 1,000 engineers on this, but if you want to use Cursor, if you want to use Lovable, if you want to use any other system, we still love you.”

You don't have to overfavor your platform. I think OpenAI would handle it very well. They can do both. A lot of companies that do this have different teams, and they let the market decide. It doesn't have to be ruthless if you have a platform. You can actually seed them both. It does happen.

The best example of that is obviously Microsoft, where they had the operating system and a dominant set of apps, but there were other apps out there. Jason, one thing I would disagree with is what you say about not having to be ruthless. You don't have to be ruthless in the short term, but one of the things you see is that, in the end, there is a grinding-you-down element to it.

After 20 years of the PC wars, there were really only 2 or 3 companies at scale selling personal productivity apps independent of Microsoft. You had Adobe, you had Quicken, and maybe some of the security apps. So I agree with you: if they buy this and favor it, everyone else will have to use their shit because it's one of the 2 things out there. But it'll be a long 10-year grind if you're the independent and they can make it better and better.

2. Multistage Crushes Seed

As a venture nerd, my takeaway was going through the cap table and seeing who did the first rounds. What do you see? You see Neal Mehta and Greenoaks. It brought me to something that I was just talking to a massive LP about today.

They said, “Harry, tell me a seed manager in San Francisco to back. Give me a seed manager.” And I said, “I wouldn't touch it. I wouldn't touch it.” The multi-stage fund product at seed is so good, so efficient, and its cost of capital is so low that it's just crushing everyone at seed. For me, this is another example: Greenoaks leading the seed and doubling down on the Series A. All of the multi-stage firms are pushing out the seed firms more than ever, and I think this is a great example of it.

How big is that fund, Harry? The Greenoaks fund?

Between $1.5 billion and $3 billion, and I know it's a big range.

So if they own 10% with dilution, going to your point last time about Insight and Wiz, even if they own 15%, how much of the fund does it return?

It reportedly returned between $500 million and $600 million, which is still less than a third, even on the smaller funds.

Man, venture's brutal.

Yeah.

Less than a third. So you need to do a 5x fund. How many Windsurfs do you need? Help me do the math.

Just 15.

15 per fund?

Yeah.

Still, it's been my experience that you'll cash the check for the first $500 million and smile. I always say that to people when they say, “Oh, it's only a 3x,” or, “It's only a 5x.” I've been doing this for 30 years. Everyone cashes the check, right?

Going back to the Greenoaks comment, I'm not sure it says something systemic about seed versus non-seed. I just think it says extraordinarily good picking from a very connected investor, most of whose investments are working really well at a much higher level. To go from later-stage deals to reach down into the early stage and unerringly pick a winner like that, all credit to him. I just have to say, great success. Go team. The only thing that counts is winning, and no one gives a damn how you do it. Well done.

I totally agree. Jason, anything to add there?

I'm with you. I don't have anything profound to add. I do think your point, though, about the squeezed white space for seed funds is a good one. If nothing else, it pushes ownership down, right? Going to Rory's point last time about risk, I think it increases the risk for seed funds.

When this multi-stage thing starts to get perfected, like on a Windsurf, if you're scraping for 2% to 3% as a seed manager instead of 12%, and your fund has doubled in size, the outcomes have to be doubled. It's a compounding set of risk pressures on seed. Can a $1 billion outcome even return the fund anymore for a seed manager? That's the old line for seed: a $1 billion outcome can return the fund.

I don't think it's true of a lot of seed managers anymore. I don't think a $1 billion outcome can return the fund, let alone do 3x net. I don't even think it can return the fund, and that's a big, big challenge for seed if a unicorn can't return the fund.

Funny, I was reflecting on our conversation last week and, Jason, your comments on seed, and I was laughing about it. Your quote on seed was, “It's a circus.” One thing I've internalized is that everyone is looking at everyone else's spot and going, “My spot is hard. My God, theirs looks easy.” Right? And we all do it.

The dirty little secret is that it's hard everywhere. There's a lot of capital and a small number of opportunities. It's just really hard to make money because of the amount of capital in the business. I don't know if structurally seed is risk-adjusted less attractive than A or B, where we play, or than late stage, where Greenoaks normally would play, or whatever. I just think every stage is wrestling.

The proof of that is that everyone is drifting into the other stages and saying, “Oh my God, I need to do that to do my thing.” It's just a very messed-up world at the moment, and no one's staying in their swim lanes. I think a lot of it is the super-big funds doing everything. Full-stack providers make everyone question what they're doing, and we'll see over time.

The one thing that I really see, though, being at seed and Series A, is that there's a lot more dilution sensitivity at seed. Very often today, the whole round is 10%, and we're able to do 7.5% with 2.5% for angels, whereas before it was 15%, with 12.5% and 2.5%. I've really seen that compression from 15% to 10% on the seed rounds.

You can afford to be dilution-sensitive as an entrepreneur if you can get cheap capital, right? You can only be as sensitive as the other side will let you. So dilution-sensitive is another way of saying, “Dude, I have 3 more people lining up down the street to give me a better term sheet, so you're only taking 10%, and if you don't like it, shove it. I've got more money.”

It's a lot easier to get better ownership when there's just less capital. Barton Biggs used to have this saying: “There's no business so good that excess capital can't ruin it.” And here we are.

3. Growth Must Justify Valuations

Speaking of excess capital, I was with the team today.

We literally just came out of an investment meeting, and the company we were looking at is doing $7 million in revenue. It's been done at $700 million by some of the big funds. And I said, “Wow, we’re back, huh? Return of the 100X.”

Well, what’s the forward multiple, though? I think 100X was always a misnomer. Maybe Rory would disagree. I think when we look at forward multiples, it’s a better way to think about this, right?

The forward multiple’s about 33X.

Yeah, which means it’s 3Xing. So let’s start with that. First of all, let’s start with the very basic. I’m going to say something here: you’re a hypocrite. I’ll tell you why I say that.

Me, not Harry. I’m the hypocrite. I’m deflecting.

No, no, Harry, because you pay the highest investment multiple. The truth is this: the entire venture business starts off with an infinite revenue multiple and gradually comes down, and multiples go down over time as growth rates accelerate. What you’re trying to do is hope to God that the growth rate stays higher long enough to de-risk the multiple before you intersect the public markets. In the end, everything trades at 5 or 6 times revenue if you’re growing at 20%.

You’ve done deals. Your last seed deal, if it’s doing $500,000 and you’re paying 33X, 50X, whatever it is, it’s a 50X, 100X. So the real question is... That’s the first thing: it’s stage-dependent. The second thing is that it’s growth-dependent. Jason asked the right question. My partner Andy says Lilly will refuse to have a conversation about revenue multiple unless you also state the growth rate. He’s like, “It’s an incomplete equation not worthy of discussion.”

So when you say 100X, 100X growing at 3 or 4 times with high growth persistence, which is a term we coined for being likely to stay growing at that rate, in 2 years you’re out of the risk zone. 100X, if that growth rate is sub 2% and starts to decline, you’re so screwed your head’ll hurt. It’s situation-dependent. So that’s what you have to say.

Going back to ’21, I think what happened in ’21 was that a lot of people paid up for growth rates at 100X and then didn’t get the growth, and that’s a fiasco. This time, it boils down to what people are paying up for: growth again. Will they get the growth? Will that $7 million become $20 or $30 million? If so, and you get just 1 more good year of growth, you’re at 60 or at 10 times. It’s a scary way to live 2 years of your life, but it’s not impossible. If it slows down, you’re screwed.

Totally agree there, and I think it’s completely understandable to need the growth multiple alongside it. My challenge goes back to our point earlier, though, which is understanding sustainability, the transience of product-market fit, and the transience of revenue.

No, you’re exactly right. If you lean in and it goes away, that’s why those 10 years of SaaS were such a good business: because it was predictable, because the input of sales and marketing to the output of revenue was predictable, and because the revenue was sticky. In retrospect, those were the golden years of just applying capital and growing into the multiple, right? And if it’s not like that, our younger partners don’t say it, but you can see it in their eyes. They’re saying to me, “You idiots. You made money when it was easy. Don’t give me shit now. It’s hard today, brother.” They’re right. I don’t like to admit it, but they’re right.

