Figma IPO：完整拆解，以及 Melio 25亿美元收购为何令人“泄气”

My rule is: if you haven't grown because of AI, you failed.

The amount of money we're investing in AI right now—$300 billion to $400 billion a year in capex—will we get the near-term ROI? Will it be an economically rational decision when you look back 2 or 3 years from now?

I don't know. I think venture is just going to rip.

You're getting a lot of money back as an LP, right?

Just massive amounts of cash coming back.

If Oracle can get AI-native by June 30 and your portfolio startup can't, I'll smile, but I'd give up.

Figma filed its S-1 last night. The numbers were pretty phenomenal: $821 million in revenue, 46% year-on-year growth, $1.5 billion in cash, and no debt. There's a lot to like about this, so I wanted to hear your thoughts on where you think it will go out at and how you analyze this S-1 dropping.

There was a lot to like—an awful lot to like. I went through it literally on my phone late at night and went, “Wow.” You didn't even mention one of the more impressive ones, which is the profitability and free cash flow margins last quarter: 40% plus. So, it's positive free cash flow, positive, growing 46%.

Again, rough math on an iPhone on a small screen, it's like Rule of 80-plus. It's a company clearly tracking a billion dollars at Rule of 80, plus or minus. I think it's going to get a great reception.

For sure, this is just a game of figuring out what the model is going to be, right? Is it going to be $20 billion, $25 billion, or $30 billion? Will they listen to Bill Gurley? I don't think so. Figma isn't consumer, but it's got a big enough brand that they don't need to leave money on the table.

It's a minor issue, but where will it trade? At $20 billion, which is just what Adobe was going to pay almost 24 months to the day, $20 billion is 20 times current revenue. Even at $20 billion, it's not cheap.

The overall Nasdaq is on fire, but Bill McDermott from ServiceNow just joined. That's about as good as it gets in the enterprise, and they're not trading at 20 times. I'm not smart enough to know exactly where it will trade, how much it will trade above $20 billion, or how that compares to the IRR of selling to Adobe for cash, having no lockup, calling it a day, and going to work as a junior SVP at Adobe for 4 years.

I think the last sentence says it all. It's probably not a better IRR because, by definition, a bird in the hand 2 years ago is better than a bird in the bush now, but it's a much better way to live your life.

For the team, frankly, and for the investors, it just looks like a great company and presumably it's going to have a great run. I think it's a great outcome for them.

Scott Belsky was right. They should have bought it. People made fun of him for overpaying—$20 billion—but I think the beauty is that if you're Adobe, you can be a little patient. It's okay if you pay 2 years ahead if it's a winner. It doesn't really matter.

People thought they were paying through the nose and that this was a flashback to a 2021 deal. Fast-forward to today: if Adobe had bought it, they would have had a pretty good deal, especially because of the synergies.

Jason, what would you pay 2 years ahead of if you were Adobe today?

I'm not sure. I'd have to do my homework. When I was reflecting, you kind of joked about questions you should ask Marc Benioff when we did this last time, right? When I interviewed him, I dug deep on the M&A.

Years ago, John Somorjai, who's run corporate development there since—he was, I think, a double-digit employee, right? He's been there for 20 years—and I caught up with him. He said, “We're just really good at the big stuff. We're really good at buying a billion dollars or more of revenue and growing it,” right? MuleSoft, Slack, and others.

Maybe that's the same lesson here. Adobe should stop screwing around. If you don't have the talent, don't screw around on the little stuff. What else can you get that's approaching a billion dollars of revenue that's synergistic?

Adobe's at $20 billion, right? So that would have been another 10%. That's something. But I don't know. I'd have to look. What else could you buy at a billion? I guess I don't want to say the obvious dumb answer of Canva, but you could do worse.

I think the beauty of the Figma acquisition was that it was market-expanding for Adobe. Reading even the S-1, I think they said something like 30% of their users are developers; only 30% or 40% are designers. It really has become a pretty pervasive piece of software for anyone involved in building software products.

Jason is absolutely correct: it was a great call by Adobe to try and buy it. There was a little bit of a thing for a while when the deal cratered, that Adobe had mentally moved on to generative AI. Obviously, they have, and figuring out their generative AI strategy—and how that reflects in media and video—is really important.

The more you look at these numbers, the more you realize that this could have been a really central, key plank for what they were doing: exposure to developers and exposure to designers in the software industry. It was a great acquisition, and I'm sure they're looking at it going, “Damn, wish that one hadn't gotten away.”

It's funny looking back on it. I'm sure Scott Belsky talked them into doing it. Dylan Field talked him into doing it, right? Scott Belsky—our startups were acquired almost at the same time as Adobe's—but he went back from Benchmark, where he had resigned as a general partner, to be chief product officer at Adobe.

I never got the call, by the way. I wasn't even fifth in line. I never got the call to go back to Adobe, even though I built a much larger business by revenue, maybe not by impact.

The life lesson for founders—it's tough—is, look, Scott's gone now. It's always a weird thing when you do M&A and it's not with the CEO, because there's a good chance you outlast them. There's a good chance you outlast your sponsor as a founder when you're acquired.

You joke, but it reminds me of venture deals and the biggest problem that I see getting done today: companies being orphaned and junior partners moving on, then having no champion in a firm and being, bluntly, up the creek without a champion internally.

You're right. Across the board, you're seeing a whole bunch of change. You'd think that any partner would stand up for the firm and represent the firm, but it is funny when it's not your deal. That willingness to go the last mile to make it happen—to get on the plane, hassle the VC and the CEO, and try to get everyone around the table together to save the company—is different.

There's an argument that says the new person is actually a little more clinical, and maybe they only invest in the very best companies and maybe the other guy should die. But there's no doubt that when you become an orphan as a company, your ability to get something done in that venture firm goes down.

Parenthetically, I would say some people wrestle with the choice: if the VC firm is still willing to say, “Hey, they can keep their board seat,” do I stick with the partner who's moved on but is still willing to help? In general, I found that doesn't work over time because you need to be in the room where the money is allocated.

I'm realistic enough to say that even if you have a venture board member who's an amazing board member but can no longer speak to the money because they're no longer at that venture firm, then, to a rounding error, they're useless. If you want a great independent board member, put them on as an independent board member, but you need to have people in the room to speak for the money when the money is needed.

Frankly, that is one of the top jobs that a venture firm has to do.

I actually know a firm that says, “Hey, reserve decision-making is so broken that we have a different partner who makes the reserve decision.” I think that's useless because they don't have as much of the historical knowledge and data about how that company has progressed and changed. Then you come in as a net-new investor.

They say, “It means you come in without bias and without any lagging indicators that could, bluntly, determine your decision-making,” which I think is a great idea: not having the partner make the reserve decision.

I like this because when I worked at a third-party venture capital firm, people would just want to bail out their Bs and Cs all the time.

Like every—like half the partner meetings were, “But, you know, getting the company back to 20% growth. Let’s put in another $4 million.” Those checks would often be gone in a year. The company was going under anyway, right? I’m not even talking about 1s and 2s. I’m talking about 3s and 4s that people want to put money into, right, J?

I don’t get their confidence, though, because every additional dollar you sink in, you’re putting your name more and more on the line. If I don’t have something that I feel great about, I don’t feel good about concentrating more and more, whether it’s my firm or not my firm. I think this is a minefield—not on the agenda today, but a minefield—because, Harry, to your point, first of all, everything in life goes back to incentives, in my opinion.

