20VC：Daniel Gross 与 Nat Friedman：被 Meta 收购｜OpenAI 的 SBC 炸弹：股票薪酬超过营收｜私募股权回归：Olo 以 20亿美元被收购｜Microsoft 裁员 9,000 人：这只是开始吗｜Sequoia 会与 Shaun Maguire 分道扬镳吗

It will be the biggest issue of 2026, I think, in B2B AI: the inability to recruit talent.

No one ever said to Winston Churchill, “Did you bring World War II in on budget?” They just said, “Did you win World War II?”

Yeah.

The truth is this: when it becomes existential, you do what you have to do to win. The run in the early AI markets has been very much attention begets more attention begets more attention. So if you start to pull ahead, provided you continue to execute, it’s very hard to catch up.

Guys, I am so glad to be doing this at a normal time. What everyone didn’t see was me at 6:00 a.m., being slightly slow to start on the last show. Doing this at a normal time: great success. Guys, it’s so good to have you back.

Great success. It’s a big week.

Yeah.

Things just keep accelerating, Harry.

Dude, things keep accelerating. I’m going to start with one that you tweeted brilliantly, by the way, Jason. You said about Daniel and Nat joining Meta, “The wild story of NFDG: two Silicon Valley legends built a $1.1 billion fund, 4Xed it in 2 years, and then abandoned it all for Meta this week.”

Bluntly, what do we make of them moving? Why don’t we start with you, Jason? How did you analyze this? Because it’s pretty big and shocking news.

1. The NFDG Meta Exit

Well, listen, I want to have Rory help me with the math here. They’re 4X on a $1.1 billion fund. The Wall Street Journal said it’s about 50% deployed, and they’re already closing another fund: 2 partners and a few other guys. For 99% of the venture world, this is beyond a dream outcome.

There’s a lot to the story. Obviously, it’s a moment in time that includes the ex-CEO of GitHub—not the founder—who’s had a run and now wants to go work for a dude. Here’s the problem: first, I get the excitement. I just flew back into Palo Alto today. I was in SoCal for a week, and I already feel the vibe.

I want to do the same, but I also worry this is going to be like the Trump administration, where everyone’s going to quit, too. They’re not going to last. The Elons and the Davids and all the techies aren’t going to be there for 4 years. Do I really want to go work for Meta for 4 years?

That’s the only weird part of it, but I get the moment. I don’t want to be friggin’ meeting founders and writing checks. This is a once-in-a-20-year moment in time. This is like 1999, except it’s not going to implode on us in 12 months.

And it was a lot riskier, Dan, because there wasn’t a whole bunch of large incumbents willing to take your quitting and monetize it for $100 million.

Yeah.

So in many respects, it’s far more zany than what went on from ’95 to ’99, when I actually was around and in the business. This is something we’ve never seen before, because there have never been incumbents just willing to plop down this kind of cash to hire people, for God’s sake.

That’s—

Right? Just to go do something. And it gets back to what we’ve discussed a couple of times, which is that a small number of high priests of AI who are deemed to have the answer just have huge market value. This is the market working through.

I don’t mean to interrupt, Harry.

Sure.

But let’s just do the math for a moment for folks who might read or watch. So you have a $1.1 billion fund. It’s 4Xed, okay? Let’s be generous on the math, but it’s true, right? Half of it is deployed, so $500 million times 4X is $2 billion. That’s $1.5 billion in profit on paper, which we can talk through.

Yep.

That’s $300 million in gains already.

Correct.

We’re ignoring some subtleties, right?

Right.

But we’re already $300 million up, the 3 of us. The 2 of them, plus a few guys that probably have 2% carry. They’re already up $300 million in 2 years. They’re up $300 million in 2 years.

This is why the offer is so good, because what you’re getting is—you had a billion-dollar fund, you put out half a billion, and you’re 4X up: a $1.5 billion gain. Remember, it’s worth pointing out, you’re not getting that money now, at least to my knowledge as a GP. It’s the LPs.

They’re basically—and I give them huge credit—taking care of their LPs. They’re saying, “You can take half your money off the table,” which effectively means even if the whole thing goes to shit, you got a 2X. So the LPs are getting the money off the table.

What they’re giving up—and there is a real give-up here—is, as investors, the right to use the other $500 million and the other billion that they were clearly going to raise in 2 weeks flat, because they’re those guys. Basically, you should think about the offer as being in 3 components.

Nat and Daniel, we will, A, take care of your existing LPs in a way that will make you feel good because you’re reputable people. Secondly, you’re walking away from $1.5 billion of investable capital, and the value of that depends on what you think you could turn it into.

Let’s just say you could do another 4X, which means 3X of gains, which means $4.5 billion. At 20% profit, there’s a credible argument that’s $800 million of value to you, Nat and Daniel. Obviously, whatever other offer they got for that had to be better than that.