Today, when you’re predicting which of these companies can keep up that growth rate, there’s much more variety.

4. AI Compresses The PMF Journey

When I look at Klaviyo, when I look at UiPath, when I look at ServiceTitan, the list goes on and on of companies that actually took a long time to get to $1 million in ARR. They really went through the idea maze and product maze to get to a good number, being at $1 million in ARR. My question is, are we in an entirely new world where today, from day 1, you’re at the starting line and you don’t have the 5 years to weave and snake, or are we still in the same world?

Well, weren’t they all... We could go through a history. Weren’t Anysphere, Codium, whatever Bolt was before—didn’t these all struggle for a year or 2 before they took off? I mean, I know Cursor almost died, right? Bolt almost died, right? And then finally it hit for Cursor. I don’t know the whole story of Windsurf. Even Windsurf was Codeium before it, right? It wasn’t even the same thing.

Everyone talks about Windsurf. 90 days ago, Windsurf barely existed. 90 days ago, it was a Chrome plugin called Codium. Now it’s taking down the market leader.

I think of it as this, man: the walk-in-the-woods period is indeterminate. It can be a year, it can be 6 months, it can be 5 years. It doesn’t matter. It’s up to them, right? You’re financing that journey at seed, and as long as they don’t run out of money and they want to keep doing it, fine.

Once you lock in, the interesting thing is that the trajectories now are different. The SaaS trajectories lock in, and Jason knows it so well: triple, triple, double, double, and that steady thing. The weird thing now is that once you lock in, as you say, you go from “This company’s not going to make it” to “Oh my God, I think I’m going to turn down $3 billion” in 90 days. That’s what’s different about today versus SaaS land.

When you get to product-market fit, the action and the odds at the craps table are pretty wild in a way that just didn’t happen in SaaS land. I mean, it’s just like, “Yippee, we won $3 billion.”

Why? Sorry, just help me. Why is that? Is that because the distribution is different, the adoption is different, and the willingness from large enterprises to pay for AI tools is different? Why is it that when you get PMF, it’s like—crack—instantly to $3 billion in 90 days, in a world where it wasn’t before?

I think it’s all of the above, actually. It’s a pretty good list. I think that stuff is working quickly and people are adopting it quickly, so you have that raw take-up. There’s a common consensus that the prize is worth taking.

The important thing to remember about AI is that if you compare PC hype, Internet hype, and AI hype, AI hype is bigger than all the other hypes put together in terms of raw belief that it’s all going to matter. 4 years into the Internet revolution, Krugman was still doing the “maybe it won’t matter.” And I’m picking on Krugman, which is not fair, because God knows we’re going to need international trade economists in today’s world. But there was still this attitude: “Does it matter? Is it all just a bunch of kids? Is it all stupid?”

2 years into AI, everyone on the planet and every company is saying, “Shit, I’ve got to do something here.” So there’s a common consensus across the entire knowledge-worker world and the entire corporate world that this shit matters. And when you have that, I don’t think you turn up to your board and say, “Well, AI really matters, but I’m a bit nervous, so we’re not going to make a big play. Maybe we’ll get someone else to run this operation.” So I think there’s just a willingness to bet big.

Harry, I caught up with Marc Benioff the other day and talked about AI, and his feedback was, “We have a group that’s all in.” He had all his logos—the Leonard Holmes and Singapore Post—and he’s like, “But I’ve got to tell you, it’s so early for others. It’s so early.”

I thought about that for a minute. What are we seeing happen? This is just my sense, okay? I think what’s happening with AI is that every early adopter in the world is looking to deploy. Every single person—whether it’s an experimentation budget, a restaurant that actually cares about AI, or the 3 of us—is looking to deploy. I mean, I’m running our SaaS through AI. I’m running on tools. I’m ready.

From 2021 to 2024, you couldn’t get me to look at anything. My life was too busy. So I honestly just think this growth is crazy, but it’s a moment in time where every early adopter is in market. 100% of the early adopters are in market. And that’s why I think it’s early, because 90% of the enterprise is not even there. They’re just playing with ServiceNow, and they’re just playing with things.

I just think this growth is crazy, and a lot of it is self-serve and product-led, easy to deploy, and cheap. These products—if you don’t use much Windsurf, it’s $20 a month, dude. This is not high risk. You have to put it into production, but this is not $400,000. They have an enterprise sales team, but all these products are cheap, guys.

My jaw drops at how cheap these products are, and they make regular B2B look like a frigging rip-off. Every early adopter is like, “I can use Higgsfield for $5? I can use Windsurf for... Why wouldn’t I? I’m in market,” right?

If it was $20,000, which is what it probably costs just to get Atlassian to engage with you for the enterprise, right? For some 20-year-old tool. But, man, $20. $20, right? So I just think every early adopter’s in market, and that’s why we’re seeing growth at the levels we have. I don’t think it’s as crazy as it sounds. Instead of 5% being in market, it’s 95% of the early adopters.

5. Hot Hands Leave Big Firms

Completely. One of the other investors in Windsurf was Kleiner Perkins, and it was surprising news when I saw that Bucky was leaving Kleiner Perkins. Bucky is heralded as one of the successes.

It's just one of another younger people in venture leaving one of the bigger brand-name firms. I'm intrigued to hear how you thought about it. Jason, why don't we start with you? How did you read this, and what are you seeing in terms of younger people leaving brand-name firms?

Well, first of all, honestly, I was accidentally an early version of this, okay? If you have a hot hand in venture and you're not running the place, I would leave the next day. That's what I did. I had the same conversation with Tomas Tunguz, who's already raised about $700 million. Everyone loves Tomas, right?

I won't share all the conversation, but one of his comments was, “I should have done it earlier,” right? As great as Redpoint is, he's basically a solo GP managing close to $1 billion. Probably better economics than being paid $400,000 to $1 million a year, plus waiting 22 years for some carry. Why? I'm not saying that's what happened with Bucky, but probably if he was going to run the place in the next 5 years, he would have stayed, right? Or whatever the dynamics are. It's just, if you can raise your own fund today, you would be silly not to.

I'm not saying it's true at Scale, Rory. But at 90% of VC funds, why would you stay and be given—when I worked for someone else's venture fund, I was told what my salary was. Fuck you. I did 10x in that fund, and you're going to tell me what my salary is? I don't even get to go to the management meetings in a tiny fund. Fuck you, right? Why would anyone stay in those environments?

Rory, listen, you're on the other side of the table. Fascinating. You have amazing young people. Why do they stay, and what would you say to them?

First of all, I'm laughing. I remember meeting Jason for the first time. You said, “If I had a hot hand and I wasn't running the place, I'd leave.” Knowing you as I do, Jason, you can delete the first part of the sentence. If I'm not running the place, I'll leave. Some people just want to run the place, and that's you. I totally respect that, right? I'm giving you shit here.

No. Sometimes you just want to be a partner. As a founder, you don't need to run the place. You just want to be a partner—a true partner, not a general partner or whatever. You want to be a true partner.

That's actually a much more actionable comment. I agree. I don't think it's as easy as, “Hey, I just want to leave,” because it's nontrivial. There's a bunch of stuff required with raising a fund and all that. If you pull it off, it's great.

I like your distinction. I think people want to work in an environment where it's fair in the sense of their return: the compensation they get is roughly commensurate with the value they put into it. That's hard to do, especially in a business like ours, which has such long lead times and such long proving-out times. But if you don't build that kind of organization, then you don't have stability. You need to do that to have generational stability.

Starting with that comment, because there's a lot in this comment, you want to make sure that, in a rational world, everyone's incentivized to stay rather than leave. Which means, brutally put—and no one ever says this—the implied statement in this is that the hotter your hand, the more incumbent it is on the leadership of the firm to make sure you're in the circle, not out. I think it's centripetal, where you're pushed in, right?