Your incentives, as the ostensible leader of your firm, are very different from your incentives if you are the 15th-most-senior GP in a much larger firm, where you’re going to rise and fall on keeping your companies alive and, frankly, the overall return profile of the firm as a whole. I’m not going to say it doesn’t matter, but it’s hard to weight it highly, right? So, as that junior partner, you have every incentive to find every marginal dollar for your companies and keep them alive because something might turn up. It’s a very different decision in that case, which might argue for Jason’s comment.

I can’t imagine having someone else make those decisions, but equally, it shouldn’t be the partner making those decisions. You have to have a combination of partner recommendation and a fairly robust group process. It becomes a very interesting discussion—reserve allocation—because it’s a finite process, and you want to allocate as much as possible into your good deals at the high prices, but at the same time, at the lower prices when you can, there is some value in having capital.

If you end up naked and defenseless on all your deals—on anything other than your best deals—and you don’t have any capital to play with, you do stand a chance of losing significant economic value if you head into a tough time.

My only point to add here is that when I did Fund I, I put out my 5 top performers that would be fund-returners after about an 18-month deployment period. None of those 5 are actual fund-returners in any way, and the 5 that will be, I never had as the fund-returners. So, bluntly, I think we overestimate our ability to predict our winners. I said this on Twitter, and Roger Ehrenberg at IA Ventures very much agreed with me.

I think it fundamentally challenges reserves as a model entirely at seed to A, because you’re unable to predict your winners. When you do allocate reserves, you essentially allocate to the fastest-growing companies, not necessarily the best long-term sustainable capital or value drivers.

Well, that’s for sure. When you have limited reserves from a seed or smaller fund, you put it all in your fastest-growing companies, 100%. There’s only 1 criterion: triple-digit growth, double-digit growth per month. That’s it. What do they do again? SaaS for haircuts? I’m in.

I think your statement is possibly true at the seed level, which is not where we play, right?

100%. I think between B and C, you have a lot more to play with.

Exactly. The truth is, this is what I say to people: we do it when you have product-market fit. Once you have product-market fit and early acceleration, if you get what you want to see for the first 2 years after the investment, you’re probably going to make money, and if you don’t, you’re going to struggle. We’ve done that.

I think I shared with you before that when I look back 2 years in, if they’re at or within 25% of the right plan, our probability of making a 5x goes from 30% to 70%. At the product-market-fit stage, I think what you said is not true. In other words, you can, in fact, tell your best—quote-unquote—your best deals. Not with 100% certainty. I would have said 70%. I’ll come back to that. I’m not sure it is 70% anymore.

That is when reserve allocation gets tricky, because, on the one hand, you have the incentive to put a lot into the good deals. On the other hand, there is some value in having capital even to defend your mediocre deals with small amounts of money when financing gets tough. You can impact—you know, it’s not going to change your life, but you can take a 0.5x to a 2.5x, nurturing it along, finding a way to make a big difference and avoid a big hole in the fund.

There is some value in having some capital there, but the vast bulk of it should go to winners.

Rory, have you ever had a pay-to-play work out? I was doing a deal the other day, and the lawyer was like, “Hey, you could lose your rights.” I basically responded, “I’ve spoken to many of the greats, and all of them have told me very simply that in pay-to-plays, when they’ve participated, it hasn’t worked well.” Has it ever worked well for you?

First of all, it never works well. My first rule of thumb is this: when you find yourself going to look at the legal documents in a venture deal, you’re probably on your way to losing money, right? The minute you start down this road, you’ve made a mistake. So, first of all, let’s start with that.

If your chance of a good outcome in a normal deal is 30–40%, and a great outcome is 30%, you’re definitely going to a different place. Whenever you’re in these kinds of trouble situations, your probability of making big money is very low.

A famous example, however, is worth pointing out: Fred Smith died last week. FedEx had a down round like 6 or 7 rounds in, and I think Weiss, Peck & Greer—which is the antecedent of Lightspeed—played pretty aggressively there and made out like bandits. That is an example of a down round in a network business where it was the tipping-point round, and the people who played made out like bandits.

It can work occasionally. But you’re right: the statistical rule when you’re dealing with these trouble situations is that you’re probably not playing for the likelihood of a home-run win. It’s low, not zero, right?

You know, the second startup I worked at, Harry, had a pay-to-play, and it IPOed.

Which one?

Unfortunately, it went bankrupt between that round and the IPO, so everyone lost all their money. You could argue whether it worked out for the VCs—not only the ones that did the bankruptcy round—but it did work. It bridged them to the IPO, but they did have to go bankrupt.

But then you’ve got the challenge of the opportunity cost of that cash, though, Rory: do you want to put in that money to do the 0.5x to 2x—which, don’t get me wrong, is an incredible transition—versus putting 2 or 3 net-new lines on the portfolio?

Yeah, no, you’re right. Actually, the real opportunity cost is the time, and I wrestle with this. I would say it’s a weakness of mine. Sometimes you spend too long working with a company, trying to work it out.

Partly, I think you get connected to the entrepreneur and you want to help them, right? If they’re willing to keep going and they have a credible plan to keep going, you kind of want to try to find a way to help them. You’re right: probably on a cold-blooded basis, you shouldn’t. But sometimes this business is not all cold-blooded.

You form these relationships with people. They crank for 6 or 7 years, they hit a tough spot, and the difference between going bust spectacularly and laying off 100 people and maybe getting a 2x for the money might well be $10, $20, or $30 million for them personally. So, I will admit that weighs in.

My rule of thumb is, I tell the founder, “Look, I’m totally willing to help,” and I’ve run—I’ll come back to the analysis—I’ve run the analysis on the money overall. It gives a decent return. It’s not amazing, but you tend not to lose money on the bridges. You just don’t make as much as you think.

My rule of thumb is, don’t make this too hard as well. If you have a clear plan that you’re going to execute and get it done, and you’re not going to dink around—you’re going to figure out a way to cash-flow positive and build value here—I will support you with a finite amount of money. If, on top of that, you’re going to make a drama, I don’t need it, because it’s the time that kills you. It’s the time.

Rory, you are going to have even more fans after this. I’m actually thinking of one deal where we’ve waited for 3 years. I haven’t had to put money in, and the guy’s just literally survived on fumes and the promise that if he ran out of money at short notice, we would cover his shutdown costs. So, I haven’t even had to put the money in.

I’m just so impressed with the guy for surviving that long, and I’m glad I did it. Who knows? We might make 0.5x, but that guy kept going, and I give him more credit.

I remember a founder I had to completely bail out from fumes with more money than I had. That company is doing over $300 million today. I caught up with the founder the other day. He forgot.

I’m actually mentally running through mine. I will never forget. I have a second story like that, too.

I wish it was the only one. I could tell you another one where the founder forgot, but this one really stunned me. It literally had no—it wasn’t like he sort of forgot.

Literally did not remember that that was the way history had occurred. It’s like childbirth: you just forget the pain.

The other thing that’s insane with this, guys, is it’s going to be $3.5 billion back to Index between Wiz and Scale in a pretty compressed amount of time. I know there’s obviously the hold and the lockup, but it’s pretty freaking phenomenal in terms of liquidity and numbers back. We have a lot of the numbers. The 2 firms that are able to 3x DPI on fund sizes this big are Founders Fund and Index. Unbelievable.

I think venture is just going to rip. Dave Clark said on LinkedIn the other day that massive amounts of cash are coming back. You’ve got to be in the good ones. You’re only going to get so much back on the Superhuman deal, but if you’re in those 2, you could have a lot coming back. If you’re in Kleiner, Index, and Sequoia, you’re getting a lot of money back as an LP, right?

