Rippling vs. Deel Lawsuit: WTF Happens Now? The Future of the Late Stage Private Market

Guest

Suddenly, we’ve drifted into some kind of theft of trade secrets, and suddenly someone opens an investigation. If my ass was on the line and I was the CEO of Deel, I’d be like, “How much money does it take to settle this thing by Friday?”

A fund returner is not enough, man. We don’t get out of bed for a fund returner. A fund returner just returns the fund. Everyone talks in venture about fund returners like they’re so great. I don’t think they’re so great. One of the depressing facts about venture is that we make an embarrassingly large percentage of our money once every 7 years, when you’re in the white heat of “must acquire, must own” high-growth venture assets.

1. Chime's IPO Announcement: Who Wins & Who Loses

Ready to go, chaps. It is the highlight of my week. Everyone knows my mother listens to this show, and she’s like, “I can visibly tell your tone is more excited. You should change that.” And I was like, “I think we’re past that after this amount of time.” Jason, this is good news, guys. IPOs, baby. This is liquidity. Thank God.

So Chime announced—well, dropped their S-1. I thought it was super interesting: 8.66 million active users. This one astounded me: 2/3 of users have Chime as a primary account. 2/3. That’s amazing in my eyes. $1.67 billion in 2024 revenue. Why are they IPOing now? Unless you have inside information, that’s, to me, the number-one question, right?

Guest

Despite the craziness of the last month and a half, we’re only 3% off our all-time high. This has been the weirdest market ever. We were doing great, then we plummeted a little bit in March—really in April—but this is the fastest bounce-back in the last 20 or 30 years.

We’re up about 17% or 18% in a matter of 2 or 3 weeks, so we’re within spitting distance of our all-time high and up on the year. What it says is that a month ago, everything was doomed. Now, everything’s back. What it just shows is that when policy changes that quickly, you really can’t triangulate on that.

I think they were smart. They had their S-1 on file, and what they said was, “Somewhere between when we filed privately and today, weird shit happened, but it appears to be over. Proceed as normal.” I think they’re exactly right. In today’s market, you’ll clearly get this done. In February’s market, you’ll clearly get this done. Oh, and in between, for about a month, things went to hell in a handbasket, but moving right along. Nothing to see here now, right?

That’s why they dropped it. I think they get points for being shrewd, keeping it on file, keeping it updated, and now, in this market, they’re ready to roll. I think it’s a great company. Stepping back, it’s clearly going to get done. It’s a $1.7 billion trailing-revenue company growing 30%.

2. The Lopphole That Means Chime Has a Better Business than JP Morgan

We’ll talk about valuation in a second, but it basically comes off 2 big ideas. The first is that, in the world of the internet, you can give people a bank account so cost-effectively that you can offer a very different product than the large US banks. You don’t have to ding them on overdraft fees. You don’t have to nickel-and-dime them on monthly fees. You can make pretty much 75% of your money on debit-card fees, which is a relatively small part of most banking revenue streams.

It’s a great idea, and the reason it works is, duh, the internet. No branches, no people in branches. They make $250 per client per year, and they can build a profitable business on that. I don’t think a JPMorgan or HSBC would be able to do so if all they were selling was just your debit-card transactions.

That’s the positive part of it. Over the next few years, they’re going to add loans, which they don’t really do a lot of today, and all the other cross-sell stuff that every financial-services company does at scale. You can look one way and say there’s a whole bunch of upside from here, because if you’re making $250 off a customer today—even if that’s a middle-income customer with sub-$100,000 income—they’re going to have other financial needs, and you’re going to be able to sell to them. That’s the upside story.

It is worth pointing out the minor negative on the story: a huge amount of this is Durbin Amendment arbitrage, which is a very arcane rule that really matters. Basically, in 2009, after the Great Financial Crisis, they were putting through a regulation package, and there’s a rule that says if you are a bank with more than $10 billion in assets, you can only charge approximately—the devil’s in the details—but approximately 50 bips on a debit card.

If you’re a sub-$10 billion bank, you can charge more, and the effective rate is typically about 1.2%. So you get more transaction revenue when 2 people go into the same shop and buy something on a debit card. If one is from JPMorgan, the shop has to pay 50 bips. If the other is from littlebank.co, they have to pay 1.2%.

The economics of a debit-card business to a small bank are much more compelling. Chime itself is not a bank, but they cleverly team with a lot of these small, sub-$10 billion deposit banks. As a result, their revenue stream—which is 75% of their total money—comes from a product where they have this kind of umbrella effect from that legislation.

Obviously, if that were to change, it would impact the economics. There’s no sign of it happening right now, but I’ve got to believe that if you’re Jamie Diamond, you wake up every morning spitting mad that these dudes are able to take your customers because you’re not allowed to charge what they can charge for exactly the same product.

That’s the only minor negative in the thing. Otherwise, it’s a great company. It’s growing nicely—30% plus. They’ve executed on a very consistent plan for 10 or 12 years, and you kind of go, “Team.” We can talk valuation in a second, but I like the company a lot. I wish we’d done the round we looked at, and all congrats to them.

What round did you look at?

Guest

Oh, way back. Way back. I can’t even remember. It was not any of these rounds.

Yeah, I heard that you wanted to do it, but your partners didn’t. You were all in, and they were like, “No.”

Guest

Yes. Good. Partnerships stick together, Harry. I love those tweets: “I would be a billionaire, but I couldn’t get it past my investment committee.” I was all in on the deal, and the founders wanted me, right? We already had a handshake deal. But let’s move on from that. It’s not that interesting.

Let’s talk about valuation, which is a more interesting thing.

Guest

Well, hold on—not to interrupt, Harry; it’s your show, but if it’s not interesting that Chime’s IPOing, then shouldn’t everyone—Chime or better—IPO now? We’ve talked about the incentives not to IPO, right, for founders to do infinite secondaries and do it. But if the markets are wide open, Figma’s out, Chime’s out. Is it time for folks to grow up? Should everybody IPO now? And will they?

3. Why Investors Who Invested at $25BN Will Make Money When it IPOs at $12BN

I’m going to override you because I didn’t say—I said talking about internal decisions on deals, not deals, wasn’t interesting. I actually think I’m going to hold your question, because I think the natural order of it is that we should talk a little bit about the valuation, which will segue us perfectly to talking about how other people should respond here, because obviously the big question here is valuation.

The last private round these guys did was at $25 billion. I saw The Information estimates of a valuation of $7 billion or $8 billion. I think that’s low. But I think if you end up plus or minus $10 billion—rough order of magnitude here—you are looking at a deal going public 50–55% below the last-round price.

That’s just a fact. You have to talk about what that means, how that works for the specific company, and then what it means, Jason, to your point, for other companies going public.

Can I ask a question here, because I’m just ignorant? If Sequoia Capital Global Equities—which I don’t think is the early-stage fund at Sequoia, based on the title—but if they did the round at $25 billion, what do you think? Is this a huge loss for Sequoia? Is it an SPV or a separate entity? Is it a small piece of the fund? Or if it goes out at $10 billion—which is epic in absolute terms—but Sequoia did it at $25 billion, what does that mean at the fund level?

Guest

Maybe Harry knows too, but I honestly don’t know. What does that mean at the fund level?

Well, just to be clear, besides not being a win, right? I actually interviewed Jeff Wang, who was the head of it, and he’s left now. He left a while ago because it’s such a great job. If it was such a great job, he’d be there today. I love Jeff.

Essentially, it’s their super-late-stage, borderline pre-IPO vehicle. Generally speaking, they’ve done phenomenally well, but it’s a completely separate vehicle run by a separate team.

Guest

Actually, this is the question that I was interested in, because I have more detail. It boils down to only 1 issue: what are the terms of the mandatory conversion in the articles of incorporation?

In other words, we sometimes reference the fact that in an M&A, even if you quote-unquote overpay, if this company sold in an M&A situation, even if you paid $25 billion pre, if you have a liquidation preference, you would get 1x your money back.

Rory O’Driscoll

The question is, what’s the equivalent term to that for an LP in an IPO? There is an equivalent term: it’s the mandatory conversion term, right? The question boils down to the following: in a qualifying IPO, which is an IPO of a certain size and scale—which, of course, this will be—is there price protection for the $25 billion round such that the price adjusts down, either fully or partially, to the IPO price or not? Is that term in there or not?