If you're sitting there as a leader and you've got a hot, talented younger partner who's killing it, if you're not putting them inside the tent as quickly as humanly possible, you're an idiot. The good thing about that is the system works. It's polite, because venture guys are politer than your hedge-fund guys. But in the end, well-run firms make damn sure that, in the main, people who are doing well get promoted and cut in. That's our job.

If we're not doing that, you're right: you'll start to lose good people, and shame on you. That's the job of “the established side of the table.” The specifics are all over the map. I know Bucky. I know that man as well. I know Mamoon very well. I remember when Mamoon was a young guy moving on from his first firm. We've all been on both sides of the table, right? I'm not going to comment on specifics.

There's a range of reasons why people leave. Sometimes it can be that I'm doing great and I'm not getting the reward I need. Sometimes it can be that I'm doing great and there is no reward because everyone else has lost all the money. So no matter how hard I work, in the next 5 years I'm just digging out of someone else's hole.

Do you think we will continue to see spin-outs from A-grade firms, from young, incredibly promising partners?

I think it's been the nature of the business for the last 30 or 40 years, so I see no reason it will change now.

It is, but the number of spin-outs has increased significantly across the board.

That's only because your window of view is fairly limited, right? If you look across 30 years plus, it's exactly when it should happen. There are 2 reasons why it should happen now, and 1 reason why it might slow down.

The reasons why it should happen are, one, you've had 5 or 6 years of slowdown—what looked like amazing performance, lots of promotions, and then a whole liquidity gap and markdowns. Everyone's looking at the last 2 funds and saying, “Oh my God, if I was here early enough, I did great in those early funds. The last 2 funds, hmm, maybe if I hang in another 5 years, we'll make a 1.7x. Maybe I'll make some money.”

If I'm a junior person, hot to trot in my career, and I think I'm good, I'm looking at that and going, “The expected value of this isn't great.” I'm a rational actor, and anyone who's wanting money should be a rational actor. So the first thing that's causing it is big-ass firms where you're not sure you're going to get money.

The second thing that's causing it is LPs still wanting to do the asset class, but wanting to do new firms while simultaneously also doing huge checks to the very same firms that people are leaving. It's just quite a funny dynamic.

I think what's really happening here is—this is perhaps too silly—that deep in their hearts, LPs are looking at these mega-platforms and going, “Hmm, I've got no choice because it's the only place to put a lot of money. But oh my God, I'm scared. I'd really like to feel good about myself in the morning. I should do some young up-and-comers too.”

So I put my $200 million into a mega-fund, and I gave $20 million to Tomas. I feel good. And I think there's quite a receptive market at the moment.

Do you think you suffer from the barbell? Which is exactly that: you want to put money in the sub-$100 million young, new firm or the platform play with multiple billions and GPs like Lightspeed and Andreessen.

I think “suffer” is an interesting word. We're all only as good as our last game. I understand what you're saying. In a world where people say, “There's only 2 things I want to do: the mega-funds and the designer new funds,” yes, in that world I would suffer.

But that's not the world that worries me, because there's something that worries me more. The world you really suffer in is if you don't perform, right? If you do perform, no one gives a damn if you're small, medium, or large. You know it, right?

Zheng Zhou paying, it doesn't matter if a mouse is black, a cat is black or white, as long as it can catch a mouse. That's why I said to my colleagues, “Guys, we're competent, but our number-one job is to be competent.” If you execute, I believe there will be a market for venture returns. If you don't, then you're right. In the absence of success, people can impose their biases.

At the margin, people love these new firms, the new stories, because it's the promise of the new. The other wonderful thing about starting a new firm right now that no one will ever say is that not only do you ditch your colleague's track record, you also ditch your own.

You literally go there and say, “I was at a mega-firm from 2016 to 2024. I did some deals. Some are great, some are shit. It's not obvious yet, but deep in my heart, I know. I'm just going to sever that thing like a stage of a rocket, move it behind, raise money now, and I will never be asked about my mega-fund return ever again as long as I make this new fund work.” It's beautiful.

You still tell the stories of your winners from the prior fund, right? You pick out your own returns and your own results from your winners, right?

To take an example of that, someone who I think is by far one of the most talented investors of the last 30 years is Fred Wilson. Flatiron was wildly unsuccessful in the dot-com crash, and he went on to do a new thing and killed it from day one.

Which is why it might also be a sensible bet. You sever your own track record, good, bad, or indifferent. That, frankly, was largely a function of the times, not you, right? People don't ever want to say that. You learn those lessons. You're way more intentional as a startup about what you're doing, and therefore you kill it, right? That is the cycle of renewal that can happen.

It’s just one of those things that happens at this stage in the cycle.

The one thing I would say is I would not want to be going out fundraising at this time. The LP appetite for new funds, I think, is lower than it’s been in a long time. LPs are not jumping at the bit to commit to new managers, either existing re-ups or net new. LPs are waiting.

They don’t want spin-outs. You would know better than me, Harry. I’m shocked that the spin-out play—maybe it’s for the reasons Rory said—I think the spin-out of a successful GP de-risks it on the 2-by-2. It’s not the spin-out, it’s not the 2021 playbook of, “I’m getting 3 buddies together to do an $80 million seed fund.” This is cherry-picking a top manager from a known brand fund. You would know better than me. There’s still appetite for that.

It depends who and where they’re from. You might have VPs from Index all day, every day, raising 10 times whatever they want to raise, 100%. But the withdrawal from endowment funds is very real. The awareness that fines are coming, very likely for many of them, and that tax-exempt status is at risk means there’s just a lot of uncertainty, and a lot of them are just waiting.

I agree. I think that, as is often the case, two things can be true at the same time, especially when you have 3 people all busy talking past each other. I think that Jason said it right. The large number of new funds that were happening in ’21, whenever it needed to be invented, is way down. First statement.

Second statement: the funds that are getting done are talented mid-career GPs from top-tier firms with a good track record. A much smaller number, but we all know them by name because they’ve been in our business for 20 years, in a way that I didn’t know Joe XYZ, who raised in ’21 and I’d never heard of, right? Every single one of these people, you go, “Yeah, that makes sense. We’re going to deal with them. They’re smart. They’re good.” You do references; they’re great.

But I do think, to your last point, that was then and this is now. The interesting thing about the next couple of years will be: is there pressure on endowments, who are typically one of the best funders of new designer funds—high-intensity, high-conviction, smaller funds? Are they in such a world of hurt in the last 2 or 3 months that, even with the best will in the world, they’re just not going to be able to do these deals?

I think that’s a legitimate question, which is why I think it won’t always be the case that every young person says, “Yeah, I’ve been successful. I should leave.” You might see, in the next 1 to 2 years, a little bit of clinging to the lifeboats here, guys, because it’s not going to be as easy as it was. I think ’23 and ’24 were a unique time. It’s never easy to raise a new fund, and these are really talented people. But ’25 and ’26, you’re right, Harry, could be tougher.

You can want to do something all you want. You can want to buy a Ferrari if you want, but if you haven’t got the money to buy a Ferrari, you can’t buy a Ferrari. These guys are going to be really strapped for cash.

6. Endowments Need Liquidity

Rory, can I ask you: do you think the endowment funds are as in crisis as people seem to make out?

I wouldn’t be surprised. It’s a terrifying set of circumstances. If you’re an endowment, you have—we’ll talk in a second about the yield thing—you have down public markets, you have illiquidity, and you have low venture returns for a long period of time. Those things alone would have put stress on the system.

What typically happens when you see stress is that an exogenous variable puts you over the top. In ’73, it’s the oil crisis. In ’25, it’s the Trump crisis. He has clearly taken it on himself to decide to significantly change, with brute force, a significant slug of the very institutions that have large endowments and are providing a lot of capital to these startups.