Yeah. I think there are 2 things. One is, at the macro level, you’re right: cash is coming. At the macro level, what’s happening here is fewer, bigger winners. It’s what I always say to people: fewer, bigger winners, right?

You’re seeing fewer winners. There are fewer IPOs, but because they’ve grown for so much longer, they’re just so much bigger. When you’re in a winner like that and you’re in early, instead of getting $400 million or $500 million, you’re returning $2 billion plus. It’s the inevitable outcome of concentration and longer holding periods.

The real skill is making sure you’re in one of those that gets it.

And you’re right. All credit to Index. It’s an amazing achievement to get that. Accel just had Scale and the stablecoin deal.

Absolutely. It’s a great time to be right. I always tell people in this business, there’s lots to like about this business: the terms, the hours, the money, the intellectual interest, and so on. There’s only 1 problem. In the end, you’ve got to be right. You’ve got to pick the right deals and be in them, and these guys did.

I think this is the problem, though, when we look at fund sizes today and say how large they are, without actually picturing what our outcome sizes will be in 10 years’ time. Bluntly, $30 billion exits for Figma, or a potential $30 billion IPO, would have been inconceivable 10 years ago when the Figma investment was made. Honestly, inconceivable. What is that outcome size in 10 years’ time? Is a trillion dollars more likely?

Yes. Inconceivable. Really inconceivable. For sure, inconceivable.

People always make mistakes, Harry, when they extrapolate trends ad infinitum. I don’t think you extrapolate just because the outcome here is $30 billion and it used to be $1 billion, so the outcome 10 years from now is $900 billion. I just don’t think the math works like that.

I think there’s been a step-function change in the stage at which companies go public. I doubt it will continue like that. I don’t think it will continue forever, but I think you’re seeing—

But I think you’re seeing—listen, Roy, you don’t need to get salty just because my friend Sam told me that you were in this shitty, messy-middle thing. Your shitty, shitty middle fund size.

I’m not shitty. I’m not shitty. I just asked Claude.

There’s an investment today that’s worth $544 million nominally. I uploaded all the financials and the investor report from today, and I said, “What will it be worth in 2029? Give me a sensitivity analysis.” It told me $3.6 billion. That’s the most likely outcome from Claude, so I feel better. It’s in the bag. All I have to do is hang out at the beach until 2029, and it will get to $3.6 billion.

Good, good to know.

Serious comment here: this is still a cyclical business, right? What you should not do is extrapolate a cyclical trend. What you see is that the investment window opens intermittently, and when it does, if you have the assets, you can do really well. But I don’t think you can extrapolate from where we were to where we are and keep that line going for another 10-plus years.

But you don’t have to. The truth is, I was thinking about fund size this morning. Seeing the strong performance from Anthropic in terms of top-line growth, and then looking at the burn—$5 billion last year going to $3 billion this year—to some extent, the fund-size argument has been answered by the burn argument.

Some of the most compelling opportunities here require a level of capital to be significant in terms of their cap table, such that it warrants at least some of the larger fund sizes. There is definitely a place to put large slugs of money. It will never be as good as when you go in expecting a $1 billion outcome, price accordingly—as all the investors did in Figma—and then end up with a $30 billion outcome.

That’s the amazing result, because you get 10, 20, or 30 times what you expected. Now that perhaps people are more adjusted to higher outcomes and bidding accordingly, you’ll probably see some deals miss to the downside. I don’t think it extrapolates forever, but we’re definitely playing on a bigger stage with bigger dollars, and there is at least some justification for being able to deploy those dollars.

We talk about bigger, fewer outcomes and the meaningful nature of them. The thing that I also found really encouraging was Melio, a company that is very good but not in the top 0.01%, being bought for $2.5 billion by Xero. I thought that was really encouraging to see—a slightly smaller, but still very meaningful, outcome. I’d love to hear: how did you guys think about this, and how did you break that one down?

I thought it was discouraging. What I mean is that Melio gets $2.5 billion. Look, that’s a lot of money. Hopefully folks will watch this and mock me for not thinking that’s a lot of money, but they’re at $153 million in ARR, growing 127%. At $153 million and growing 127%, if you knew nothing else, why would you advise your portfolio company to sell for $2.5 billion?

We all have deals like this. I just got pitched an AI startup at $2 million in ARR, and it’s only growing that fast, worth the same amount.

Harry, it’s not—I don’t think this is a big one. I could imagine what happened if you guys know the story, but this is not a deceleration story. This is crazy.

I’m not convinced that the growth rate was that high on a sustainable basis. I’d be surprised. Maybe it is, and the market might not be sustainable, but I’m literally reading Xero’s slides where they’re selling their own shareholders on it.

That’s what they claimed, right? $153 million in revenue, whether it’s growing 127% or 100%, is this not the greatest multiple of all time?

That intimidates me as an investor, actually, because I have plenty of deals that are great and not as good as Melio.

No, look, yes, it’s like 13 or 14 times, depending on NTM revenue. I thought it made a ton of sense. I had it pegged for a lower sustainable growth rate, and we were investors in Bill.com—very happy investors in Bill.com. It’s public, it’s much bigger, and the market’s been tough on it recently. It’s a tough comp today.

That kind of bill-payment and accounts-payable space is fairly crowded. There are a number of plays, a lot of go-to-market, and consolidation. Being acquired by an ERP-adjacent competitor made industrial sense to me. To me, I was like, “Yeah, that’s about the right price, and that all makes sense.”

Again, I’m disconnecting from the growth-rate comment. You have the advantage over me because I haven’t read the press release, and no one, I’m sure, would ever lie in a press release.

No, I don’t think Xero can lie in what they’re saying. They’re not showing the forward growth, so your point is probably correct. But on trailing velocity, it’s a force of nature. I think they had some interesting strategic deals that are always challenging at scale.

The other interesting comment about it just worries me for M&A. We all throw out, “They’ll buy my portfolio company for $900 million or $2 billion,” but are you sure you’re better than Melio? I’m not so sure.

I think the interesting thing about that one also was that they’d raised a couple of rounds significantly above that in value. In 2021, they raised 2 rounds, the last of which was $4.5 billion.

On the one hand, it shows what a lovely, rigged game the late-stage business is. If you always get a 1x on your losers and you have enough winners, by definition, you have a positive IRR. It’s a nice thing.

On the other hand, it just shows how wrong you can be. You sat there in 2021 thinking, “I should buy at $4.5 billion in the expectation of making a 2x or 3x,” which implies $13 billion plus. Then you fast-forward 3 or 4 years, and you’re happily taking $2 billion.

I mean, the preference stack here was $650 million—$650 million total preference stack—and then the last round was at $4 billion, led by GC. So, to your point about the preference stack-to-outcome size being the risk that you're taking, 100% here, you had pretty minimal risk when going in at that, I think.

Yeah, I agree. You're still getting a 0% IRR, but, yeah, I guess it's agreed: you're getting a 0% IRR and you lose again.

The interesting thing to me—I guess this is Captain Obvious—is that this is actually a little bit good for founders. I'm a little slower than the 2 of you, but at Melio, the last round was at $4 billion. Okay, $500 million in revenue or something like that, right? And then there was a corporate round at $2.5 billion.

But if you're getting your preference back and it's below your $4 billion, you don't really care what the price is. I mean, I almost killed myself as a founder fretting. I had to at least quintuple my investors' money, right? But if they're only getting 1x, they don't really care what the headline number is, do they? In a way, the pressure is off a little bit.