I went to the S-1. It wasn’t clear, but you can actually get the articles of incorporation, which is where it will be. I ran out of time. I’m actually very interested in that because, if you think about this late-stage business, there’s only one thing that can go wrong in a late-stage company. Probably 99% of them won’t blow up. The only risk you’re running is the risk that you overpay.

If you can negotiate a term that effectively says, “Hey, if I overpay, you’ve got to give me more shares such that I didn’t overpay,” then it’s the world’s best business. There’s only one thing that can go wrong, and now it can’t go wrong anymore, right? It’ll be really interesting to see what these guys did on this deal and, just in general, Jason, to your wider point on all these deals, what are the terms of the late-stage rounds in terms of IPO protections and blocking rights?

The one thing I do know is that I know the General Atlantic team very well, who were also part of this round. They also took part in the Shein round and led it at $100 billion, and I know that they are incredibly diligent about putting those protections in place in the case of a mispricing happening, like Shein, which is not hitting the $100 billion price that was paid.

Guest

Yes, and we’ll find out, because before you file the final S-1, it’ll be very clear. Let’s say it starts to be priced in the $10–$12 billion range. As part of the S-1, they’re going to have to disclose the adjustment and quantify the adjustment. It’ll all be there at the end.

The great thing about S-1s and going public is that all the facts come out, because otherwise the CFO goes to prison, right? We’ll know exactly how many shares get issued. If these guys have full protection, then that’s a win. You didn’t.

An interesting lesson for the founder: you didn’t raise money at $25 billion. You thought you raised money at $25 billion, but in fact, if you go public at $12 billion, you raised money at $12 billion and just didn’t know it. Harry Stebbings

The more I listen, the more I think Jason’s right that seed is for suckers.

Rory O’Driscoll

Well, not only is it for suckers, but I know a lot of seed folks who have all these great opinions on how SAFEs are terrible and everything’s terrible. I don’t think ratchets for a late-stage deal are such a bad deal. I don’t think it’s a big deal.

Obviously, if you’re an early-stage investor, you’d prefer there not to be a ratchet or an adjustment. Let’s be clear: we can’t argue with that. But if you need the money, totally, and you’re splitting the difference. Sequoia, SoftBank, Tiger Global, and Dragoneer are coming in at $25 billion. If you’re worth north of $25 billion, you win. You won the bet.

If Chime ends up at $10 billion and, let’s say, it’s a full ratchet and they’re ratcheted down to $10 billion, they probably only bought 3% of the company, right? So, you have 3% dilution because you lost the bet, but you still won the bet because you got the money, right?

I don’t know why people get so emotional about these ratchets. I understand why they’re toxic at the early stage, but these are just seed investors who are grumpy that seed’s a sucker bet, including me. I’m not grumpy about it anymore. I just sign the documents. I don’t even read them anymore, because it doesn’t matter what’s in them. I just sign them. It doesn’t matter what I think.

Guest

I’ve got to say, I find myself astonished at agreeing with you again, but you’re exactly right. I’ve been through the drama of one of these. We gave a late-stage ratchet, then you’re pricing the IPO and everything gets bent out of shape. I can’t remember one IPO recently that had one where people were writing, “Oh my God, they have to go public because of the ratchet.”

I ran the numbers, and the truth is, Jason’s exactly right: you end up saying, you gave away 3% of the company. You were in around, let’s just say, you were wrong by 50%, so you gave away 6%, not 3%. It sucks. I’d prefer to have 100% of my position, not 97% of my position, but it’s not the end of the world. It’s an economic term, not an emotional thing. Provided it’s not out of control, it’s survivable.

It does mean that, on the other side of the table, they have a wildly attractive business, because we’ve just agreed they get full price protection. They get liquidation preference in M&A. They get full price protection in an IPO. And, by the way, it’s priced at the IPO price.

The IPO pop puts them back up 30% the same day. That’s the little bit that’ll rub you as the CEO. “Let me get this straight. They paid $25 billion. It’s being marked down to $10 billion, so I’m giving them an extra 3%. Then, when my share opens in 2 hours and it pops 30%, they’re going to be up 30% on that recorrected price.” That sucks.

But, Rory, for sure, they made this investment in 2021. What’s the IRR on this deal? Not so great, right?

Rory O’Driscoll

Agreed. But that’s the point. Even if they have a full ratchet, and they get a 30% pop and distribute by 2027, that’s 6 years to having a modest return, right?

Guest

Yes, no, you’re exactly right. Totally agree. The risk you’re running on these kinds of transactions is primarily IRR risk, not loss-of-capital risk.

In the business you’re in, you have a substantial risk of loss of capital. My guess is you see 60% or more. At our stage, 30%–35% of our deals don’t work out, right? At the stage these guys are at, most of their deals—90% or more of their deals—when they’re writing those kinds of checks should be a 1x-plus IPO pop.

They’re not running the risk of getting it all wrong. What they are running, to your point, Rory, is, “Oh my God, we were 4 years too early and our IRR is going to be pitiful.”

Rory O’Driscoll

When do you think you transition to an IRR game? For us, we don’t really play the IRR risk game, so to speak. When does that become crucial?

Guest

Probably when you raise enough. It’s hard to answer that question. I think there’s a size. I remember realizing, when you watch the late-stage hedge fund guys come in, that they fundamentally run their entire life on IRR. They have yearly high-water marks and compensation schemes, and therefore they’re competing for a deal. They’re not saying to themselves, “I need a 2x or a 3x.” They literally use different language. They say, “I want to return 30% a year.”

You definitely see that with the late-stage hedge-fund guys crossing over. It’s probably in those last 3 or 4 years—what is now 3 or 4 years before the IPO—when you’re looking for that kind of return. It’s that kind of money, when the alternative use of your capital is public stocks. Maybe that’s a way to think about it.

Rory O’Driscoll

4. Are IPOs Dead & The Future of the Late Stage Private Market

When we look at this price that it could go out at being significantly lower than what we all said it was going to be, going to Jason’s question, does this mean that everyone should IPO? Now’s the time. Markets are receptive. Bite the bullet and go?

Guest

I think more people should. It’s worth pointing out that this is still a $1.7 billion-revenue company. It’ll probably be larger than 80% of the companies that are unicorns, and maybe 90% of the companies that have been talked about for IPO. It’s not like this is a mid-tier, marginal play. This is a top-of-the-line revenue-scale company.

I’m not going back to this and saying, “I have a company doing $200 million in revenue. Chime went public. The window’s open, guys. Let’s get ready.” These things happen incrementally over time.

There are a bunch of other late-stage companies who now clearly have, at the very minimum, a choice: I could easily go public. Do I want to or not? That set of decisions is, let’s call them, the billion-plus-revenue guys.

Interestingly, some of them, even if the window’s open, are choosing not to. There’s SpaceX and Stripe. They’re literally saying, “That’s not what we’re doing right now. Thank you very much.” Then there are others. Klarna, another example, is at that scale, where I think they’re saying, “Let’s push for the line, get the capital, and go for it.”

Oddly enough, both decisions make sense. If you’re a Klarna, if you’re a financial player where access to capital is really important, you are fundamentally a lender. I think being public and having access to money in all its different ways makes sense.

If you’re a high-growth AI company or something like Stripe, you have infinite private capital at dirt-cheap rates. Why would you bother?

Is that not fundamentally it? Respectfully, your Klarnas and your Chimes of the world cannot raise infinite amounts of capital at good terms from the private markets. Stripe, Databricks, Anthropic, and OpenAI can. That’s why those go public and the others don’t.

Yes. Klarna just said their growth substantially decelerated. They just published their numbers, right? My partner Paul did an analysis on it.

It’s 13%. He’s a phenomenal analyst: 13% growth. The cost of borrowing is way up. If Klarna was doing well—and broadly speaking, I think it is doing well—they should go public much sooner than OpenAI or Stripe because I just think they need to have access to continuous capital.

They’re a lender at scale. They’re not a cash-flow machine. They probably have more financing alternatives as a public company. I think it’s more appropriate for a financial-services company like that to be public than, say, an OpenAI or Anthropic.

And you’re right, there is cheaper capital available, too. Even if Klarna were, let’s say, growing at 20%, like Chime, or 26%, like Chime, there are more options for cheap private capital if you’ve got the sex appeal of OpenAI than if you don’t have the sex appeal of buy now, pay later, right? One of them just has more intrinsic ability to raise cheap private capital.