So, deliberately not commenting on the merits of it for a second, if I was the CFO of an Ivy League university, let’s just say my cash planning for this year would be dramatically different than my cash planning normally. If someone sauntered into my office and said, “We need more illiquid assets,” I would say, “Get the frick out of my office. We don’t.” It’d be like, “No, I’m thinking bonds here, dude. I’m thinking index funds. I’m thinking accessible cash at a moment’s notice when 30% or 40% of my revenue could disappear.”

Rory, I literally had, across channels, 50-plus LPs in my inbox after our last show. They will all be screaming, “Okay, but if I don’t do these venture funds, I’m going to lose that trusted relationship with Mamoon at KP, with Danny at Index, with Brian at Founders—you name it. I can’t just go, ‘No, I don’t want more illiquids.’” So what would you advise them with that in mind?

2 things. Going back to your point on the new funds, what they’re saying is, “I don’t want to lose what I have.” What you see then is the bias toward, “I’ve got to start by protecting what I have, the relationships I have.”

Yeah, it would be a mistake if you’ve been in Sequoia for 30 years and you’ve got 1 chit left this year, and option A is to nuke Sequoia and do this new fund—2 really smart people, they could be amazing—and option B is to keep a Sequoia relationship. You know, because those guys do vindictive like no one else, that if you pull out, you’re done forever. What are you going to do? You’re probably going to stick with your existing relationships.

So, first of all, you’re right: they themselves have to make choices. Then, as I think we mentioned it last year, the second thing is that, at some level, some choices get made 1 level above you. They can say, “I want to keep rather than add a new one,” and that’s 1 choice. Then someone 1 level up says, “I want liquid assets rather than illiquid assets.”

Liquidity premium is one of those words that doesn’t mean shit until it means everything. When you need money—when you need money to fund your students, or to pay your professors, or to fund your research—you end up selling. You end up giving up on upside either by not pursuing new deals or even, shock horror, selling existing assets because you just need money.

And we are seeing that. Yale are reportedly selling a reported $6 billion pool of different assets in a secondary sale. Are we going to see that? Is that the start of a new trend for endowment funds to get the liquidity they need for the outflows they have?

I mean—and Jason, chime in here, because I’ve been on a rant there, so I don’t want to hog the mic. But I was thinking about this because it’s a really big deal. I was listening to your list of questions, and this question is a big deal.

Yale has been the intellectual godfather of the endowment model. David Swensen’s book, I’m sure we’ve all read it. I read it 20 years ago, and I’m like, “Hmm, that’s the definitive book. I don’t need to read any other.” This guy nailed it cold.

The Yale endowment alumni have gone all over the world. They’ve been hired, they’ve been at the court of the king, they know how to do it. It’s spread across many endowments. Intellectually, as I said, this would be like if Vanguard said, “You know, we’ve been thinking: active management is the way to go.” Right? You know, right? The question is, what’s really going on here?

There are mitigating circumstances. I was mentally running through it. First of all, the rumor might not be true. I think it is, but I don’t know. But there are 2 or 3 reasons why they could be doing it, in ascending order of severity.

The least severe is, “Hey, we just think we’re going to need money. It kind of sucks. We love these assets, but we just need capital.” It’s not a knock on the model; it’s just a knock on, “I didn’t plan for the fact that the President of the United States might try and effectively take away our federal income.”

So there’s no collapse of the intellectual theory. You just misjudged the amount of illiquidity you could afford because you misjudged the variability of your cash flows. That would be a conclusion that it’s bad short term because it’s bad, and it speaks to other people having the same problem, but it’s not saying the whole model falls to pieces.

Obviously, an even worse conclusion would be that they’ve been looking and thinking and saying, “The entire private sector’s overfunded. I want to pull back a little.” We just think systemically, long term, that’s not the case. But it is a big deal because they’ve been so damn good for so long.

You guys would know better than me, but in the few conversations I’ve had, everyone got their distribution planning and their cash planning wrong. That’s what happened in the industry. This is even pre-Trump, pre-everything. These are conversations I had late last year with LPs: “We’re cool with our paper returns. We’re cool with our gross and our net IRR, but our cash planning was just wrong. We did not plan for this liquidity drought to last this long.” That’s what I heard. So cash planning was off.

I had heard 2 things. The second thing I heard—and I know this is 20VC, not 20PE—is, “Venture, whatever; it’s PE that’s the big problem.” They’re really the same thing. Venture is a subset of PE, right? PE is so much bigger, and the fact that these deals did not go public in 2 to 3 years is the bigger stressor.

They don’t love venture, but they’re modeling 20-year illiquidity with regular cash-outs, right? They don’t sweat VC as much as PE.

I think we're suffering for that—for the bigger cousin.

What does that mean? Sorry, we're suffering for the bigger cousin? What, because PE hasn't had the liquidity that VC—

Because they're putting 5 times as much into PE, or 10 times as much as venture. Venture's a rounding error in most endowments, right? It's a subset, just like seed is a subset of venture; venture's a subset of PE. It's not that important. It's just juice, right? It's just a way to juice your returns.

PE is where you deploy more capital, and if it's been 5 years and your cash flow models are off there, they haven't brought the cash back. It's great we bought Zendesk, Anaplan, and Shmanaplan, but if none of them are returning cash, that's an order-of-magnitude bigger issue than these little 8- or 9-figure checks into Scale or 20VC. Those are rounding errors. Those are just juice. Those are to get some extra basis points on the overall endowment, right?

I mean, there are exceptions, right? But mostly it's juice. Mostly it's a little extra alpha on the endowment.

Don't forget Cooper and Zuper and Duper. I mean, that's—

Yeah.

They were really damaging.

Yeah, there, that's the stress. This is pre-Trump, but that's the bigger stress than venture illiquidity, right?

We constantly go in and out of the denominator effect on their public books, based on how it's weighted according to their privates. Every time we have a big swing in public markets, everyone's like, “Ah! Ah!” They're feeling the denominator effect. I'm like, is that really a thing given the increasing volatility of public markets today, meaning that you're constantly in and out of denominator-effect danger?

It's a thing, but it's not the thing that we're wrestling with now. I mean, a lot of things in life can be problems, but the question is how serious a problem, right?

The denominator effect—just in case any of the readers don't know—is that you have a target allocation to private equity. Let's say it's 10%. Everything's going great, but then the public markets take a bath and you're at an allocation of 10/90. Public markets take a bath, that 90 goes down, and your allocation changes. Privates don't mark to market as aggressively on either side, so your private allocation goes from 10 to 12. You have, quote, a denominator effect.

It's an issue, and I've definitely had conversations with people over the decades where that's a thing. But if that's all that was happening, I think people would power through, right? I think Jason's exactly right: it's a combination of, at a minimum, maybe 3 things. At a minimum, the cash models have been wrong and it's all taking longer. That's what we know for sure, because it's true.

The second thing we don't know—and that's the scary thing—is whether our models are wrong on timing but the IRRs are good. In other words, are we still going to get the return we want from this asset class, just over a longer period of time? Same IRR, but compounding for 6 years rather than 4 in the case of PE, or 12 in the case of 8 for venture, right? In other words, am I still good for my 17%, which is 600 basis points above small-cap return? Or, more concerning than just timing, has that 17% gone down to 15, 14, or 13? Am I getting paid? Because now I'm not getting paid for the risk. I'm taking the risk and I'm not getting paid.

But the third issue, and most catastrophic, which I think is particular to the endowments now and not anyone else, is I actually need the damn money.

I think that is actually a very pertinent note to the endowments, because they have a lot of the time-mandated outflows that they have to spend on upkeep of community facilities and scholarships, whereas a lot of other institutions do not have those mandated outflows.

But yeah, when that happens, things just get harder. Six months ago, I'm willing to bet, if you looked at the plans—because endowments run from June 30 to June 30—no one had on their plan for FY 6/24 to 6/25, which is the current period, the President of the United States taking away two-thirds of our funding. No one had that in the plan.

We're just dealing with something so extreme and so far off the norm that I'm sure everyone got caught. It's like saying COVID: I didn't have a COVID plan in January 2020, right? Neither did you. Neither did anyone, right?