Well, they don't. But I thought you were going to say something different, which is that the earlier guys care. The early guys care and, more importantly, the founder cares. I mean, you know, the—

Yeah. But they're just going to make their own decision. It's binary, right? They're going to make their own decision. I think it's liberating for a founder to say, “Goodness, I can sell my company for $2.5 billion, and these bozos with the blazers aren't going to say no.”

Agreed. Right. That was the fear when I grew up as a kid, right? I was terrified my VCs would say no to everything. I was terrified every day.

Totally. I can tell you, if you're a founder stuck in that situation right now, no one who wrote a check in 2021 at $4 billion, if the company is underperforming, is going to stop a 1x right now. They'll be as grateful as could be: “Give me my money back and move on,” right?

So it doesn't matter if they did it at $40 billion or $4 billion; they're going to get the same outcome in Melio, right? They're going to get the same 1x back, which—separate comment—creates a large amount of weird dynamics on various boards as people wrestle with these prices.

But you're right: you're in the room considering it priced at $2 billion, and you've got some people who are blankly indifferent and just want to sell because nothing's going to impact them either way. Then you've got people like the early investors and the founders, for whom the difference between $1.5 billion and $2 billion is just a huge amount of money, all of which is going to them. And that's great, because the founders created the value here.

You say, “Give me the cash back.” The only thing I'll say to that is: yeah, give me the cash back—or even a discount as the cash back—but don't give me X billion of company stock at some ridiculous price.

I think I have more Airtable than anyone at $11 billion because they've acquired about 4 of my companies. It just shows a couple of things. One is, private-to-private is hard, right? And you always have this “Am I getting shafted?” feeling, because it's all notional.

Which, again, going back to first principles, is one of the beauties of being public. When you get bought by a public company, you don't have to have this existential debate—or, in your case, Harry, from our conversations prior, this existential whine—about how much or how little you value the stock. You just read The Wall Street Journal and know exactly what the damn thing is worth, right? It's another beauty, from a speed-of-acquisition perspective, being able to do public deals.

Rory, I'm sitting here in Frinton-on-Sea in Essex. Chamath is in Portofino. I have a reason to whine when my stock is given at a $15 billion price.

Yes, you do.

Okay, Jason. I really appreciate that, because it totally made me think differently about the Melio outcome in a way that I hadn't before, which I love. But $187 million in March, actually. So they could have been at $200 million—well over $200 million—when the deal got signed, right? $187 million in March. I can't get my head around why you do that then. But okay, fine.

Another thing I'm looking for—you can also look it up—at that $4 billion round, which was $500 million in revenue. I think it was $500 million. So, listen, I know what the Wiz slide says. My understanding is that a lot of the early guys sold very quickly. It was 2 years to getting out at $5 billion for the small guys, not for Bessemer. I think Bessemer might have funded it. I mean, that guy finds all the good Israeli ones, right?

My understanding is that maybe 9 figures of that quick 2021 round was out at $4 billion, right? Good for them. I mean, there's no Hopin secondary, but it's pretty good.

I was on a sales call this week. Some kid was yelling at me, Harry, and I was trying to explain to him something that I knew more about AI than he did. He's like, “Well, that's not how I did it at Hopin.”

I'm like, “What did you learn in the 3 months that that market existed?” Exactly when you went from infinite demand to zero demand, right? I mean, the easiest job in the world was Hopin in May 2020, right? Easiest job in the world. Hardest job in the world: May 2021.

I'm just going to say one thing, and I get in trouble for this, and then we can put a pin in this one. The founder always gets chastised for founder secondaries. I remember when he was being forced to take cash off the table by hungry growth investors who were diluting the hell out of him, and he was going from 49% to 42%, or whatever it was. It was a completely rational decision.

I think it's okay if the investors got to sell, too. Listen, you're not the founder, but my point is, if he's being FOMO'd, okay? And that does happen. It's happening again today. People are getting FOMO'd, right? For folks, it's when, even though the investors want you to be conservative, they want to put so much money in the hot company that the only way they can get it is by giving it to the founders—buying their stock.

But it sure would be nice if the investors had the same option. I had this happen to me just once, Harry, where there was a FOMO round and I said, “Listen, if you're out, I'm out.” This is what I said to him: “Whatever—I just want, proportionally, whatever you're going to sell. If you're going to sell more than $10 million, that's like you're making the right decision. I just want to sell, too.”

That makes sense to me. That was my only ask.

Me too.

Reversing my comment earlier, I would point out to you both that you do have that right. It's a co-sale right, right? Read yours, Harry.

That's not enough. You can't always get as much as you might think.

But you're right. You have a co-sale right, and it's relative to the sale. The reason it's there is for exactly this purpose.

You assume we have any rights, Rory? I think the challenge of 2021 is that we didn't have any rights, but typically there would have been a—

I mean, look, it comes up because, you know, can't they be waived by the majority of the investors, though?

Yes, they can, if you choose. I'm getting waived by some guy bigger on the cap table than me. I guarantee they don't even—The last deal I had, they didn't even tell me when my rights were waived.

You guys are both sounding a little punchy and a little bitter. Remember, no one's going to cry for either of you.

Not bitter, but good criticism. Actually, you're lucky to be able to invest, right?

Exactly right. Some stuff works, some stuff doesn't. Move on.

But I would like a super co-sale in that situation. That's all. If you're selling more than $20 million, I would just ask for a super. It's a genuine comment.

And it's funny you use the expression, the FOMOing, which I love, but it's odd. This is a FOMO process where frequently the person doing the FOMOing ends up dying because they—

Well, my point in the case of Hopin is that you FOMO'd the people, and yes, the company died, but you died too because you lost all your money. But you're right.

The thing that blows it up, Jason, you're right, is I love that expression for FOMOing: some late-stage investor who just has $100 million to put to work and is just going to do it no matter what. The question that I have is actually: do AI researchers have the same challenges that FOMO'd startups have? When you stuff them with $100 million, do they become less efficient and less focused? That's my question for Zach, that no one seems to be able to find out about 12 months from now.

Seriously, you give these guys $100 million and they go and buy a massive pad in Atherton. I don't know. I'm worried.

You all have the founders that bought the massive pads in Atherton. It's a real issue. The massive pad in Atherton is almost 100% correlated with a decline in revenue. It's almost a 1-to-1. I can't prove causation, but I can prove correlation. Almost 1-to-1.

Rory's just like, “I can't believe I'm with these 2 degenerates.”

No. Yeah, you are, in general. No, I mean, look, I don't think large amounts of money necessarily demotivate people across the board. I think what large amounts of money do is reveal what you really want to do, right?

That's why you see some people make lots of money in venture and decide what they really want to do is double down and keep doing this. And you see other people say, “I made large amounts of money. All along, I really wanted to be—I really wanted to tour the world and play golf.”

It just liberates you to do whatever it is you choose to do.

And thus, by definition, your real person comes out. I think the risk you’re running here is—you’re exactly right. It would be absurd to suggest that no one who gets $100 million won’t be a little bit disincentivized. So yes, there will definitely be some hit from that, right? I don’t think it’s a rule, a law of nature, is what I’m trying to say. I think it’s more a law that it doesn’t change people as much; it reveals what they really are.

Maybe some people have been coding for the last 5 years going, “God, I hate AI. I wish I could get out of this.” And here’s my $100 million, and I’m done. But in all fairness, if you think about it, this is tough stuff to get right. The incentives are so hard. Incentives are hard in anything.

I mean, Zuck made it work with Bret Taylor, right? He made it work with Kevin Systrom. Eventually, they flamed out, but those worked for a long time. He got his money’s worth out of those deals.