I think the interesting thing, to Jason’s point, is most companies are more like Klarna than OpenAI. No surprise, most companies aren’t singularities. And I do think, as the window opens, you will see the people doing $1bn in revenue, then $750m in revenue, and $500m in revenue thinking, “Maybe I should do this.”

Every one of them is going to have to wrestle with their version of: Did I give away price protection? Did I raise at a high price? Am I willing to take that kind of hit in the public markets? But I think if the window stays open, more of them will start to investigate this, and should.

Guest

Listen, you’ve said this a million times, Harry. I was at EF’s demo day a couple of weeks back in the US, and I saw one of my LPs there who had retired—one of the best. You know all the LPs; I don’t. I’ve had the same LP since inception. I’ve never added one, and I don’t plan to.

But this guy’s legendary in the industry, and he just retired, so he had no axe to grind or no game. He went through all my portfolio companies and others. He was like, “These guys just have to sell or go IPO. It’s just time.” And this is someone who pioneered a lot of this going long, you know.

So when I feel that vibe check, I think it’s got to water down to—or cascade down to—the GPs, which has got to cascade down to the portfolio companies: “Should I keep rolling the dice on anything sub-OpenAI or not?” If the LPs are saying that, for a variety of reasons, the GPs won’t keep tripling down, right? They will suggest IPOing, and it may be subtle, but when I hear that from one of the top 10 LPs of all time, that may push the pressure to IPO, too. It may trickle down to the CEOs.

He’s about to tell me that my question is stupid, Jason, so he’s going to reshape it in a minute. But you said they’re like, “Sell or IPO.” Thoma Bravo—or Orlando Bravo—is saying it’ll be a cold, quiet year for M&A, and then we’re looking at Convergence, a company that’s less than a year old in London, selling for 9 figures to Salesforce, and seeing more and more M&A.

I’m just confused. How do you guys think about those 2 opposing truths?

Rory O’Driscoll

People’s prognostications of what’s going to happen in the future are pretty worthless, including mine, to be clear. All you can say is what is happening right now.

What’s happening right now is Convergence. Salesforce bought a small, interesting AI company because they want to be an AI company. Duh. No surprise. There’ll be a whole ton of these over the next 2 to 3 years as these large software companies listen to Jason telling them they’re screwed on this podcast and decide, “I don’t want to be screwed. I want to be a contender.”

The best way to be a contender is to pick up some of these small acquisitions and fit them into your product. That’s clearly a trend that is happening. A fact-based statement: a much bigger one, Moveworks, and it’ll probably continue to happen, right?

Totally separate trend: What are Thoma Bravo and people like that making of software roll-ups? Are they going to buy a whole bunch of venture-backed portfolio companies? As we’ve talked about this before, I don’t think they are. I think they’ve got a fair amount of indigestion from the stuff they already have, and I don’t think the companies that venture makes are naturally a great candidate for PE purchase as much as people think. So his statement could be correct, too.

Guest

Yeah. I’ll say that the PE bummer still hangs over all of this. The fact that he doesn’t want to buy all of our portfolio companies is a big bummer. It is a big bummer on this Convergence thing.

I’ll tell you my view of what he said, just from my tiny lens. I’ve had 2 portfolio companies recently that got offers to buy them at $500m. In isolation, that sounds great, but these are very good companies. This is not OpenAI; these are very good companies.

One was just a smidge above the last round, and one was a smidge below the last round. What I mean from the Orlando thing is, listen, this sounds good, but these are tech leaders who want to make AI-adjacent deals, and they’re not willing to go all in. Maybe in 6 months those deals would be $1bn, and they both would have cleared at $1bn.

But they both said no, and the companies just walked. One of them bought a company instead for just under $100m because it was just easier. Instead, they bought someone at $2m for $100m. Great deal for the founders, right? It did raise a seed round, but they walked from buying a leader because $500m was the limit, right? And they all could afford infinity, right?

In this case, for all intents and purposes, that was—I wasn’t shocked, but I was like, I’m just watching the sign of the times.

But usually that accelerates, right, Rory? In a typical phase transition, what wasn’t clear to me was: were these acquirers PE-backed platforms or big-tech leaders? They offered to buy 2 different companies for $500m. Convergent evolution: the same number, but one was just above the last round—a little profit.

The other, they didn’t care about the last round. It was, “We don’t care.” And then they bought a much smaller competitor and will lose years because of it. They didn’t step up in the way I would expect—not that they have to; it was just a sign.

When you see that, to make a lot—for us all to make money in M&A—you need folks really stepping up in these deals, right? They’re like, “Rory did the last deal at $700m; I’m going to pay $2.1bn.” That’s the way venture works. If that doesn’t happen, like the Yammer and other deals, we don’t make any money if they don’t pay 3× the price around, guys. Otherwise, it kind of collapses venture a little bit, I think.

Rory O’Driscoll

Yeah. I mean, look, the truth is this: There are times in the market when the euphoria takes off and people are willing to lean in, and then there are times when the other side feels it has leverage and doesn’t want to do it as much.

Right now, in a few cases like Wiz, I think they created all the leverage. They played it perfectly, and then the other side did what they needed to do. They paid the big step-up from the last round.

I think in a lot of cases, all these people read the same press we read. They’re saying, “Venture guys, it’s a little bit tough. No one’s had liquidity,” and they’re probably in the mode of, “I don’t need to overpay.”

The only thing that changes that is if some of the companies that buy and buy successfully have success with those acquisitions. If you wake up 2 years from now, you’re competing with ServiceNow and the Moveworks acquisition is killing it, and you’re now 2nd or 3rd in the space—I don’t know who that would be; would it be Zendesk or whomever?—then by God, you’re going to do what you have to do, right?

5. Exits are Larger Than Ever: So What? What Happens? Who Wins? Who Loses?

You’re exactly right. One of the pressing facts about venture is we make an embarrassingly large percentage of our money once every 7 years, when you’re in the white heat of “must acquire, must own” high-growth venture assets. The trick in the other 6 years is surviving and keeping all the little companies alive and growing nicely so that when that moment comes, you have inventory to sell. That’s probably not this year.

Guest

Yeah, that’s my sense. It’s not—it’s getting there, right? Because these offers happen. $500m is not—I mean, we’re having a little bit of fun here. In any absolute sense, it’s an insane amount of money, right? But it’s not enough in venture, right?

They thought they had enough, or they were like, “Maybe…” There’s stress in the system.

How much ownership do you have in those 2 companies?

Guest

Let’s just average them to 10%.

Nice. Wow. Good for you, Jason.

Guest

No, dude. That, in your fund, is what—70% for these?

Yeah. Yeah.

Guest

But, dude, a 1× isn’t a fund returner. It’s not enough, man. What’s the point? We don’t get out of bed for a fund returner. A fund returner just returns the fund. Everyone talks in venture about fund returners like they’re so great. I don’t think they’re so great.

Well, it depends. Look, you’re in that business, though. I mean, obviously, the perspective of someone who has an $8bn fund, where they’re humble enough to recognize that the poor guy’s just going to have to chip away $2bn at a time.

Guest

Yeah. Poor guys. Yeah. I’m not into that vibe. It’s just not worth it.

But really, I don’t think a fund returner for seed is really enough. I mean, yeah, it’s the classic 1×, and then 2 deals do 0.5×, right? The rest do another, and you dribble and drab to the 3×, right? You dribble and drab, but I just don’t want to be in a world where that’s enough.

Yeah. I need a nice place in Marylebone. Is that where 20VC is? Is 20VC in Marylebone? Marylebone. Yeah. I want one of those carriage houses, and I want a nice one, you know? I want a nice one. That’s going to be my fourth house in Marylebone, down those cobbled streets with those carriage houses. You know what I’m talking about.

Well, then you should start a podcast, buddy. That’s one way to do it.

That’s one way to do it. Can I ask—well, I thought we were speaking about exit values. There was something fascinating: VCs and LPs, interestingly, basically did this analysis of exit values, and they found that in the 99th-percentile exit, so the top 1%, the price, or the value, grew from $1.4 billion in 2005 to 2009 to $10.2 billion in the most recent 5-year period.