Plus, they had premium carry in the fund, so let’s call it an even billion. They gave up $800 million to a billion, in theory, to join Meta.

I personally think not having to put that money out in today’s frenzied market and instead getting—if they got anything like a comparable equivalent in terms of capital return—that’s a pretty damn good deal. It’s also an excellent deal for the LPs.

They are losing the stewardship of those 2 guys on their investments, which is why there are no new investments. But on their existing money in the ground, if they’re putting $1 in, they’re getting $2 back, and they still own the other $2.

So if SSI goes great, hooray. If it doesn’t go great, they at least got a 2X. For the LPs, it’s an interesting one, with a subtle nuance. I believe the terms are that, overall, Meta will buy 49% of the fund, but each individual LP can opt in to sell as much or as little as they like, and then it just aggregates.

So there are a lot of LPs going, “Hmm, do I take my 49? Do I hold tough and sell nothing? Or do I ask, ‘Hey, can I take it all if it’s available?’” This is a chance for liquidity in a stellar fund early. It’s an interesting question we could talk about.

I remember a couple of months ago, I was having lunch with one of my LPs that I share with Harry. After lunch, the LP was going to meet with them for this fund—for fund 2—to put it all in.

I’m like, “You’re doing what?” He’s like, “Jason, it’s not even the same game you’re playing.” He’s like, “Don’t worry about those guys. We’re not in the same box. You’re not in the same bucket.”

I do not think LPs are happy about this outcome, by the way. Don’t get me wrong: they’re happy to get cash back and good economics back. But I think there was such excitement and fervor around them as a partnership and what they were building that they will be sad to lose the stewardship and the future funds.

Yeah. And I think at that level, that's probably true, but there are a lot of people. Look, getting jilted is one thing, but getting jilted with a 2X is a lot less painful.

When funds end, most of the time they end in weirdness. It ends up being a pain in the butt for LPs. I'm giving Daniel and Nat very great credit. This is a very clean ending. They can look everyone in the eye and say, “Yes, simply put, you, Mr. LP, would like us to do venture capital for the next 4 or 5 years. It would appear, in the market system that we all live in, that our highest and best use is building AI for Meta.” So the market has spoken, and that's what we're going to do.

I'm genuinely not surprised. I can say that with some credibility because I competed for a deal with those guys, and we lost to them. It all worked out in the end.

Rory, they beat every single person in the market who wanted to win.

I know. I agree. I ended up getting to work with them, and I think they're wonderfully talented guys. But I remember saying to the CEO, “They won't be doing this in 4 years,” because it was obvious to me.

Why?

That's interesting.

Venture—look, it's a perfectly good gig, but it's not... They have so much more talent. If, when you meet 2 people, one of whom has been a CEO and a talented entrepreneur early on—we looked at that round at Xamarin—and has been the CEO of GitHub and built the most compelling product, and then his colleague has been involved in that kind of early AI stuff, capable of being a founder at SSI, your highest and best use is not being the 50th venture guy.

Even if you're the best venture guy, there are other things you can do. No, I'm totally not surprised.

No, I've never felt so unspecial, Jason.

You should.

We're not. I felt this way going into venture myself. I felt like I was walking into quicksand, a world with fungible sources of capital, where clearly some GPs are better than others in terms of adding value, no question. Some funds are modestly better than others.

The only value I think you can add in venture in the world really is if you discover talent that would otherwise not get funded. That's the mitzvah in it, isn't it? You find the young Rory. No one would fund him, and he builds Cursor. You've done a good thing for the world. No one else is really adding significant value to the world in venture.

Well, dude, that's why I did Project Europe. We just funded this kid in Athens with Project Europe who's doing humanoid robotics from his grandmother's garage. Never, ever would he have been found.

If I'm an investor in SSI and I've just put in a big check, and Daniel led that fundraise, I am a bit pissed off now.

It's interesting because that was—it’s going to sound zany—but why are you pissed off this week rather than a couple of weeks back, when he stepped away from SSI, right?

No, I'm saying it because he led the round as an operator, and he was in charge of the fundraising, committed to it as an operator. I think that is a responsibility there that you don't piss off a couple of weeks later.

I think it's interesting, Harry. When I wrote it up, the other thing I said was that it was an interesting parallel to Gary Tan, because Gary Tan left Initialized to run YC 2 years ago. I instantly got it, right? I don't know Garry very well, but I've had a chance to watch him a little bit over the years. We have a lot of LPs in common, and people thought it was crazy.

But their fund, too, is going to be a 10X fund, I think, with Flock Safety, Rippling, and others. So he left them with a 10X fund, right, a great position. A lot of them are involved in YC today as LPs and otherwise. But some people were upset. They're like, “We love Garry, and we wanted to go another couple of rounds.”

The LPs we had in common—don't get me wrong—were all happy for him, but they were upset briefly because they thought he was going to do this for another 20 years, right? I'm like, “This guy's a builder, right?”