Yeah, right. So Zuck doesn’t need 10 years out of this team, right? He has a history. Maybe sometimes we look too much in the past, but he’s made it work before.

I think if it doesn’t work, frankly, it won’t be because XYZ engineer got $100 million and went weird. I think, as someone pointed out, it’ll be because some other engineers got really pissed off. It’ll be an internal disruption—the weirdness of different people and vastly different comp—but, B, the wider question is: will it be a great ROI anyway?

I’m not sure. It’s not clear to me that this money will yield—this is almost going to be heresy here—the amount of money we’re investing in AI right now, at $300 to $400 billion a year in capex, and then lumping in a bunch more salaries on top. Sometimes I do wonder: in the end, it’ll be magnificent, but will you get the near-term ROI? Will it be an economically rational decision when you look back 2 or 3 years from now? I don’t know.

I think it has to be an economically rational decision when you put it in proportion to market cap. I know all those arguments.

You’re right. I mean, the proportion-to-market-cap argument, which I’ve made as recently as last week—I get it. But exactly what you’re saying is, you’re redefining economically rational. There are 2 definitions of economically rational. One is that it will yield a positive NPV. That’s kind of a very boring, accountant-logical thing, and I’m questioning whether it will.

What you’re saying is some grand, theoretic, game-theory version of economically rational. In other words, I’ve got to show up at the party with GenAI, with AGI, otherwise I won’t get invited to the party. So I’m going to spend $60 billion to make sure I’m there.

As I say, it really feels like heresy even to say it. But at some point in the process, we will become overinvested because that’s just what humans do when faced with this kind of opportunity, right? Is it now? Is it 2 years from now? Was it last year? I don’t know.

But if you look at what this capex is doing to the balance sheets of the largest companies on the planet, what it’s doing to free cash flow, it’s just astonishing—the amount of investment that’s going on here.

Totally. I mean, you said it turns Warren Buffett’s idea of a dream business into a cash-incineration machine. Absolutely. The poor man would be like, “People, have you lost the plot? I only left 6 weeks ago. Goddamn you all. What are you doing?”

Okay, but we talked about getting incentives right. I think an interesting one from that perspective is Menlo raised another $1.5 billion—a record fundraise in venture. When I looked at that, I thought, “Amazing. For Menlo, fantastic.”

My question to you is, first, how do you think about that generally? Second, do you think LPs are more excited by incredible returns driven by Chime and some other historically great investments, or by Anthropic and a forward-looking AI lens that Menlo, I think, have pioneered well?

I think it’s good to have both. I think they’ve done a great job and deserve the money, of course, right? I think Chime is a big-ass return. All credit for that. It’s hard-dollar money on the table.

On top of that, I agree: Menlo has done an amazing job glomming on to Anthropic, doubling down on Anthropic, and telling a strong AI story. It totally makes sense. All credit to them. You win, you get the prize. That’s how America works.

Totally agree with that. Jason, anything to add there?

No. The only thing I would say—and listen, you guys have a broader LP base than I do—is that I think the LPs I talk to are acutely, maybe certainly even more than me, aware of the math we’ve discussed here.

They’re acutely aware of what it takes to return these size funds. They’re acutely aware of what their managers’ asks are and what the commitments are in terms of the size of returns they need to generate. I’ve certainly heard skeptics of whether these outcomes are there, but they understand where the market has to go to achieve the outcomes.

They know the math on the back of their hand. They’re participating in this, right? They’re participating in the game on the field. You want to sit out Index, Kleiner, and Sequoia after Figma? Probably you don’t. You’re not allowed to sit it out.

The math that you do, Rory—and you’re my favorite here—my LPs can do versions of that instantly. They have an encyclopedic memory of every company I’ve invested in, somehow, because they have a lot of managers and they know everything. They know the AI issues and the challenges, and they know the exact amounts of how all this capital has to play out.

They get the risks. They get that you need huge outcomes to make this work. It doesn’t work looking backwards.

Yeah, agreed. They know it. The only proof you can have that someone will be in the deals in the future—there are only 2 proofs. Either you have some kind of story for a new fund, or you can say they’ve been in the deals in the past.

You can look at Menlo and say Uber, Chime, and Anthropic. You can join the dots in those 3 sentences and say they probably will show up in the right place.

We mentioned the numbers needed to make it great. It’s an unsexy discussion topic, but it did actually, again, give me hope. And Jason, I hope you don’t dash my hopes on this one like you did with Menlo. Couchbase, a company most people haven’t heard of, was acquired for $1.5 billion.

Again, not massive, not huge, but still very important and meaningful. The question there is: will we see a plethora of these PE buyouts or PE-led buyouts at a similar scale, which will actually drive a meaningful amount of liquidity, or is this relatively one of few?

I’m hoping. I think—yeah, I don’t think it’s indicative of a trend. I took a look at it, and Couchbase was bought by Haveli, I think it’s the firm in Austin, which was founded by Brian Sheth, the former co-founder of Vista, who was an extraordinarily good investor at Vista and then spun out in 2020 after a split with Robert Smith.

He’s a talented guy. It looks like he’s building a very concentrated, thematic portfolio, kind of going out with a certain number. I get the impression he went out to find that asset because he has a perspective on what can be done in this space with that particular product.

I don’t think it’s, “Hey, every $200 million software infrastructure company with 16% growth is going to get hoovered up at 5.7x.” I think it’s much more thematic. I think this can be relevant in AI. They have a newer product that’s growing much more quickly.

So I think it’s someone taking a very rifle-shot bet, which may be right or wrong. We’ll see. But I don’t think it’s indicative. If you have 10 other $200 million database companies, don’t hold your breath waiting for PE to come along.

I remember, for example, we were in DataStax, which was a similar-size private company, sold to IBM. Great outcome for all concerned, very happy to get it done, but there wasn’t a plethora of PE buyers. It’s not typically a PE asset because they’re complex technical products.

You always wonder. The thing that encouraged me here was that it was $215 million in revenue, growing at 12%, so not stellar growth. I love how firms will make those bets with mediocre growth, bless them, and not be profitable. I’m encouraged that this is the one.

No, I agree, and that’s why I said I don’t think it’s typical. I think it was much more, “Hey, there’s a perception that this asset they have is of value.” We’ll see.

You’re exactly right, because every venture firm—including us, including everyone—has lots of $100 million, $200 million, and $300 million-revenue companies, subscale for the new world of IPOs and subscale in terms of growth, trying to figure out where they go. I’m not expecting PE to save all of us, put it that way.

This is such a small number. It’s barely useful other than anecdotally. But I have 2 portfolio companies that are kind of in that intersection where a tech company and a PE buyer might buy them, right? They both got mediocre M&A offers recently. Mediocre—not bad, not great.

The first thing I asked the founders was, “Well, what have the PE firms said to you?” No one had contacted them—ever. Even 2 years ago, it was like every week you were getting a call, right? “Who’s Driscoll Stebbins and Limited [?]?”

Neither of them had gotten any calls from PE firms. None. Zero. So that worries me. Like Rory said, these are ones that could go either way, right? They’re cash-flow neutral.

They have good growth. They have strategic and non-strategic value. You could mash them into somebody else. The point is, you can mash them into somebody else, right? You’d think you’d at least be getting the VP or the analyst calling them.

I’m surprised we haven’t seen more Bending Spoons-like models in SaaS. Bending Spoons obviously does consumer-subscription rollups for subscale outcomes for venture firms, but also for too-small outcomes for PE. I’m surprised we haven’t seen that more in venture.