Almost like a 5 or 6x increase in the top 1%’s exit value. I wanted to hear your take on this first, and, Harry, please slam the question in whatever way you feel is relevant.

What was my take on it? First of all, I thought it was great analysis, to the point where I actually emailed David and said, “Send me the underlying data,” which now gives me the advantage over you because I have it in front of me here, so I can keep you honest.

I love the way David put this down as mine, and you knew that I didn’t do it—absolutely, because it was mathematically correct, so I knew it wasn’t Harry.

Look, I think it was great analysis, really profoundly great. The question is—and what it’s being used to do is try to hypothesize: If this trend continues, how big will exits be in 2, 3, 4, or 5 years, and thus, what’s the ability of venture to raise ever-larger aggregate amounts and still make the math work? That’s the embedded question in this, right?

I think, unfortunately, the answer is one of degree. It definitely points to exits getting larger, with a small number of winners getting larger, though the trend is not as pronounced as you think. The first period of time was 2000 to 2004, where the 99th-percentile exit was as high as $3.3 billion. In other words, they went down. You had a cyclical downturn for 10 years before they started going up in 2015 to 2019 and then exploding up to $10 billion in 2020 to 2024.

It’s not linear growth for all time. It’s a dip and then a growth back, so that’s the first comment. It’s not as clear a trend as you think. But I do believe at some macro level it’s correct, and it’s simple: It’s not that things are getting better; it’s just that the longer you hold the company, the more compounding takes place, the more dispersion takes place, the big get bigger, and the shittier ones are crap. It’s just math.

Therefore, by definition, if the window to stay private stays longer, the size of the largest exit will be higher. I have no doubt. The largest single exit in this database in 2020 to 2024—the largest single outcome—was actually 2 at $65 billion or above. If there aren’t 4 exits above $65 billion in the next 5 or 7 years, then the people who bought Stripe, SpaceX, Databricks, OpenAI, and Anthropic are screwed.

I don’t think they’re screwed, so I think this trend is going to continue over the next 4 to 5 years. There’s no doubt in my mind, and it’s not for magical reasons. It’s just because you’re holding longer.

Let me give you another example of that. I looked at the historical data, and the biggest single exit in the period 2000 to 2004 was a single—you called it the 99th percentile. I don’t think that’s a useful term. The more useful point is that it was the largest single exit because there was only 1 in that group. It was a $23 billion exit. It was Google in 2004.

If Google had stayed private 1 year longer, Google’s market cap at the end of 2005 was about $140 billion. If they had just stayed private another year, the entire data would be swamped by the fact that the biggest venture exit ever was $140 billion in 2005.

Simply because if the Google CFO had had a heart attack in 2004 and they postponed their IPO for a year and a half, then the largest exit would have been in 2005, and it would have been Google. The point here is that all of this is a derivative of small numbers and how long the very best companies stay private. That’s all that’s going on here, right? It’s true, and it’s a thing.

What it means is that the bigger your fund, the more imperative it is that you have to be in those 6 deals, which explains why capital is so easy to raise for those companies. It all makes sense. It was great analysis. I’m not sure it points to everybody being able to do great in venture because everything’s going great. I think it points to the top end of a power law, where it really matters to be in probably 5 or 6 companies at almost any price.

But if the top end of the power law is so much larger than it was previously—which I think we’ll all agree it will be in 10 years’ time, in 2035—are we wrongly negative on the size of Lightspeed, General Catalyst, and your mega-funds because we’re considering today’s exit size when we should consider it as a 10-years-out exercise?

I’m not negative. No, I’m not. I think those firms have a great business. I think that if you can get your capital into those companies, then you’re going to do great. I’m not sure it supports everyone being able to do that, though, and the scale of late-stage money relative to the opportunity is much more nuanced.

The direction of travel is clear. As long as companies stay private longer, there are more opportunities—not just at the 99th percentile, but at the 99.9th percentile, at the very tippy top, with 1 or 2 deals per decade to be in them and compound for a long period of time. That’s definitely true.

Does that translate into all the funds making enough return on all the deals to make late-stage money work? That’s not as clear.

I totally agree. My concern is just that there are so few companies in that 99.9th percentile. It’s a world of concentration unlike any that I’ve—

You’re exactly right. So, Harry, how do you—maybe there isn’t anything in the study that answers this—but how do you know in a unicorn whether you should keep in or sell? Where’s the line?

Obviously, it always looks so clear in retrospect. You obviously shouldn’t have sold any SpaceX ever. You shouldn’t have sold any of the others. So the question is, how do you know at the time? I think it really is a function of market size and momentum.

It’s the classic problem. If you’re presented with an exit opportunity at $5 billion—framing it clearly, let’s say all your unicorns get to $5 billion—and the great secondary guys come and say, “You can cash out anything you want now,” statistically, let’s say the last round was around $9 billion.

Statistically, 80% of the time—I mean, Mary Meeker used to do this analysis of IPOs, which was excellent—and because I believe private late-stage companies in 2025 are just the same asset class as IPOs in 1995 to 2005, the same analysis applies. Most companies barely beat their IPO price a year later, or ever again, and a small number of companies compound and do amazing things.

How does that become actionable if you’re sitting there and you can sell all your private companies at $5 billion, or at the last-round price? Statistically, 80% of the time you should sell, right? Because that’s what the stats say. But 20% of the time will cover everything.

The interesting thing is that those 20% will not just make money themselves; they will cover everything else. The wonderful thing about this business is that compounding is a very forgiving thing. So, obviously, if you’re smart enough to be able to tell the 1 good one from the 4 bad ones, sell the shitty ones, keep the good one, and you’ll be rich.

But if you can’t do that, it is a matter of mathematical truth that the second-best alternative is holding them all, provided you have 1 of the good ones in there. I think what’s going on with all these late-stage funds is some version of holding them all. If I do enough, and I’m in the good ones, and I double down on the good one, the long-term trend toward a massive tail in the power law is going to make me money.

My new rule is—and listen, you’re lucky to be there, right?—when I was looking at this on a spreadsheet, my new rule was: at $2 billion, sell unless you’re 100% sure you shouldn’t. As a seed manager, at $2 billion, I know it sounds goofy, but it sort of ties to doing better than returning the fund.

The risk is that unless you’re sure it’s a SpaceX, being in that 80% is not so great, is it?

Again, you’re exactly right. The trends toward private for longer have made it harder. The truth is, it’s a hard comment, but they’ve made it harder for most investors and most funds because you’re exactly right.

Now you have to figure out a $2 billion valuation. You have to make these choices, whereas before you got liquidity on them all. Now you’re still in private land, and you’ve got to try and figure it out. You’re having to make those decisions with a smaller portfolio count.

The truth is, if you need 20 deals when you’re starting at $1 billion to compound to $100 billion at your stage, Harry, the portfolio count you need is much higher. The way that manifests itself is that if you constructed your portfolio 10 years ago assuming you got your exits at $100 million to $200 million, the risky thing is that now you still have to double down 1, 2, or 3 more years from here.

And as you say that, the bad outcome is you have the 80% but not the 20%, and you don't have the compounder that forgives all sins. I think it's pretty smart. That's why the whole push toward taking money off the table as a secondary is just smart. I don't think you can take the risk of doubling down ad nauseam when you're 10 or 12 years in. It sucks, but there you are.

Yeah. I remember back in the day—I’m dating myself—but I was at Emergence with Peter Gassner at Veeva, right? You might remember it better, but I saw a little bit of it happening, and their LPs were very mad when they held, right? Because they were the only real investor in Veeva, right? They owned 30% at IPO. I was just looking it up: it was worth $2.4 billion at IPO, which was a lot of money back then.

So, let the partners hold, but distribute to the LPs. They’re going to get $750 million, and it was probably a $250 million fund. We could look it up, right? It’s a multiple-fund returner on that one, right? When I was just starting, we had some LPs coming, and they were kind of mad at what you guys held. But today, even with some volatility, it’s worth $40 billion.

Agreed. $25 billion, right?

It created billionaires out of that—out of GPs. Created billionaires by holding.

But how do you know?

I knew Peter was the best one out of our class, by the way—our batch. So maybe you can know. David Sacks was 10 times better than me, right? There was René Lacerte and others, but Peter was [expletive] off the planet in terms of quality as a CEO. But I didn’t have the numbers, right?