I hope they're happy too. I hope Harvard and everyone's happy, because I know I'm not going to write a big check because they poked the bear. I'm not sure whether I have empathy or not, but all the emails saying, “Please give us more money”—it ain't going to work on me. I ain't going to write a huge check because they poked the bear. It's not my problem.

Yeah. Well, the last thing my wife said to me, because we are paranoid former green card holders who have now, thank God, become naturalized citizens, was, “Don't say anything that will get you singled out to the president.” Literally the last thing she said before I walked upstairs.

I'm not being political here. I'm not joining your bandwagon, I'm not joining in on that, and I'm not joining in against it. I'm just a simple, humble naturalized citizen who wants to stay in this country. I do feel a little empathy for some of these organizations, despite their prior sins.

The one thing I will say is, I was surprised. When we raised the fund last year—whatever it was, 9 months ago—we obviously spoke to a lot of tier-one endowment funds. The number that had over 30% in privates was shocking to me, and over 30% even in venture was shocking to me. In my head, it was always—

I didn't know there were that many with 30% in venture. That, I did not know.

Yeah. In my head, it was 6% to 10%, honestly, Jason. When I heard 30% from 5-plus big names that you would know—

Yeah.

Reciting my David Swensen: one, it's not shocking if you have the perspective—which you should have correctly—that the longest-lived political and corporate institutions in the world, other than the papacy, are the universities.

These guys have multi-century timelines. Bologna was in, uh, Bologna was 30. Oxford and Cambridge were 13th- and 14th-century institutions. Harvard is, I think, 18th century. If the longer timeline you have, the more you can take on illiquidity risk, provided you're getting paid for it. So it's not crazy for these guys to have done that.

In a world where your plan was to disperse at most 3% of your endowment every year to fund scholarship needs, you just didn't plan for the situation where the world could change utterly. So I get why they're there. I didn't go to any of these universities, but I'm going to come back to it, Jason: I am more than mildly sympathetic despite their past sins.

I think there's pounding that's trying to drive change, and then there's pounding too hard. I'm just looking here. My God. As a non-American university graduate, I never thought I'd be giving this commercial, but this is one of the best products we have in the country. We get foreign students to come over here, put down $60,000 or $70,000 a year without blinking an eye and pony up for our education, and we get them to feel good about us afterwards, right?

It's not clear to me why killing this particular golden goose is a good idea, but that's not my mandate. So I've got that off my chest and I can move on now.

I think that was fantastic. I think your wife will be thrilled.

7. AI Rounds Keep Growing

Another one that I saw this week, which I thought was really important, was actually, I think it was Bryce from Oat VC. He was essentially saying, if we're building single-person billion-dollar companies and AI makes it so much cheaper to run companies, why are rounds bigger than ever? Traditional seed-company rounds are bigger than ever. Why, if everything is much more efficient?

Everyone wants to invest in the companies that don't need their money. As valuations inflate, they're just going to absorb more capital. So that's number 1. We all want to invest in things that don't want us.

Two, founders are utterly insensitive today to raising at astronomical valuations. There is no sensitivity to the risk of raising at a $100 million valuation. I mean, $100 million to me is the last chance to not go for it, okay? So I tell every founder to stop at a $100 million valuation if you're not sure you're going to IPO.

Now, you may get it wrong, but if your gut says, “I don't know, man. 50% of $500 million, that's not me,” don't raise north of a $100 million. One billion, two billion, three billion, $10 billion valuations—the kids these days, Harry, the kids, the generation after you, they don't care. They see no risk in raising at billion-, multibillion-, $3 billion-, or $10 billion valuations. They just don't see it.

The combination of that and wanting to get into the hot deals means they'll absorb, up to a dilution threshold, lots of capital. They'll just absorb it. I don't think it's any more complicated than that. There's other reasons, but I don't think it's any more complicated than that.

I don't think VCs want to invest in companies that are capital-efficient and don't need their money. Accel figuring out how to buy 30% of Atlassian back in the day was the genius move. You know, when I started investing, I think I met with Rich Huang.

I'm like, “Why don't you do all these deals?” He's like, “We just can't find enough. We don't want to do anything except Atlassian. We want to be the only investor and own 20% to 30% of a bootstrap company,” right? VCs love companies that don't need their money.

I love it, Jason. We VCs want to invest in people who don't want us, who don't need us. We want to invest in capital-efficient companies, and we want to do that in a capital-inefficient way. It's a paradox, but it's true, right?

8. AI Makes Rollups Investable

Do you know what I see more than ever? Insane amounts of AI-incentivized roll-up plays, whether in legal, accounting, or professional services. A lot of home real estate plays. Is this a venture model? Is this not a venture model? How do you guys feel about the AI-incentivized roll-up play?

I'm modestly skeptical, which means, ironically, I'm going to start by citing a success. We were investors in SpeechWorks, which became Nuance in 2000, and they were a generic speech-recognition company—AI, a prior generation—and did a lot of broad corporate stuff. Then they found this vein of gold in medical transcription, and the way they built that business over the decade from 2005 to 2015 was that they bought crappy little mom-and-pop transcription companies, injected the AI, and made it work.

So there's an example of where it did work over an extended period of time. But—and this is my but—I think it's a crappy model. Just for contrast, right? You weren't expecting that, were you, Harry?

Why do you think it's a crappy model?

Because the bet you're taking is that you're buying a set of customers who weren't picked by you because they were suited for your product. They were picked by some mom-and-pop founder as the 10 customers they could sell to. Then you come in, and maybe your AI is so good that it can address all the needs of all 10 customers.

But in my view, it's much more likely that, of the 10 customers, 3 or 4 of them are a perfect sweet spot. You get them across to being all software, and it's great. The other 5 or 6, to a greater or lesser extent, have slightly different needs, because, remember, they didn't pick you because of your AI—you didn't have your AI when they picked you.

So you're going to have a lot of churn and people for whom you can't make it work. It's going to take longer to bring them across, and you're not going to develop any new deal muscle. I would worry that you end up just piling in this sudden mass of services with a relatively low multiple, and I don't know if it'll be a compelling business.

I love that insight: they didn't pick you. I mean, it's obvious now, but I'm going to take that with me, because this isn't something new. It's just, in some ways, been accelerated by AI, right? This has been true since the first job I ever had. If you were a value investor, you always looked to buy a traditional asset on the cheap, right, and tack on 8 figures of revenue.

This has been true since the dawn of the internet, but they didn't pick you. That's the problem. They didn't pick you, did they? They're not going to stick. It's not durable revenue in any way, shape, or form, right? So financial engineering is all it is, in most cases.

I would actually push back on both of you. One of my fastest-growing companies has gone from $0 to $30 million in revenue in 2 years with a pure roll-up play that's helped by AI tooling. To your point about the customers not picking you, the customers are all pretty much identical. It's a real estate management product. They're identical in the service they require and the product they engage with. There is zero ambiguity.

The ability to roll out to a uniform customer base makes it a very efficient model, actually. The thing that matters most, then, is just acquisition price. Can you acquire it effectively at a good enough price? What's your speed of turnaround in terms of your payback? What's your margin juicing? We go from 5% to 40% in 6 weeks. That's a big increase in a short time.

Agreed, and the key sentence was the first one: all the customers are exactly the same, and it's all tuned to your technology. I buy that. There will be examples of this that work. There are examples of everything that work, right?

But the more broad-based you go, the less likely it is to be efficient. I'm not saying it's never going to work. I'm just saying it's probably a lot harder than you think, and it really does boil down to this very clearly defined use case where you can be certain that most of them will come across.

Everyone's looking at buying BPOs, right? Everyone's looking at that in the contact center.

9. Enterprise White Space Remains

Where is no one going that more people should be going? If we're seeing roll-up plays be massively over-invested, I see them more than ever. Where do you think not enough people are going?

This is the moment where your little heart inside says, “If I knew for sure, I'd be damned if I'm going to tell you, Harry Stebbings.”