I think you’re going to see a lot of it. You had mentioned Visma, which is a company that’s going public in the UK, and we’ll come back to that. But the aha prior to that is, it’s a roll-up of hundreds—literally hundreds—of subscale software companies. Constellation Software out of Canada does the same thing. It’s going to have to happen once the sellers are willing to take price, because all these assets can’t go on forever subscale.

The interesting thing about something like Constellation is, at scale, they’re interesting. A $100 million revenue business where [?] has nowhere to go, but if you assemble 20 of them and you’re at $2 billion and you have 10% EBITDA, it might be the sexiest business alive. The market will price it, it will value it, and you’ll be able to get liquidity.

But Constellation wants to pay 2x. That’s the slight.

Yes, yes. That’s their model: 2x, right? I know that. I had them present at SaaStr Annual this year. Their team came, and I think that was the title of their presentation: “2X.” Who wants 2X? Show up at 2 PM on the west lawn. Quite a few people came, which is maybe telling, right? They’ll do some deals at 3x, but 2x is the model: 2x revenue.

Agreed. They’ve been able to build a model whereby they can take all the risk out of the deal on their side. The question is, going back to even the Couchbase acquisition, where it’s 5.7x, is there something more sensible, a little more growthy, that you can do with these assets combined together, maybe with a more thematic approach than just Constellation, such that the clearing price can be 3x or 4x? I don’t know. But you’re right: someone’s going to have to figure this out, because these 300 or 400 companies aren’t going anywhere until somebody does.

I think someone who’s really well placed to do it is actually Grammarly. They’ve got—

No, just keep going.

I’m being serious. If you can build a next-generation productivity suite and do a number of great acquisitions, hopefully Superhuman being one of them, then maybe you can piece together something exciting.

Don’t jump out of your seat, boys.

Okay, fine. I’m hopeful. My Grammarly stock is going to the moon.

Good for you.

Well, listen, we can talk about it forever. I want to believe in Rory, but for as long as I’ve been in tech, I’ve heard this story: “It’s going to happen. Someone’s going to do this. More of these companies are going to get rolled up and mashed up.” VCs say it’s going to happen over lunch and cocktails, but I don’t see anybody stepping up. Just because it makes sense on paper doesn’t mean it happens in the real world, right?

True. But I don’t know if it’s stepping up. There are 2 people that have to step up: the buyer and the seller. From the buyer side, people have made money doing this throughout tech. Computer Associates, way back in the day in mainframe land, made money doing this. You’re right: you have Constellation, you have Platinum, and you have a bunch of others. There is willingness to do it on the buy side at the right price.

The question, to your point, Jason, is what is that price? Is that price one at which sellers are prepared to transact?

Don’t you think roughly everyone that hasn’t raised a round since 2021, that’s still doing okay and is north of 9 figures, is cool with selling for 5x? I think everyone is ready. They’ve already rationalized the 2021 valuation away, just like the Melio example. Everyone’s thinking, “It’s growing 18%. It’s a 5x. Let’s just sort the public companies—it’s a 5x deal.” I think they’ll take that deal. I don’t think anyone’s saying no after 4 years of this money sitting there making no interest in the bank. Are they?

I think on average you’re probably right at 5x or 6x. But no one’s buying all these 100 nine-figure B2B companies for 5x. No one’s rolling them up, right?

What you’re saying is, the market thought—or the owners thought—they were worth 10x to 20x-plus. Now they think they’re worth 5x, but the buyer still thinks they’re only worth 2x. So there’s still a gap. That worries me.

I think a lot of them would sell for 5x right now without arguing. I think they would sell for 5x.

Yeah. Well, as I always say, price clears all markets. We will find out. As you get into these older and older funds and as the growth rate becomes more locked in, your willingness to get realistic has to go up. As I mentally run through my portfolio and the things I’m involved with, there’s the stuff that’s growing at mid-teens, where I know why it’s growing at mid-teens and there’s a credible story of reacceleration. Then there’s stuff that’s growing at mid-teens and it’s never going to reaccelerate. You don’t know with 100% fidelity, but you should have some sense of what that is. For that latter category, you’re right: you should just seek an exit.

I just worry that AI changes so many markets that a lot of these candidates are viewed as hopeless now. It was one thing in 2021, when the software hadn’t changed in 7 or 8 years.

No, you’re right. But it’s now the second half of 2025. If you haven’t grown because of AI, you’ve failed. It’s not just that you have a copilot; you have to have grown. You have to have done an Intercom. You have to have reaccelerated your business. You had 18 months since ChatGPT launched, or whatever it was, to do this. If you didn’t get it done in those 18 or 20 months, it’s never going to happen.

I like your line, Jason. If you’re not reaccelerating, you’re losing, because someone else is accelerating. On a relative-market-share basis, you’re nowhere. If you don’t have that story and some of those facts, you’re nowhere.

I’d buy that. ICONIQ just put out this set of metrics showing how everybody’s growing across roughly 300 B2B companies. This isn’t causation, and it may not be correlation, but roughly 300 or 400 companies—including all the AI leaders, all the hot AI startups, and all the Harveys and Schmarveys and all the other ones—are growing exactly the same as they were 12 months ago, just about, when you throw them in.

We are stealing budget from each other. Sure, CIOs are adding an AI budget and this and that, but it’s clear it’s net zero. If you haven’t pulled yourself out by now, I don’t mean to agree with Sam Lessin, but you might be utterly irrelevant. It’s probably too late.

I think it consistently goes back to one of Rory’s most pressing statements: are we able to transition labor budgets into technology budgets in this next wave of AI? If we are, the markets are huge. If not, then it’s net zero. That’s how I read it.

The thing I think about is that being AI-native is such an advantage. We’re in a company similar to Superhuman that will be at more than their revenue in 9 months than it took them 8 years to reach. That’s no discredit to Superhuman; it’s just that the benefits and tailwinds that come from being AI-first versus layering it on are massive.

Another fun one—and maybe you guys know more about this than I do—is Clio, which just raised at $3 billion. It’s a 20-year-old company with a great CEO, Jack Newton. They just bought a 25-year-old company, vLex, out of Barcelona for $1 billion because it AI-ified its legal libraries. It did that by June 30 of this year. It got the message from ChatGPT, right? It became a data source at scale for legal.

I’m not a total expert, but in some minor ways it became a scale platform for legal. All of a sudden, it’s worth $1 billion. It’s an AI leader, but they got it done. Clio gave up a third of its market cap—or 25%, depending on how you do the math. That’s a big deal to do, isn’t it, for a Barcelona company founded in 2000?

And are you saying—just curious, Jason, because you’re not being clear—are you saying, “I totally get that. That’s magnificent. Well done, Clio”? Or are you saying, “I can say that vLex, a 25-year-old Spanish company, made the AI jump”?

They became AI-relevant. They had the time. They got it done, and now they’re worth something again. Harvey tried to buy them. Clio then bought them. They became a hot property, right? But if your other plain-vanilla B2B company hasn’t done it by June 30, I don’t think June 30, 2026—or that bridge note or that Series C7—is going to make it.

I probably would vote Harry and Rory down at the reserves meeting on that one.

But to be clear, what you're saying, which I think is this, is that there are examples of pre-GenAI companies being able to move quickly enough to become relevant in the GenAI world.

Yeah, I spoke too quickly. This one's amazing: vLex, founded in 2000 in Barcelona, could become AI-relevant today, right? I mean, it's impressive, right? At least superficially, it's very impressive.

The interesting thing will be—

I agree with that sometimes.

Yeah. But at the same time, we're wrestling, stepping back rather than pretending certainty, with whether the company that has scale can become relevant. In some cases it can. On the other hand, are you better off with the brand-new company that's less than 2 years old, that started off the day after ChatGPT was invented and doesn't know any other world?