But they made the bet, didn’t they?

Yes, they did. And remember that the thing they had to their advantage—or disadvantage—was that when you’re public, you can make that decision. You can distribute and allow different people to make different decisions. That’s the beauty of the public markets.

If Veeva had compounded as a private company, it would have just been—I mean, the company would still have been the same, but those choices would have been harder to make and different. It wouldn’t have been as easy; there’s no distribution as a concept. Interestingly enough, I hadn’t processed this until real time: the incentive—and this is where I go back to my monothematic theme of the death of IPOs—is just bad news all around for capital allocation.

Now, as a GP, you’re sitting there going, “I want to hold this thing forever.” My LP would probably like to get some liquidity. If it was public, I’d distribute and I’d keep mine. They’d sell theirs. Everybody would be happy because everyone can make a choice, and choice leads to optimal outcomes.

Because we can’t go public, I either have to sell now, which maybe is not what I want, or I have to ride it out for the next 5 years, which is maybe not what my LP wants. I think that’s inducing some tension in the system, which would go away if these things were public.

Taxes alone can be an incentive for a GP to hold.

Yes, taxes alone. Most of our LPs don’t pay any taxes, right? I mean, there are many other reasons, right? But taxes alone, especially if you don’t have QSBS, you’re like, “Huh, I have to pay 50% in San Francisco. Maybe I’ll hold for another year. See what happens.”

And again, that’s the beauty of it: I can get to Puerto Rico.

What? You have 50% capital gains?

Well, if it’s short-term capital gains, right? It’s still going to be—even with long-term, you’re going to pay 15% in California, 22%. You’re still going to pay 40% in California on long-term capital gains.

Well, this is what I find nuts, though. Look, everyone—some of the folks you’ve interviewed on 20VC have moved to Miami. Half your capital gains.

What are you saying? We have half your capital gains.

6. Is Europe Totally F

Yeah, but Harry, that’s a very true statement. You have half our capital gains. In fact, you have less than 10% of our capital gains. I don’t know if you saw the Wall Street Journal today, just dissing on Europe. You don’t have any capital gains, so it doesn’t matter. The taxation rate is purely notional, for entertainment only. I’m being mean.

Sometimes I’m surprised there isn’t more capital flight. The US makes it pretty hard to leave the country, and it’s a pretty tough country to leave, but I’m surprised there aren’t more people in Miami than there are. I’m really surprised there aren’t more, because, Harry, you save 15% just by moving over to Miami. There are some asterisks and daggers with it, but, yeah.

I’ve clarified my internal situation. If I were to reduce my income tax by 20% or 25% by moving to a nontax-based state, I would also reduce my net worth by 50% because my wife would be staying behind. So it’s just not an option for me.

I am at peace with paying whatever Gavin Newsom needs to keep this kind of bloated, overpaid show on the road. It’s a great place to live. So, genuine comment: not that I wish it were lower, but tax is not the reason to leave California. We should all be so lucky to have capital gains.

But the serious comment you made, Luke, is that the beauty of public markets is that they allow everyone to make their own choices about their own economic decisions. They can hold if they want to build wealth, or they can sell and pay taxes if they want to sell. It’s a lot harder to do that in the private markets.

I do believe we will look back and say there are reasons why the very best companies choose not to go public, but it’s a darn shame that the public markets haven’t addressed those concerns such that all this stuff could be done in the public markets.

7. Challenges of Going Public & What Needs to Change?

What is the number 1 thing that you would change if you ran the public markets to make them more appetizing for companies to go public?

I love the idea of time-based voting that they floated. I think it’s the Texas Stock Exchange—don’t quote me—that’s trying to get together an exchange with some of the leading companies and startups and some other folks. I don’t remember the people, but the idea that your share weighting is, in part, based on how long you’ve been an investor in the company is interesting. It’s an extended version of founder voting.

What happens in a public company is that sometimes all the arbitrageurs and all the short-term investors pile in, and they really push the company to make short-term decisions. But if your vote were partly predicated on how long you’d been an owner of the company, I think that could lead to very different results. That’s one random comment.

I don’t know how to get away from the randomness and the noise level around quarterly calls and all that process. I’ve never found a way. Google, for a long while, did that by simply not doing them, which may be one approach. I wish you could reduce just the anxiety and tension of that part of being public and always being on display. I don’t know how to do that part of it. I do think forcing longer-term holders is part of it.

I wonder—are you sure? I know it will be over. I mean, I’m not an expert. Sometimes I wonder if it’s really so broken, right?

I’ll give you an example. Some of our favorite companies from your portfolio—Aaron Levie, he’s all in, right? But he’s certainly spoken of the headaches of activist shareholders and all that at Box, right? Huge headache for him. Huge headache.

Then I was interviewing Brian Halligan a little while ago. He has some perspective on this now that he’s chairman. He’s like, “Honestly, it’s not much more work being a public company, doing this, than it was being private.” It’s not that much more work. But HubSpot’s a $30 billion company.

So maybe the bar should just be high. If you hit your numbers, if you grow 50% at $500 million in revenue, it’s not really much of a huge, huge deal. You have a huge finance team. It’s not the end of the world to go public, is it?

No, it’s not. But the odd thing you have—so, a couple of comments. One is that the companies that do precisely the best are precisely the ones who are in a position not to do it at all, right? You have the Stripe situation.

So it’s the companies who want access to capital at the $200 million, $300 million, $400 million level for whom it’s still a relatively big burden. But I do agree your point is actually the right one, not mine, frankly, which is that even though it’s a bit of a pain in the ass, I do wish you could deal with things like activism, silly regulations around boards and board composition, and all that.

The real point is that people respond to economic stimuli. Going back to the same thing we say over and over again: if capital were more expensive in the private markets than in the public markets, then most CEOs would go to the public markets, right?

The core reason it works is because there’s a lot of capital available in private markets for companies doing $200 million or $300 million in revenue, with a lot less hassle than getting that same capital on the public side. And that’s the reason they do it. People respond to price signals. CEOs respond to cost-of-capital signals. And there’s no doubt that, bizarrely enough, the cost of capital in the private markets remains cheaper than in the public markets.

Now, it's interesting when you look back—as a random comment—at the Chime $25 billion valuation. An objective fact is this: the cost of that capital was twice as high as you thought at the time because you didn't give away 4%; you gave away 8%.

8. OpenAI's Future and Predictions

Right? So, it may well be that we're in this little bubble where we actually don't know the cost of capital for some of these late-stage rounds. If you get highly priced rounds with lots of price protection and you ultimately go public, you may, in fact, discover that the last couple of rounds were way more expensive than you thought. I don't think it'll ever go back to $100 million IPOs, but I think there'll be more of a normalization in your choices between public and private.

I want to finish today, if that's okay, with one final segment. My team loves Kalshi, a predictions marketplace where you place bets in the real world, and I chose 3 that I liked. Number 1: Will OpenAI stop being a nonprofit, yes or no?

I mean, man, I wish I knew. Listen, the news cycle is so fast, as you point out. What a loss for Sam versus Elon Musk in the short term, right? What a loss, man. We already forgot about it, but I think it's a total capitulation to the idea of ever being a traditional for-profit company, right?

I'm going to go with yes because the question is weakly phrased enough that I can answer yes. Will it stop being a nonprofit? It doesn't give a time. At some point, this company is going to go public. It's going to have a PBC-type structure. It will get there. It'll take a lot of lawyers, but it will get there. So, I bet yes on that one because there's no timing.

I've never seen a dysfunctional company that's more successful than OpenAI. All the founders left, they fired the CEO and brought him back, and it went from nonprofit to for-profit, for-profit to public benefit corporation. The momentum is crazy.

I'm not a nonprofit guy, and I don't mean to crap on them, but the motivations are really weird in a nonprofit, right? As for giving that up, the folks I've seen on nonprofit boards aren't going to give up this power. No one I've seen on a nonprofit board wants to give up the power because there's no money in it. So, it's all about the power.

I'm going to disagree a little bit on that because I think the motivations in terms of being a nonprofit were much more important for the engineering staff and the early employees, who really had a profound belief that they were doing something important for mankind. I might not share that belief in the slightest, to be clear, but I think that was an animating factor in attracting the very best intelligence into this business early on.