You're giving me a hard time today, Rory. You're like, “Delete your filters.” You're like, “No, you're a hypocrite.”

Yeah, sorry. I just came off a board meeting. It's been a tense day already, and it's only 10 o'clock. I think we talked a little bit last time about how both Jason and I would still do the triple-triple, double-double core SaaS company in enterprise.

Normally, you try and trot out esoteric areas, but all the esoteric areas are full. There are people doing rockets, defense, and healthcare. So there's a little part of it that says, “It's not our focus area.” But a partner said to me yesterday, “Guys, everyone's running away from consumer. Maybe you should spend some time there.”

It's not our thing, but if you were a personal investor, whenever everyone is running away, if there's still a technology that's animating progress as this moves from just being a trailing-edge tech thing, you do have to say to yourself, maybe. But I don't have a ton of amazing new-place insight, especially there.

It's so interesting you said that. I tweeted this week and thought, 5 years ago, there were 2 to 3 competitors for everything that we looked at. Now there are 10 to 15. I just got out of an IC where we were looking at an L&D tool—a learning and development tool—for GenAI security.

It came up with red flags, and then it came up with the market map. I was like, “Well, wait. In the last 12 months, these are the competitors for learning and development in large enterprise for GenAI tool security? Oh, God.”

But if you just step back for a minute, the LMS space was already overcrowded before AI, right? LMS was one of these classic spaces with too many vendors for a mid-sized TAM. That's the worst area to invest in. I've done several investments like that, but you have to be intentional if it's lots of vendors and a smaller TAM.

I will say, just being very tactical, there's not as much innovation in the true enterprise—not B2B, not mid-market, but gnarly, big problems. It's just not what all the kids in San Francisco know. So there's always going to be less investment with A- and S-tier teams solving gnarly enterprise problems, especially outside of security.

There aren't going to be 100 kids who want to build the next ServiceNow. There just aren't that many, and there are many other examples. The other one that's obvious, but I think people miss, is that there's so much excitement around vertical agents, whatever these are.

I still believe vertical SaaS is under-invested in, because the AI wave is just coming to vertical SaaS. It's just starting in a lot of these categories. You're going to see 5 competitors, but you're not going to see 500 in a lot of categories. In legal you are, and in sales tools you are, but a lot of categories aren't going to have 500 AI competitors. They just aren't going to know the markets well enough.

So this deep market expertise in enterprise—I still think you're going to have fewer competitors.

Do you guys speak to all the companies in the space? Will you really map it that effectively when there are 10 to 15 competitors? Will you do that work?

You'll try to, because everyone says you do it all. But everyone has their model on what works, and if you're doing seed, by definition, I think it's really hard to.

At the stage we invest at, first of all, there's an observation: when you look back at success, when you have this market and can pick the winner going in, to state the obvious, you get disproportionately greater outcomes. It's the first stage at which it's vaguely knowable. Maybe that's a better comment: pre-seed, it's all unknowable, right?

But there's this early proto-market where everyone has $1 million to $3 million in revenue. You can at least try and figure out who the winner is, right? Do you actually get in front of every company before you pull the trigger? It's hard to do that because, let's get real, you talk to the first one and think it's interesting. You want to talk to the others, but you might have to make a decision now.

So I'm not going to lie and say I never pull the trigger until I've seen all the players. But I'll tell you what you have to be damn sure of: when you're pulling that trigger, do I know who the universe of competitors is, and do I have a decent sense of how they're doing through jungle telegraph, customer references, third-party references—whatever you can get?

You should do that. You shouldn't be blundering into deals. Therefore, the fatal error by definition at our stage is this: you fast-forward 12 months, and it turns out the number one competitor is someone you haven't heard of. At that point, you should be committing ritual suicide on the boardroom table because you screwed up.

You do want to have a sense of the market map. Maybe, to say the earlier comment more succinctly, you often go into deals and you don't end up with the winner. Duh. That's what losing looks like. But you really don't want to go into a deal knowing upfront that you haven't got the number one. That's like saying, “Let's lose money here. We have a good plan, people.” By definition, that's one of the key things you've got to do diligence on at this stage: is there some compelling reason why these guys are ahead?

When you look at the distributions that you've had, have they been markets where it is winner-take-all or much more distributed? Is it an Uber and a Lyft, or is it Salesforce, HubSpot, lots of CRM plays, and Veeva for a specialized industry—much more fragmented?

The data says consumer tends to be more winner-take-all, and enterprises tend to be more oligopolies, right? We were fortunate enough to be early investors in HubSpot, so yes, we can speak to oligopoly. That's just the nature of the beast.

Even within enterprise, I would argue infrastructure, where my colleagues invest, tends to be a little more winner-take-all because you don't need a separate router for healthcare versus banking. You just need a router. You just need a GPU. At the apps level, the reason it tends to be more fragmented is that there are markets and submarkets with real nuances between them.

You threw out HubSpot and Salesforce, and you're right, they're both winners in CRM, but they're very different—to the point where Salesforce was actually an investor in HubSpot early on. The markets were sufficiently fragmented that they could both build huge companies in markets that, at one sound-bite level, are the same, but one level down were very different. So, there are many more oligopolies and multiple winners in enterprise apps than in either infrastructure or consumer. Jason, does that jibe with you?

Yeah. When I look at the billion-dollar exits I've had, they're all in brutally competitive markets that were oligopolistic or similar. I wish it was marketplaces, because you'd prefer that, right? I mean—

Jason and Rory, can I be blunt? How many 10Xs have you had?

You can. I've had 4 or 5 above a 10X.

Jason?

What does 10X mean? Sorry. You mean cash distribution? I've only had 3 billion-dollar exits, I think, right? You mean 10X deals?

No, 10X distributed cash back on deals.

Well, only 3, but I hope—hopefully there's more, right? I think only 3. Maybe 4.

That's not the only, dude. It's so damn hard to have 4. You've been doing it a decade less than me, which is depressing in itself, but I remember: it's so damn hard to get one of those, especially across a cycle. Enough with the “only.” So many people don't get any; so many people get 1 or 2. I had a decade where I had none.

I am more proud of my 2Xs and 3Xs in 2004, 2005, and 2006 than any 10X that sailed into the 2021 bubble and made me a fortune.

10. Investors Must Survive Doubt

Rory, did you ever have a crisis of confidence as an investor, and what would you say to young people now who are looking at 2020 to 2023 and going, “Fuck, am I actually any good?”

I have crises of confidence all the time, and my most recent was yesterday. Absolutely. It's actually an interesting question. This is such a hard business, and if you don't have angst about your ability to do it, you're missing the point.

Starting out, out of my first 5 deals, there was one point where I thought I'd lose money on 4. I spent 3 years of my life pretty much not sleeping. Terrifying, right? I was dreadful at this business. It was 1995, early on. So yes, that was one crisis of confidence.

Was there a crisis of confidence from 2000 to 2010, when I was vaguely confident but no one was making any money? Yes. Were there 5 or 6 years when you felt amazing? Yes. Did I have more crises of confidence in the last 3 or 4 years, when, despite having 25-plus years of doing this, you start making dumb decisions? You have deals that make it clear you got it wrong again. Of course you do. I think anyone running money in venture or anything else does that.

The question is, what you can't do is say, “Hey, I was great then, so it'll work out.” You have to say, “What am I doing wrong right now in today's market? Am I playing the game correctly for where it is today? What did I miss?” Then just go back to basics.

I do believe that if you do the right steps in the right order, you can't stop mistakes, but you can minimize them and probably do okay across a cycle. When I look back at recent mistakes, I think I skipped a step. Shame on me. When you skip a step, it bites you in the ass.

Where do you think your scale is not playing the game correctly today? I'll give you an example for me: I have access to amazing hot rounds, like some of the names you hear about. I don't do them because I just think they're crazy-priced, they seem completely detached from reality, and then I'm proved wrong consistently.

I see you're turning it on me in revenge for your hypocrite comment, which I respect.