Implicitly, Harry, you're implying the latter. You're saying it's just easier to start at ground zero, on November 22, and crank than it is in Jason's example, which is someone who was around beforehand and then successfully inserted themselves into relevance. I'll admit I don't have a doctrinaire answer here. We've seen both work, right? But it is probably one of the big investing questions whenever we look at a deal.

Mine's pretty simple: for the vast majority, I want the AI-native company, except for the exceptional product builder who genuinely is like a Des Traynor. I've interviewed the best product builders in the world. Des is one of the best of the freaking best. Him and Eoghan, I think, are so phenomenal at this transition.

But they got it done by June 30.

They got it done.

Yeah. So basically, whether it's Intercom from 2008 or vLex from 2000, they had until June 30. They got it done, right?

But the ones in our portfolio that haven't—they're still talking about it. I've given up on them. They had their time.

Got it. If you're still going to the board meeting arguing that you need to do something in AI, just stop going to the board meetings, right? It's no good.

He deserves credit as well, though—Howie at Airtable. TBD on whether it works, but what he's done in terms of the unbelievable shift in product strategy, very, very quickly, is impressive. Again, TBD on whether it works, but Eoghan and Des really burned the boats to make it happen.

Bravo, Howie.

Yeah, you have to do it. It's bold.

I just know—I will freely admit to biases, cognitive biases—and I just know that's the board meeting where you get up with a sinking feeling. For 3 or 4 years, you thought you had a home run on the existing thing, right? You could smell the money.

Now you're coming in and saying, “Oh my God, I'm in a $14 billion startup and I just want to throw up.” I just admire the courage of anyone who can do that. There's a lot of cognitive dissonance for about a day there, where you just go home and say, “Wow, that hurts. That hurts.”

Sorry, just to tie it back into one fun thing: if Oracle just closed a $30 billion-a-year deal with OpenAI, right—an extra $30 billion—I mean, if Oracle can get AI-native by June 30 and your portfolio startup can't, I'll smile, but I'd give up.

If Oracle can do it, founded in 1972, it could become AI-native-ish, right? And you can't. You're still working on it. Your team's still arguing. You're not sure about hallucinations. You're making fun of Fyxer—is that what it's called? You're saying that's not real, that's stupid, those emails don't really work. If you're doing that, Oracle went from $0 to $50 million in 12 months and then closed $30 billion with OpenAI. $30 billion a year.

Larry got the message, didn't he?

Yes. I'll say yes, he did. One version of the message is that he stopped brilliantly buying his stock back with his free cash flow and instead spent it on Nvidia GPUs. It seems to have worked because it's converted into cloud business for Oracle.

But he did it as early as Intercom and earlier than Airtable. He went all in. It's so funny if you look at the P&L. Cloud and PaaS is only a small percentage of the total revenues, and within that, AI is an even smaller percentage. It totally shows up in the capex budget because all the money is going out.

He's got this wonderful upside-down AI business inside a wildly profitable old-school business. But you're right, Jason. A, the market loves it because the market thinks it's leaning in. And B, if he can get this kind of revenue—and I'm going to say the sentence here in Debbie Downer fashion—and the investment boom in AI continues for 3 or 4 more years, then it'll be a great deal.

I'm looking at Oracle's stock price. This is the best stock I've ever seen, just watching Yahoo Finance or Google Finance. Absolutely. This is a dream stock.

Do you know what I also found interesting? I was reading this morning, before this: Surge AI, which is a Scale competitor, is now raising first-time money—$1 billion at a $15 billion valuation. We had Garrett on from Handshake, who said literally he's staying up night and day. I'm in Mercor, which is absolutely crushing. Turing reportedly raised a new round. Who knows?

There's a tailwind. Who's picking up the business of Scale? Does the Scale acquisition create an unbelievable net-new economy for these businesses—an influx of cash, do you think? Is it setting a benchmark?

Short answer: yes. Jason said it last time, literally the week or 2—I think the day or 2—after the deal was announced. The remaining asset is an empty husk. Maybe we're wrong. Maybe the new CEO figures out what to do with it.

Stop, Rory. Forget the end of that sentence. It's a sentence.

2 weeks ago, you could maybe have said, “Perhaps the new CEO of Scale AI will be able to figure out what to do with the business, build it back, get relationships with the other LLM companies, and continue to sell training data,” blah, blah, blah.

Since then, the 49% owner of Scale AI has spent their time stealing your best people and offering them gazillion-dollar packages. No one from any of the other high-end cloud providers is going to give Scale AI the time of day. Why would they? They're not going to give them confidential information. Why would they?

So, no, there's no business there, which by definition means all the business that was there has just gone to someone else. I assume $1 billion in revenue has been allocated out between Turing, Surge, Mercor, and whoever else got there.

Once that's reallocated, then you're back to equilibrium again, as it were. Then the bet remains the continued growth of capex spend on the AI boom, and that's what all these companies are riding.

Surge just did $1 billion last year, bootstrapped, right? Or pseudo-bootstrapped. Did everybody know about them except me?

Dude, no one knew about them. I didn't know about them.

Well, what excuse is there for venture capitalists and technologists to be talking about the $800 million one just because it raised all the money, and not know about the one that did $1 billion without any outside capital? Shame on everybody for not knowing about a billion-dollar competitor, and for everyone talking out of their ass on social media about Scale as if they even know what it does.

I'm sure you'll find that all the big firms had—I'm willing to bet the guys at Insight had a call placed to Surge once a week for the last 3 years.

I would hope so, because all the big—

Yeah, their website is badass. Look at it. It's nothing. This is a good company. “The quality of your data determines the ceiling of your ambitions.” That's it in 4 paragraphs. We're making so much money. We don't need anything on our page.

Look, and good for them. I assume the funding is genuine. I assume what happened here was that they had, like all these companies, a small profitable business 5 or 6 years ago, and then only 7 customers turned up and basically said to them, “We're each going to spend $100 million, or you don't need to spend anything on marketing. You only need 7 sales reps.”

It's like, why make a whole bunch of noise? Just put your head down, hire these people, and make some money. I assume that's what they've been doing.

Well, it looks like he's super smart. He was in core search at Google, ads at Twitter, research at Facebook, and founded this in 2020, right? He just quietly gets it to $1 billion.

Totally. Dude, that is the most Thrive Capital benchmark-style website I've ever seen. Look at this. Surge is doing $1 billion. It has a 1-page website which says, “What made Hemingway extraordinary? His life experiences. War, love, triumph, and loss.”

This is the most badass startup. Let's get him on the pod next week. War, love.

Can you track him down, Harry?

Dude, I'm going to track him down. I imagine he only goes by Edwin C., which is badass too, right? He goes by Edwin C., founder. He doesn't even say CEO. It doesn't say, “I'm the CEO.” And this is the greatest founder-story website of all time.

Good for him. I go back to first Twitter. Do your magic.

Yeah. When you only have 7 customers and they’re all desperate to spend money with you, you just don’t spend a lot of time wasting time doing sales and marketing. Good for him.

Listen, there are so many places we could take this. I want to do one more topic before we do a quickfire. We’ve got the LaunchDarkly CEO quitting to be CEO of Asana, Stanford layoffs with the 2.9% endowment tax, and Ben saying that about 50% of work is being done by AI. Where do you guys want to take it before we wrap up with the quickfire? Those are 3 different choices, to be clear. If there’s some mysterious parlor game where I was meant to have something in common between those 3, I want to put it out there: I failed.