I think the “not-for-profit” halo mattered, and it's no accident, in my view, that the 2 companies that have been most successful—OpenAI and Anthropic—embraced that because they recognized that the most important audience for both of them was talented AI engineers. All of them shared the religion that said this thing could change the world and be dangerous, so therefore they embraced the religion too.

I don't think today—I don't think it's a question of the board of OpenAI not wanting to give it up. I think Brett Taylor is just such a smart dude. I think it's a question of untangling the mess. When you've got litigation on every side, Elon busting your chops, a bunch of state attorneys general, and you kind of know where you want to go, it's going to be hard to get there.

In the end, the value of the asset is so high that there'll be some half-assed, cobbled compromise whereby the entity will be a PBC, the not-for-profit will be 1 level up, and the only question is how much they get. How much will Microsoft get? How much will the investors get? Somehow, they'll figure it out. So, I think they'll get there. It'll just be a wild and wacky journey.

Okay, the second one, and we're going to cap it at, like, a 1-minute response max: ChatGPT-5 revealed this year, yes or no?

The folks at OpenAI—the engineering talent is so much better than you realize. It's so next-level, right? The talent that OpenAI, Anthropic, Cursor, and Windsurf attract is epic. If they want to merge all their models into 1 model, which would make my life easier because I can't even tell them apart, that would be great, right?

At a consumer level, it makes no sense, right? But, man, deferring any feature requests into the backlog. Merging all of these—is it core enough to happen? Maybe there's a reason it hasn't been announced, right?

It will happen. But given all the activity, it wouldn't surprise me if it pushes a year or longer, right? But they have the best in the world. The best. We underestimate the engineering talent.

So, you're saying no.

I'm saying I think no, but it would be great.

9. Rippling vs. Deel Lawsuit: Is Deel Screwed?

Okay, you'll know. I'll take you on that one. I'll say yes. The velocity? Yes, and a don't know. Jason, baby, you want o3 minis and maxis? I'll give you Rippling and Deel as the final.

Will Rippling beat Deel in the lawsuit?

100% yes. There's no chance they'll lose. It could get settled, right, which is always the right outcome. Even when there's a motion, the right outcome is always to settle it, right?

There's no way they lose. 100%. The facts are too bad. They stole trade secrets. This is a classic case. They're going to lose.

So, Jason, you're going to lose. Rory, you're going to lose. Okay, why am I going to lose?

I do agree, Jason. I think Rippling prevails in this lawsuit if it goes to court. I think it probably should settle because most civil litigation does settle. But it's hard to imagine, from the facts as stated and the kind of witness statements—I did see the counterclaim just recently—how they win.

Did you think the counterclaim was weak?

It was some version of, “You guys did it too,” right? The first filing that Deel made back was pretty blah and weird. It was about 4 or 5 weeks ago, and it was full of a lot of weird stuff: “You went to Harvard,” or “You were angry about something.” I didn't quite understand it.

This is a counterclaim that said, “The guy that we hired—I think you guys hired him.” It was all very, “You did it too.” But I just think I'm with Jason. I think the facts—and you have the person who made his affidavit, and, by the way, you've chosen to run to another country and hide—I just think the fact pattern looks crap.

At some point, sense prevails and you say, “Whoopsie, sorry,” and settle. So, yes, I would buy the yes on this one.

The other thing, also, for what it's worth—I don't know how it works in the UK, but in the US, counterclaims are not what they look like, just as an FYI. Why do they do it? Because, listen, I'm not a litigator, but I've been on the other side. If you do tech long enough, you're going to be on both sides of these, right?

Counterclaims can offset any claims, even if they're outside of the statute of limitations, even if they couldn't be brought on their own. So, let's say Rippling wins a billion dollars against Deel, okay? Or a lot of damages. They could get, even if they couldn't bring their own lawsuit, an offset for $900 million.

There are so many incentives to bring counterclaims that you would not bring as a claim. So, there's always a lot of drama, and some of them won't even make sense. Some of the counterclaims don't make sense, right? They're not doing it because they think they're going to win. They're doing it because they know they're going to lose.

So, you put everything in a counterclaim because everything counts as an offset, right? It's a sign of losing, these counterclaims. It's not a sign of winning. And it's really stressful when you're on the other side and you get 10,000 counterclaims back because there's always some truth in it, right? There's always some truth in it, but it's a litigation game to get offsets.

It's a sign, Harry, that they're going to lose. All those counterclaims, ironically, are a sign they're going to lose. If you're 100% in the clear, you just say, “Not guilty.” You ignore it and, as a CEO, you go back to work.

You don't flee to other countries. You go back to work and you say, “Parker, you do what Sam's doing with Elon. Sorry we misunderstood each other, Parker. Happy to have a beer and talk it out.” That's what an innocent CEO says: “Let's talk about it.”

Let's talk about it, wherever it's at. Harry's dying to come in. Let's hear it, Harry. Let's hear it.

As Alex is a dear friend, and I'm also a Deel shareholder, I would just like to add that Alex has actually been abroad for many years. The media amplified him being somewhere where he's been for years.

Yes, fair point. But your point is right. What you do is say, “Plane ticket to Dublin.” He probably wouldn't take it even if it was first-class.

Okay, dude. He ain't coming to Dublin to testify on this puppy. So, thank you. No, because no one's going to Dublin, Rory.

No, but if you're going to win—if, in all seriousness, if Deel's going to win—you do what Sam did to Elon. You say to Parker, “I know we disagree. Sorry, as friends. Let's get together. We used to be partners. They used to be integrated. They used to be partners. We'll work together again. Sorry we disagree. Sorry, things happened that shouldn't have happened. Let's talk about it.”

And you don't say anything else, but that's what you say if there's really nothing there. That's the power play.

That’s the power play. What Sam did to Elon is—I’d go further. It’s what you say if there’s nothing there, and it’s what you say if it’s there as well, because you settled this thing. Nothing good is going to come from this kind of litigation, and I checked it—not now, but a few weeks ago, because we thought we were going to talk about this. All of this is just civil litigation.

But the scary thing is that some of the allegations could be interpreted in a criminal fashion. If I was the CEO of Deel, I would want to get this behind me so fast my head would hurt. I’d want to settle and bury it deep and say, “Whoopsie, sorry. Donation to the charity, your choice, whatever.” I would want this done.

Look, I don’t know if this is going to escalate to criminal, because one of the things that’s very funny when you see the difference between civil and criminal is, in civil, people yell at each other. I think criminal authorities—the FBI, etc.—are way more jaundiced and hard-nosed, and they’re just like, “We’re not going to get involved just because you kiddies are fighting with each other.” They’re only going to get involved if they see malfeasance.

But the more noise you make and the more yelling you do, the more risk you have that someone looks at the file and says, “Let me think again. Yeah, it’s in Ireland, but both of these are U.S. companies. There was an allegation here. That would be an espionage issue. Suddenly, we’ve drifted into some kind of theft of trade secrets, and suddenly someone opens an investigation. Then you’re fucked.”

If my ass was on the line and I was the CEO of Deel, I’d be like, “How much money does it take to settle this thing by Friday?”

Rory O’Driscoll

Absolutely. Get this done. Get it done. Popcorn out, boys. I know, Alex, this is not going to happen.

So that’s right. Hey, look, it’s a free country. People do what they’ve got to do. To all founders out there: settle everything. Especially when you’re not in the wrong, settle it, right? Because it’s so hard, if you’re not in the wrong, to settle it, right? It feels so wrong. That’s the number 1 reason to settle when you’re not in the wrong.

But I will say this: I have 2 observations. One is that everyone gets really emotional about litigation and starts getting personally invested in it. So that’s the first thing. But then the second thing is even more important: the lawyer that you engage will tell you you’ve got a great case when you start out.

As you get closer to the courtroom door and as you spend more and more money, they start changing their tune slightly, just because suddenly—and maybe you’re just not hearing what they’re saying on day 1. The day before court, they’ll be saying, “Remember I told you it was a 50/50 bet.” And you’re like, “I’ve just spent $2 million and 6 months preparing for this trial, and you’re telling me it’s a 50/50 bet?” You’ll sit there and go, “If I knew then what I know now, I’d have settled at the start.” That’s the best advice for this.