I wouldn't say it's revenge. I would say it's more tit for tat.

Yeah, totally. Well, I'm Catholic. I've done confession. I can do this.

One of the things you're wrestling with is that you have your strategy and you want to stick with it. We all get some version of a bright, shiny object, right? It's like, “Oh my God, are the returns...” You're looking at other things, and you do try and keep it focused on the main chance: A and B rounds with product-market fit, early revenue, and looking to scale. Broadly speaking, we have kept on mission.

I'll say there's sometimes when you say to yourself, “I don't want to drift off all the time, but I'd love to be able to do one of the hardest things I think there is to do in an investment management firm: make the occasional exception without making it the rule.” That's really hard to do.

Can you reach for that $1 billion deal that you see, where you have good connections, should do it, and it makes a ton of sense, without doing the 5 other deals that are massively overpriced at $1 billion? The ability to move beyond your strike zone once in a while is a muscle that's hard to develop, but it would have economic advantages if you do it and would be catastrophic if you did it wrong. That's one area. I'm sure you have the same.

Because you implied it yourself, Harry: you see this later-stage deal and go, “It's not what I said I'd do. I shouldn't do it.” But every once in a while, you say to yourself, “Should I?” I wrestle with that.

ElevenLabs at $25 million—I could have had 1% of the company, and I was like, “1%? 1%?” And, like, “No way.” It was a $25 million fund. It's a $3 billion company now. That would've been a fun return on 1%.

Well, look, for what it's worth, there's a lot of complexity here, right? But I think any deal where you have 100% conviction that you'll 5X, you should do it irrespective of ownership or valuation, just to do it.

If you've already met with the founders, you believe in them, and you're not thinking you can make money, but you're like, “I am 100% sure I'll 5X it,” you'll always like to get an extra 5X out of X millions in your fund. You'll never regret an extra 5X. It may not return the fund, but you'll never look back when you're in carry mode and say, “You know what? If that 5 had become 25, that's an extra $5 million in my pocket.” You'll never regret the sure-thing 5X.

I know it sounds silly, but this is where I started investing, and I lost track of it by going to, “This is my rule.” This is my rule today again: if I'm 100% sure I'm going to make 5X, then I'll do it, period. No matter what—

But dude—

—I’ll just do it.

I have a $125 million seed fund. If someone on my team comes to me and goes, “I want to write a $3 or $4 million check,” I'm going to 5X it, I know, but I don't know if it's going to be much more—

No, this is what you get to do running the place. Your team isn't allowed to use this heuristic. You're allowed to use it.

To join the dots, that's why Jason should be in his own firm. I actually think Jason had the right answer because, if you play back what I said earlier, if you want to establish an exception strategy in a team organization where there's a broadly equal team, then it's really hard to rein it in. That's what I'm saying.

We're a broadly equal team at Scale. We're 7 people who can write checks. If you start breaking the rule once, then you've broken it for everyone. It's no accident that the firms that have high position-betting variance, where they're willing to go anywhere, are single-leader-dominated. Jason, you're exactly right.

If you're running your own shop, that's the joy of running your own shop. You can range high, range low. You're not trying to follow a model. You're not trying to build a thing.

So Harry, he's right. It might be corrosive to your entire culture and piss off your unique people, but the only person who can pull out the exception card is you. We've chosen not to do that.

But Rory, Jason's got no other investors. It is just him.

Well, that's exactly it. And then he can. There's a reason why the guy in Omaha who doesn't listen to anyone is the richest man on the planet, because he's like, "Thank you all for your opinion. I don't give a shit."

This is always the tension between building a firm, having a consistent strategy, versus, on the other hand, doing the reach. Going back to what Jason said, Jason is actually right about his comment on high conviction, with the caveat that it's not an institution-building strategy.

Something you said, Jason, resonated with me. There's a rule in engineering that you're only as accurate as your least accurate variable, and everyone gets caught up in the revenue model and the price. It actually is the conviction level you have, if it's informed conviction—not bullshit, "I swing from the gut"—but if you have high knowledge that this is a thing, you've got to weight that extraordinarily highly.

The future's so damn uncertain. Most things, you don't know what's going to happen. A lot of things don't happen. So if you get to that unique insight of, "This is a thing and it's going to run and run," then finding a way to monetize that bet is actually your job, right? There are very few things about the future you know.

I remember realizing, "Oh my God, every single app for the next 20 years is going to be rewritten as SaaS. I should monetize that bet." In the same way, I give all credit to the people who said, going back to what I said earlier, "These AI models are a thing. You just have to get me a piece of the 1 or 2 that are going to work." When you have the high conviction and you're not trying to build a consistent internal strategy for management reasons, make the bet.

So why aren't you doing billion-dollar rounds if you can see 5X-plus in them with the core fund?

When you're trying as a firm to build a strategy, that's a very idiosyncratic, leader-as-genius approach. As Jason said, when you're trying to do leader-as-genius, making all the decisions, you do that. When you're trying to build a peer team that are doing deals, broadly speaking, you stick to what you're doing, you have a plan and a strategy, and you accept the fact that there are going to be deals outside your core competence that work really well.

Your job, which is hard enough, is to execute your core strategy really well. I'm far more worried about missing a 5X or a 10X deal that was in the Scale sweet spot of $1 million to $10 million in revenue, Series A or B, enterprise software. When you miss that, or even worse, when you turn it down, that's when you have a bigger problem.

It's not, "Oh my God, I missed the Hail Mary that was outside my sweet spot," which might bother me personally. But in terms of building a team and a firm that's functioning, if you're not playing the system the way you want it and seeing the deals you want to see in your sweet spot—especially if you've defined that sweet spot so you can be successful—that's where you should agonize a lot more.

It's the difference between personal investing, frankly, and building a firm.

Rory, I'm aware you've got to rock and roll. This has been a pleasure, my friend.

Good to see you guys again. Fun to be back, and fun to defend Harvard, which I never thought I'd do.

Do you know what, Rory? You were fantastic, and I'm a hypocrite, you know?

Well, we knew that already, Harry. No surprise there.

My favorite moment on the last show was when you were like, "You'll just edit us all out and just make yourself sound disciplined." I was like, "Yep, this guy's smart."

Yes, you did last time. As they say in the civil service, why attend the meeting when you can just write the minutes?

Take care.

Have fun. Jason, I wanted to ask you one more question before we wrap.

Sure.

11. The Bay Rewards Density

You mentioned it on this sheet, which is that the Bay now has 82 tech billionaires, per Henley & Partners, a firm, and that's going up. Is there any point in being outside of the Valley? Is this the ultimate centralization of talent back toward Silicon Valley, unlike any other time?

I loved living at the beach during that global pandemic. I have a beach house in Southern California. I know all the best brew pubs. It's good living.

But, man, I'm literally just—I got a DM yesterday from a new-generation tech billionaire. I'm just going to go meet him tomorrow downtown. I'm going to walk to this meeting, and it would never have happened if I wasn't here. It's just 1 example from this week, right? The density here.

Listen, you asked the question: How often do you meet all the companies in the space in investment? I only did it once, on my first investment in Pipedrive, and I gave up. I've never reached out to a competitor. I've never done any competitive diligence, right?

What?

Not once. And so the way I invest, the Bay Area is just perfect for it because I can get to know people for real. I can understand the space. I have some time, and I'm just not this Zoom guy. San Francisco is so back.

Why would you not do competitive analysis? My partner Paul would literally have a heart attack right now.

Why don't I do any? For 2 reasons. First, do you have outlier growth? Only so many can have outlier growth. If you're an inception investor, if you invest pre-product-market fit, I get it.

But if you're saying, "Listen, I want to see someone growing ideally from 1 to 10 in 5 quarters or less," here's the thing, Harry: I only meet a handful of them.

When you do Algolia—

Yeah.

When you do Talkdesk—

Yeah.

You name any of them—

No one grew, grows, growing at—it's a rocket ship today.

But your Algolias, your Talkdesks, when you invested, your RevenueCasts, they were not rocket ships. They were good and exciting, but they weren't obvious rocket ships.