Yeah, listen. I think they’re all interesting. We’re running out of time, but I do think this quitting is underdiscussed. A lot of folks are quitting—there are a lot of co-CEO and CEO resignations, and founder resignations. I mean, quitting LaunchDarkly to go to Asana is weird, man, but I guess it’s okay today after 2 years.

Why do you think so? For context, the LaunchDarkly CEO quit to be CEO of Asana. Asana is obviously a public company; LaunchDarkly is not. Asana is obviously much bigger than LaunchDarkly. So why do you think it’s weird, Jason, and how do you think about what it means more broadly?

It’s not weird. It’s just that turnover everywhere is accelerating. If you’re not in the elite of the elite, people are bailing faster and faster, and the capital is only going to the top companies. Crunchbase did an article this week: there’s no such thing as unicorns anymore because, forget about AI, only the money goes to $5 billion and up. It’s got to be 5-billion-corns. About 50% of all the unicorn money is north of $5 billion.

So it’s either those—even bigger than LaunchDarkly. Asana might not even be worth $5 billion today. Let’s look it up, but I bet—I mean, it’s probably above the line, $3 billion. Okay, so it’s tough. Everyone going to join Meta for $100 million is mercenary, and that’s the world we’re living in. At the same time, AI is going to lead to a lot of layoffs. I don’t think Mark meant to say 50% the way it sounded, but that’s what every CEO talks about behind closed doors at a public company. They say, “I need to reskill 20% to 30% of my people,” but they really think, “I might not even need half.”

I think there’s a lot in that. Let’s start with the turnover comment. There’s a lot even there. I do think one thing we’re seeing is founders looking at the fact that if you signed up for a 4- to 6-year journey and it turns into a 12- or 13-year journey, it’s not surprising. You are seeing people say, “This is more than I thought,” for a whole bunch of reasons: life reasons, exhaustion, whatever. So we’re definitely seeing founders saying, “Maybe I need to get a president. Maybe I need to not be the CEO.” It’s just a lot longer and harder than I think it looked when you took $500,000 from Y Combinator and said, “This is a fun journey.”

Fast-forward 10 years: that’s probably a third of your working life. If you’re still 5 years away from an exit, that’s half your working life. If it’s not your passion, it’s almost impossible to keep doing, and even if it is your passion, it gets hard. It gets hard. We’re definitely seeing that, and we’re trying to deal with that in lots of different ways. Sometimes you have to reincentivize founders; sometimes you have to accept that a transition is appropriate. Then you have to figure out whether you go through the pain and hassle of finding someone else or actually just say, “You should look for liquidity for the company.”

And then separately, Jason, you’re exactly right. There’s this weird kind of concentration-at-the-top effect, where only the winners seem to matter, or the payouts of the winners are becoming so skewed. I saw a tweet that said something like, “Just go from Instacart to OpenAI to make money.” That’s exactly right. That’s what happens when you’re in extreme wealth and extreme volatility. Even being 10th doesn’t look nearly as good as being 5th, and the differences are so stark at the outer edge of the curve that you’re seeing all sorts of weird deal-hopping and people-hopping, as you say.

I mean, Cursor is raising more money now, and I think it’s a reported $28 billion. That just goes to your point about the escape velocity that this 0.01% has, being so much greater than ever before. According to The Wall Street Journal, last year saw a record number of CEO departures—or the all-time high tracked from public companies. Overall, 2,221 CEOs quit last year, up 24%.

It is weird. It’s a little bit like someone did the really fun thing about the Ronaldo post, besides some dude who just got hired by Facebook and then got $100 million. It definitely is. First of all, it’s Revenge of the Nerds, which, on balance, I’m good with because no one ever confused me with Ronaldo.

More importantly, they made the point that anything that becomes that high-stakes maybe ends up like that. Are we going to end up like English soccer? Take any sport—take the PGA. You win the PGA, you get $5 million. If you’re 20th, 30th, or 50th in the PGA, you’re barely scraping along and covering your expenses. That’s what it’s like in these winner-take-most economies. It’s a weird, weird time.

I think what’s even more challenging about that for us at the earlier stages is that when you have that situation arise, the opportunity cost of not being in those winner-take-all deals becomes even higher. Your willingness to pay whatever to get in at your entry price becomes anything. Your $10 million for a seed for Andreessen or Insight, or any of the others—whatever, who gives a fuck? The opportunity cost of not being in those deals, with the outcome sizes being so much bigger, is so much higher.

I agree, and it raises the risk for you. Even at our stages, you can be a perfectly conscientious little seed investor, do 30 deals per fund, and just not be in the one that matters. If everything else doesn’t matter, then you don’t matter, right? That has all sorts of downstream implications in terms of how you should think about your portfolio and how you should think about fund size. It’s quite crazy. We lost a Series 8, actually, in the last 10 days. It was doing $5 million in ARR. It started at around $225 million and ended up at $600 million.

Eventually, it will overshoot, but along the way it’s really hard. We should do some quickfire, because I’m going to wrap up at whatever it is—30.

Okay, let’s do it. Will Elon Musk create a new political party this year? Yes gets you $230 if correct; no gets you $160 if correct.

Does it count if he creates a party?

He can create a party. The question is, will it poll, right? Quickfire: will he create the party?

I say no. I think he’s just—I empathize, in a sense. He’s too emotional about all this. He’s shooting from the hip. I’ve been there myself; I have the same flaws, so I get it. But it’s not going to happen because he knows it’s not going to work. He’s so smart. He knows it’s impossible. It’s impossible to win in the U.S.

I’ll go with no because there’s not an option that says, “I don’t care,” because that’s really what I think. I just don’t give a damn.

Okay. No, you just don’t care. Fine. You know this one, Rory—I’ve been looking forward to it. I found this one and I was like, “Jason’s going to have a banger.” Will Roy Lee, founder of Cluely, become a billionaire before 2029? Yes gets you $534; no gets you $109.

How do you value a paper?

By paper, we mean—just to be clear—it doesn’t count the same. We can cut a little slack for our $2 billion friends here, Jason. Is it worth $5 billion? Is that the question? By 2029?

Listen, I think he’s badass smart, but everyone’s a little full of shit about AI and revenue, right? If he claimed it was a $5 million run rate before Andreessen invested, let’s put some asterisks and daggers on that. Could someone do that deal at a $1 billion valuation? I think someone might. He owns half, so he’s worth half a billion already. Let’s assume he owns 25%.

How much does he own? This is starting to look more and more plausible the more we work through the math, doesn’t it? Wow. If he’s really at $5 billion, growing at this rate, and it doesn’t crater—200x, 300x AI deals—2029 seems plausible. We could probably move this to 2028. Wow. I’ll see you in 2029—on paper, on paper, on paper, right, Jason?

I mean, I would have bet no, but Jason’s convinced me. On paper, I’m not getting $100; I’m getting $100 to $109. If I say no, I’m only getting $9 for a win, so it’s barely worth it. But that’s a terrible bet if that’s the way it works. It’s a terrible bet.

I just want to be a contrarian in my heart. I think no. But I suppose this is proof that the story is, you know, pessimists sound smart and optimists make money. I'm going to join Jason in the optimist camp. I'm going to go with yes despite myself.

I'm going to go with yes, too.

Boys, this has been fantastic. Rory, thank you so much for joining us from the holidays. Jason, thank you so much for putting up with the late night. The Essex coast has never been sunnier. It's raining outside, which is why you don't go to Essex. Nobody—there are so many reasons why you don't go to Essex, but I'm willing to put weather in the top 5.