People don’t get it. I’m going to put this clipping in a post I did on it. Rory’s right. Every single time you’ve been through it, you meet with the lawyers and they tell you your case is super strong, right? They tell you in the meeting your case is super strong. $2 million later, it gets close to trial and they’re like, “Well, those counterclaims—I mean, they are silly, but a jury might not see it that way, right? The judge might.” And all of a sudden, it always gets closer to 50/50 when you get to trial.

Guest

The stress goes up. It’s never worth it for either. It’s almost never worth it. But that story—every time you get your dander up and a lawyer tells you you’ve got a super-strong case, I literally just went through this with a CEO. I said, “Hire someone great and have them play the other side.” Play the other side, because they’re going to tell you Rory’s story and you’re not going to want to do it anymore. 100% of the time, they change their tune.

It’s about $2 million in when they change their tune, isn’t it? It’s right around $2 million in legal expenses.

Rory O’Driscoll

That’s so true. It’s funny. My wife was a criminal defense attorney, and she would always say the worst defendants are defendants who start talking about principles. “I don’t want to hear about your principles. I want to just hear what it takes to settle this thing,” right? And don’t get on your dignity.

You guys said before about the amount of kids being put through college because of OpenAI’s legal bills and everything around it—legal bills. I think the same applies here. In any case, invest in Wilson Sonsini and Cooley.

10. What Were Your Biggest Takeaways from the SaaStr Event?

Jason, what a week. I want to start with you. You had the biggest and the best there. What are your big takeaways from seeing the world of SaaS come together in one place?

Guest

The biggest takeaway, I think—you’ve had both of them on 20VC. What Yamin from HubSpot and Aaron Levie both said is, “I’m super excited, but I’m anxious.” That’s the theme.

There was literally 50 times more energy at Saster this year than last year, 2024. I didn’t realize it until I was there. It was the end of the Debbie Downer era. It was the end of folks saying, “Woe is me. I was growing 70% in 2021, selling my fungible sales automation tool, and now I’m growing 2%.” Last year, these people still did not show up.

This was your idea, Harry. We had 300 sessions. I didn’t allow a single session to talk about the past. It was banned. The past was banned. You were only allowed to talk about AI, and you were allowed to talk about AI today and tomorrow. And that was it.

From the CEO of Snowflake to HubSpot, we banned the past this year. You were right. It created incredible energy and anxiety—but good anxiety, right? You’ve got to work twice as hard or twice as fast.

I think that’s great positioning, because I remember either after 2008 or 2009—or I think it might have been after the dot-com crash—what you saw 2 or 3 years later is that even the survivors were scarred. There was just no ability to talk about upside. You’re like, “Oh my God, I got from 10 to 15 to 18, but I’m just so shattered from the pain of the last 3 years that I lack the capacity to think big again,” right? Many of those companies, as a result, didn’t make it.

I think that’s really good positioning. It’s coming out, hopefully, into a picking-up time, and I think that’s what it takes to win this year. This whole “SaaS is dying” bullshit—it’s changing, and you better be AI-forward or dead. But if you are, I’m very certain that if you make the right moves, you can grab hold of this thing and not just grow but reaccelerate growth, which I think is what 2025 has to be all about.

Jason, you mentioned last week about buying Clay out of fear and said, “Hey, Rory, come to the CMO event.” I’m sure Rory was there cheering from the front row.

Rory O’Driscoll

I was there in spirit.

But what were the takeaways from the CMO event?

Guest

The basic vibe was everyone recognized that 20% to 30% of their teams are going to be replaced with AI, and they’re happy for it. They’re waiting for it. Whether they’re ready, whether the tools are ready today, or whether it’s going to be 6 months, everyone is ready and, behind closed doors, borderline excited for the bottom 20% or 30% of their teams to be replaced by AI.

No one’s regretting it. No one’s wondering, “My God, my culture is going to be impacted at my company if the sales rep that takes a week to get back to somebody loses their job to AI.” No, no one was regretting the impact on culture. Seriously, no one was regretting it. They were embracing it.

“How soon can I deploy tools to manage out the bottom 30% of my company? I don’t want them.”

No, what’s interesting about Jason is that it’s not just that he’s the Grim Reaper; he’s the happy Grim Reaper. He’s like, “I love my work. Let’s do some reaping here.”

Rory O’Driscoll

And, yeah, maybe the positive spin on that is you see it: you can grow 30% next year and not add any headcount. But it’s the same story, just perhaps a little more benignly placed, and we all know it won’t actually happen that way. There will be some churn.

But, yes, it’s all versions of the same story, which is that AI is a productivity lift at some level.

Guest

They’re stressed about the change, though, because almost everybody also recognized that moats are weaker. I buy that. Now, I think over time moats will emerge, but you’re exactly right. This is what we often refer to as a run-fast deal. A good portion of the value that will be created over the next 3 years is just by running fast.

11. How Does Microsoft Open Sourcing VS Code Affect WindSurf?

You talk about running fast. I mean, to be fair, Windsurf ran faster than anyone in terms of what they built as quickly as they did, and then to get to the sale position that they did.

Then, last night—or yesterday, whenever it was—Microsoft announced that they’re open-sourcing VS Code, really putting a dent, I think some would say, in the hopes, enthusiasms, and valuations of Windsurf and Cursor. How did you guys analyze that one when you saw that announcement about open-sourcing VS Code?

Guest

It’s interesting because, yes, it amps up the competitive tension, but it is also worth stepping back. If you were the developer behemoth, which is Microsoft, the fact that you have to do this now to remain relevant, at a zoomed-out level, is a sign of relative weakness, not absolute weakness.

There’s still absolute strength, but if you had dominance already, you wouldn’t feel the need to, right? They’re not going to open-source Windows OS. There’s been talk over the years, right? When you have a dominant position, you don’t have to be nice.

This is a manifestation of them feeling competitive heat, always recognizing, as they have for 40 or 50 years now, “Developers, developers.” So they’re doing something to stay competitive. It might dent your perceived valuation if you’re one of the competitors, but you have to give yourself an attaboy. You punched hard enough on a $2 trillion or $3 trillion market-cap company that they felt the need to make this move.

Did anyone have an AI SDR that actually worked?

Whether they work and whether value accrues are 2 different questions, and they occur in roughly that sequence. In other words, if you don’t have clarity on whether they work or not, then you shouldn’t be thinking about where value accrues, because no one’s going to make any value anywhere. Right? First, you’ve got to say to yourself, what parts of the sales process can be automated?

I’m 100% with Jason. There are parts that work well today. There are parts that will work well in 6 or 9 months as the technology progresses, and there are parts that don’t work well. You probably don’t do your marquee email to your very best prospect, but you all have, coming back from a trade show or something like that, masses and masses of leads that just don’t get followed up on. Right? There’s a whole bucketization of it. It’s not so much the marquee work you’re already doing as all the work you should be doing but never get around to doing. That alone is a significant lift.

That’s back to what Jason said. It’s a 50% idea. If you make the reps 50% more productive, if you get them to focus on their key tasks, you’re giving yourself a 50% lift. And that’s not nothing. But I do agree with you, Harry: you have to be very granular about what works and what doesn’t work.

We’ve seen that in all, I think, in all our AI companies, which is why, going back to the investing thing, step 1 on all these deals—and every AI deal—is: do you, the vendor, and you, the customer, have a mutually agreed figure of merit on what success looks like that you can both track? If not, at some point you’re going to churn, because either you’re going to underperform or they’re going to think you’re going to underperform. You have to be aligned around that, and then you have to be really honest: are you delivering it? Time everything—your pacing, your aggression—to when you are in fact delivering enough value to be able to keep them happy and keep them moving forward.

To Harry’s point, though, I’m not smart enough to know which 100 to invest in, right? On the other hand, Otter.ai—how many note-taking apps are there? I believe, based on CB Insights, there are over 18,000 note-taking apps. I mean, I’m in 7 of them.

But on the other hand, Otter.ai has a privileged position at Zoom, but Otter.ai just announced it crossed $100 million in revenue, right? And there are several over $10 million. So I’m like, this stuff’s bundled in, and I can’t tell the difference. Note-taking is becoming like voice, like Gong. For a while, Gong was disruptive, and it still is, but now that functionality is built into everything.

But go back, because I think Gong is interesting, actually, because you made the comment. I think that’s an example of someone who took voice as an entry point and built a compelling multimillion-dollar business. But if you look at what they do today, voice recording—let’s call it sales-specific note-taking, which was core to what they did 5 years ago—is only a part of what they do today.