But they had top 0.1% growth. I can pull up the old numbers. No, I mean, Talkdesk went from 1 to 15 in 5 quarters, okay? Algolia was growing 20% a month for the first 2 years, okay? Pipedrive, as flawed as it was, was the fastest-growing one in the space.

So listen, I'm not saying this is the right way to invest. What I'm saying is, I have to be honest: I don't meet every founder growing that quickly whom I believe in. I have to believe in the founder, and they have to be growing at top 0.1% rates.

Maybe with your network, every hour you're meeting someone going from 1 to 100 in a week. But I find that plus a founder you believe in, plus the opportunity—not every opportunity exists for a variety of reasons. You might not meet them. There might be ownership issues, fund-size issues. There are a million issues, too, because I have such a narrow sweet spot.

I have to rule them out if they're too late or too early. I feel like I do every single deal where every box is checked, and I don't have the luxury of deciding whether there's one that's even better than top 0.1% growth. I'm not saying it's not flawed, but it's worked okay.

Is there any box that you're less strict about being checked?

Yeah, unfortunately, it's the competition box. I'm super loose on that. As long as the founder has a large piece of white space in the space, I'll do it no matter how competitive the space is, even if I would prefer not to.

Everyone would prefer no competition, à la Peter Thiel, right? I just don't feel like, in B2B, to Rory's point, where we're building oligarchies, we have this luxury often of having no competition. Windsurf, starting our conversation, doesn't have no competition, right?

I would love to have no competition, but that's the box I've completely given up on, despite wishing I could check it. No competition. I would love it. If I could, that would be a gift, but it's the one I've given up on, and it's harder than ever. Everything is more competitive today, right? Going to your point earlier.

So that's the one I give up on. I won't give up on growth. I won't give up on the CEO. I no longer will give up on the CTO. We've talked about that in the past. That's one I will never give up on again. S-tier CTO or I'm off. I'm off it.

But competition—I don't care about college. I don't care if you went to high school. I don't care about any of that. I give up on all the educational crap.

If you have a super-competitive market, is the core skill set that you look for in a CEO different from a noncompetitive market?

No, because if you're post-revenue and you're growing at outlier rates, you figured something out. And I will be honest, Harry: for every single investment I've made, every single investment I've made, I have not truly understood its competitive positioning until after I invested. I'll do a lot of internet diligence.

Don't get me wrong. Anything you can do on Google or ChatGPT, I'll do it. But to really understand the market, if you didn't come out of that space, it could take you the better part of a year—6 months—to really understand that market. How are you gonna figure that out for sure before you invest?

I would tell you, the guys will do 30 references: 10 with the team, past and present; 10 with customers; 10 with—

Oh my God, I love it. Yeah, I just did an investment with Marratech and saw their due diligence. I've never seen something this good in my life. Oh my God, they talked to 100 customers.

I literally said to the founder, “You gotta share this with everyone in the capital. I've never seen such good diligence in my life,” right? I talked to 2 customers when I invested in this company. And actually, after the term sheet, I talked to 2.

Do you know what's so shocking, though? This is just the entry ticket for growth firms to get a meeting now. That's how competitive—

That's why the diligence is so good?

Yeah. It's so fucking difficult for them to get in the door with a founder at $1 million to $7 million in annual ARR. That's just the ticket. “Hey, it's worth your time to meet me because look at all the work that I've done on your company.”

Yeah. Well, look, for what it's worth, I'll tell you why I have this strategy, right? At this point, I really do love a lot of the founders I invested in. You talked about RevenueCat. You're right, I invested crazily early. I love the founders. No matter what happens, I love them. I love a lot of the founders, right?

But as a business model, I'm only in it for one big, massive win. That's it. Everything I'm gonna do for the next X years, it's just for a Wiz or better. I don't care about anything else from a business model. I don't care. So I'm gonna do the best I can to find a Wiz or better going forward, and anything else, these are just means to an end, right?

So I gotta check all those boxes, and I think if you invest in enough folks in that top 0.1%, and you're lucky enough to do so, one of them will hit. One of them will be worth another true $8 billion outcome—not a fake $8 billion on paper, but a real $8 billion outcome, right? That's it. Nothing else matters.

How able do you think you are to know whether you'll invest before even meeting the founder? I know that seems strange, but if you know the market, the traction, the competition, where they sit, the background, revenue, revenue growth—

Every single time, I've known I wanted to invest before the first meeting. 100%.

Dude, if you get everything you need and you meet the founder, and they are just boring, they're uninspiring and dry, do you do the deal?

Dry?

You're just like, “Ugh.” They don't excite you.

Never met one. I've never met a founder who could write an incredible email, express incredible excitement about a boring industry, and get me excited by email, only for me to meet him and think, “This guy's just dull,” because they're already so passionate about their business.

Anyone that's that passionate about their business is fascinating. Whoever makes the best mugs in the world is gonna be fascinating. I would love to talk to the CEO of Shure. Let's bring the CEO of Shure on 20VC. It's gonna be fascinating.

So I've never met a CEO with this kind of crazy, outlier growth—whether they didn't go to college, dropped out after a month to do a podcast, or whatever—who wasn't interesting. I've never met this person and thought they weren't interesting.

I've gone to other people's meetings. I've gone to other VCs' meetings where I literally wanted to cry from boredom and bang my head, but I've never once had a founder who legitimately passed the pre-meeting bar and then met them and thought, “This guy's pretty interesting.”

Owner is going to be huge. It's a huge success, okay? But I don't think I've actually invested in a company with more competition than Owner. Not a single one. Direct competitors, adjacent competitors, virtually identical competitors. I've never—never.

Well, I don't know why, and Dean, the CTO, is so good, and that's why I invested. But why is that interesting? Just because you can play with the app doesn't make it fun to invest in, does it?

And in a shit market. Every investor does this, and they always go, “Look at who's the winner now.” Dude, Olo, Noah is great. Love Noah.

Yeah. Great founder.

It's $1 billion, just $1 billion, and it's been 15 years. Fucking brutal slog.

Listen, and I'm only an expert in Olo from afar, right? But Olo is a story of doing the wrong end of the tail. Right? Olo is trying to do basically an enterprise play in an SMB market. That's really brutal, right?

It's the same problem in anything in B2B e-commerce. If you're not doing some SMB in e-commerce, yeah, there's a few niche players, but it doesn't make sense to fund it. Even with Shopify, only 25% of the revenue is big brands.

So if you're only big in restaurants, in anything that's the consumer end of B2B, you're gonna be niche, right? You're gonna be niche, right? And the niche competitors to Shopify are dead or dying, like Salesforce, right?

Olo is great, but it's high-end, right? It's chains and stuff. Who—it's hard. There's just not enough TAM.

Do you know what I love? The number of amazing founders—I don't think I've told you this, and I didn't mean to rub it in—the number of amazing SaaS founders who tell me, “You're friends with Jason? I emailed him, and he never responded. And now I'm a $1 billion company.”

And I'm like, “Ah, yeah. I would love to have invested in everyone Jason didn't respond to.”

Well, that's why. Really, I just wanna do it. All I wanna do is 1 more Wiz. One Wiz, right? That's all I wanna do for the next 1 year or 10 years or 15. That's the math, right? None of the rest matters, right?

But, yeah, I gotta do better. Actually, honestly, Harry, the AI is already gonna solve half that problem for me. It has a rule, and I'm still tweaking the rule, but if the metrics are good enough, it forwards an alert to me to look at the deal.

So now it's better than an associate. Honestly, it's better than an associate. It will review your deck. It will provide you feedback. You can iterate. There's no pressure. You're not being judged—the AI does not judge you.

But the AI is better. It will give you all the feedback on your TAM. It'll tell you how good your growth is. It will compare it to other investments. It will tell you whether you're in the sweet spot. It'll tell you the check size we do and the ownership size we do.

And then you just set up a trigger: if it hits these numbers, just send it to me.