What they’ve done is use that entry point, build a stack, add forecasting, add CRM updating, add a whole bunch of related functionality, and now, I think, they’re reaccelerating. Now they have a defensible business where AI was the wedge product, and then you had to hustle your way and add the other stuff. I think Gong has done a very good job of that.

Amazon, Apple, Google—we can name the most defining companies of our time, and the list can go on and on. You can even choose big consumer brands. The point is, Fireflies, Otter, Granola—every subvertical. Just for me as an investor, they bring them to the IC with me, and I’m like, “This is not a deal that I want to be in. One of 100.”

I can remember being a snarky little 30-year-old VC making snide comments that Amazon was just a bookseller. Sometimes, if you get—and I remember in 1996, when Kleiner did the round, and I think they went public in 1997—that was when companies actually went public quickly, and you could have got your head around the teeny-tiny TAM, right? It was a wedge. The thing is, sometimes you have to see the wedge point, the wedge, and then where you can go from there. So I do hear you.

I think note-taking is a crowded vertical. It’s a crowded space. It’s one I actually love and would like to find a play in, and I could talk about it for a while later. But don’t make the mistake—Amazon is an example of that—of looking just at the market today. Look at where you can go with that product, and ask whether you can articulate a longer vision.

But going to Harry’s point, when you’re looking at deals at scale today, has defensibility and unique selling proposition gone down? When you score a deal, has it gone down? That’s kind of the question Harry’s asking. He’s brought these deals, the revenue’s there, he likes the founder, but he can’t—the defensibility appears to approach zero in some of these deals.

We just actually had a polite version of an argument on this in our partners’ meeting just yesterday, right? I’m looking at a deal—I won’t say which one—which was in a very commodified market with lots of revenue, lots of good growth, but, oh my God, they can name 10 competitors. And one of my other partners that I’ve worked with for 20-plus years is looking at a high-risk, high-intellectual-property deal with no revenues right now, but clearly—and lots of technical risk—but clearly, if you pull it off, massive defensibility. And we’re sitting here going, “How do we trade off these 2 things? How?”

We’ve definitely skewed more to, yeah, we will do the run-fast deals, but you’ve got to go in with your eyes open and know that you have the team that will run fast. From a portfolio-construction perspective, I really like the fact that we’re trying to add some more singular, different companies where it’s an N of 1. It’s got, ideally, some more product-market fit, but where you kind of go, “This has defensibility and a much longer run,” because I don’t want to wake up with every deal being exactly GPT plus 27 competitors in 3 years. So it’s a tough and competitive investing environment from those perspectives.

Listen, behind it all are the model providers, the ones that are the best businesses, right? We go through the peaks and troughs of, “Oh, they’re great. They’re commodities.” And now we’re in the realization that OpenAI and Anthropic are actually phenomenal businesses.

Anthropic’s run-rate revenue grew from $1 billion in Q4 2024 to $2 billion in Q1 2025. Its 100,000-plus customers grew 8×. How did we analyze this? Is this way better than we thought? Was this what you thought? Does it change your perspective on Anthropic?

I mean, listen, you can segment the market, but it’s also that Anthropic is a distant number 2 in some ways, right? The growth is jaw-dropping.

On the other hand, listen, I’m not—I’m sure you guys have chatted about this—I mean, OpenAI also said they’re going to burn at least another $44 billion until they’re profitable in 2029. At least another $44 billion. It’s jaw-dropping. It’s the old Amazon thing from the old days that Rory said, on like 18 doses of steroids, right? Everyone’s all in. It’s another $44 billion, but it’s clearly a multitrillion-dollar company, so the math ties.

Is it a new world, or will it collapse on itself? I don’t think so. I mean, listen, I’m so all-in on AI. I’m totally bullish. But $44 billion is a lot, isn’t it, Luke?

Well, I interviewed a growth investor this morning. Sorry to interrupt you, Luke. He’s already a holder in Anthropic, and he said, “I’m trying to buy every employee’s options. I’m trying to buy everything around Anthropic that I can. I’ll get a billion dollars’ worth. I’ve got supply from my LPs for $5 billion.”

Yeah, I’d believe it because they’re great businesses. And, you know, I haven’t cycled through the up-and-down way you describe it. They always have to be great businesses because they’re the most defensible part of the stack. I don’t buy the commodification argument. There are going to be 2 or 3 of them; I don’t think there’s going to be 10.

People are going to build around that, and they’re going to make technical choices. You’re going to fast-forward 5 years, and just like lots of people could do Amazon Web Services in 2007 or 2008, once you’re 10 years in and you have the scale and you have the development environments and all the rest, those are going to be great businesses. They’re going to be great businesses at scale.

Will they be worth $300 billion or $1 trillion? I don’t have a developed opinion, but no, they’re the anchor tenants of the AI economy, just as Amazon Web Services, Azure, and whatever Google Cloud was were kind of the anchors of the cloud economy.

With the rise of great companies comes the fall of others. Chegg: $12 billion to $95 million, third round of layoffs. Is this just exactly what we said—that moats are smaller than ever and companies can be changed overnight?

Guest

Yes, short answer. Why would I pay Chegg for X when I can literally type it into ChatGPT?

Yeah, that was a business that was dead roadkill in front of us. Jason, what’s the most obvious next Chegg?

Guest

The tough one is if MCP really works, right? It’s very interesting seeing HubSpot, Box, and Dropbox be excited about MCP. If it really works, we won’t even really need these applications.

12. Why So Many Companies Are About To Become Database Companies

If I can use MCP to grab my content from Box and do better, more powerful things with it by combining it with my own model, I’ll barely even know Box exists. Look, HubSpot is great, but if my own AI can pull out all the structured information from HubSpot and run my own AI the way I want to do it, I might not even ask, “What is HubSpot going to be worth to me?”

If I can put my own AI and my own agents on top of these apps, all these ones that we think are so great because they’re databases, I think they instantly become vulnerable because they just become databases. They just become databases, I think. And I think they know it. I think that’s why everyone’s stressed.

The only pushback I’d give is the word “instantly.” I do believe AI will exert pressure on some of these apps in terms of newer solutions to do it. But I do think we should assume that the replacement cycle plays out over 10-plus years.

I think the existing dominant vendors have a chance to ride it out. Look at what ServiceNow is doing very aggressively. But I think you’re right: if you are asleep then, or if your app doesn’t deliver a ton of value, you will get ground down slowly and painfully.

13. The Future of Salesforce: Buy or Sell?

So, you mentioned they’re turning very quickly into a database. People say that about Salesforce. If you were to look forward, would you be a buy or a sell on Salesforce, with the potential for it to be a database overnight?

Rory O’Driscoll

Yeah. I mean, I hold Salesforce stock. I’m not a seller, right? I do think Jason is right on the analysis. I don’t think they’re going to be an explosive grower from here, to say the least.

But I think one thing we underestimate is the power of incumbency—not to be a venture-type IRR performer, where you’re hyper-growth and track into an IPO, but just the ability of these big tech companies to extract massive profits and massive cash flow at scale.

If Salesforce is, quote, “just a database” in 10 or 15 years, it doesn’t mean they can’t still kick off gobs and gobs of cash. All you have to do is look at Oracle. Not only is it just a database, it is a freaking database—that’s what they did first.

And Larry Ellison, depending on the day, somewhere between the 1st and the 4th richest man on the planet, isn’t sitting there going, “Oh, I wish I was cool in AI,” right? He’s going, “I know how to optimize every damn dime from these corporate customers. We’ll ship just enough new stuff to keep them on board, and we will build a wildly profitable, I think, 43% operating-margin business.”

So, I’m not looking at Salesforce to provide the oomph in my portfolio, but I wouldn’t want to offload that stock with those cash flows because I just don’t think it’s going away.

We use Salesforce in our shop here. I think I’ll be dead before we rip it out. Guys, you’re amazing. Thank you so much. I so appreciate you guys.

Rory O’Driscoll

Harry, I just want to be clear. I was just bringing it back to the questions that were in the notes that I prepared, too. You kind of went off script in the first 2 minutes. It took you 30 minutes to even get back to the second question. I was like, “Dude, I didn’t prepare for freaking VS Code forking. Give me a break here.”

I blame Jason. They were in his notes.