Totally. People with those talents are going to be drawn to the thing that can most allow them to instantiate those talents. You shouldn't make decisions based on someone else, quote-unquote, doing the right thing. All you can do is evaluate their incentives and motivations. Are they aligned with you? If they are, most of the time it'll be fine.

But even then, kind of back to what you said last week, Jason, we're seeing a lot of founders tap out. Walking away from something, or in this case walking towards something, is just going to happen in this market. You can be pissed for a day, but I go back to what I said: if getting screwed over is getting 2X in cash and a ticket to ride on the other half in 2 years, we should all be so lucky.

2. The AI Talent War

I'll just end it with 2 statements. 1, I really don't think this is about money. Both of them were worth half a billion before. Another half billion—I don't think they give a shit about that.

The final thing I'll leave on is: do we think this talent-accumulation machine that Zuck is building is going to work?

I think it will work. I think it already worked at X. I think they're running a little bit of the Elon-X playbook of just being insane, cracked, creating this mega-mecca for talent. I think you almost have to do this playbook. You have to create a mega-mecca for talent somehow, right?

I think it will probably work. The best want to work for the best. It's always been true our whole careers. The very best only want to work for the very best. They won't tolerate anything else. That's why so many struggling unicorns are in an existential death spiral, because they can't attract anyone great. Not a single great person is going to join unless the founders are great.

Some folks don't care that the growth has fallen to 10%, but they want to work for the prominent best in many cases. This is in the 2-by-2: the best and super-prominent. I'm going to go 80% of people want to go there. The best people want to go there. And you need soldiers, not just captains and generals, right? So you have to attract them.

Broadly agree with that. One of the things that's funny is that I'm just reading one of the many books about OpenAI. I think it's called Empires or something, and they talk a little bit about how they deliberately and ostentatiously wanted to raise their profile for exactly that reason, Jason: to be able to hire the very best people.

So this dynamic has been there. Literally, one of the first emails—I think it was sent to Elon—was basically some version of, “If you would lend your name to this project, it would be cool because we would hire more engineers quicker because you're cool.” End of an email, right? Back in 2015 or 2016. So I totally agree with that.

At that level, I think, Harry, the answer to the question is yes, it will work. These wildly smart people—they are members of, as I say, the mythic inner circle of people who know the magic spell. They will make the product.

I think the interesting question, as a business decision, will being the 4th or 5th broadly capable LLM be a compelling business for Meta? To me, that's a much more interesting and difficult, unclear question. But I don't think they're agonizing about that now. They just feel the existential need to play.

I was with a founder this morning of a multi-billion-dollar company. It's a very good company. They said, “My single biggest challenge today is Cursor. Cursor's just paying insane amounts of money for everyone.”

No disrespect to Cursor, but the question for me was just: wow, if this is a tidal wave of just incredibly well-funded AI companies paying through the nose, and it's not just Meta—

Totally.

—but suddenly 10 others have to compete, where does this end?

With people losing money. I mean, obviously.

The talent war is under-discussed. How the hell, if you're not at a top vibe-coding company, are you going to compete for talent? You better find somewhere where being very good is enough, because you cannot take these teams head-on. There is just no way.

You also can't pay them $800,000 a year plus RSUs plus guarantees. You just can't do it, right? The biggest issue of 2026, I think, in B2B AI will be the inability to recruit talent. I don't think firing up Lovable or Replit is going to solve this problem. I love both tools, don't get me wrong.

I see too many startups saying, “I'm going to hire an AI guy.” My worst one is when a startup says, “I want to hire a VP of AI.” I'd like to sell all my shares on any secondary market that exists. If you think your answer is to go hire a VP of AI who wears a tie and is studying things, just shut the startup down. Sell it for anything. Go sell it to Grammarly if you can, while there's time.

My God, you're just piling on. I think that's true. Obviously, at the application level, if you're a user of the models, you don't need to have the same caliber of people as it takes to build the models, provided you have people who can deploy them and use them. So I see Jason making his I-disagree face.

It's true, but the problem is, in a lot of these categories, sure, you have exposure to the same models, but your ability to do more with them requires an S-tier team.

Otherwise, you're just lost in the sea of sameness.

Yeah, and that's the point I'm trying to make, Jason. You're exactly right: they're not doing it by paying $800,000 or $1 million in cold, hard cash to the 50 people in Silicon Valley who can do that. You build a center of excellence somewhere else. You have to be very, very good, but you don't have to be quite as on it as if you're building Cursor.

I was actually talking this last week with the CEO of a really good B2B company. It's adjacent, but it's coming up on $200 million in ARR. Doing well, okay? And he's like, "Yeah, I basically have to give half a percent or a percent of my company to each AI engineer now—to get who I need."

Wow.

You can't do that for 50 engineers, can you? How does that math work out at the Series D dilution stage?

Pretty badly, to be honest, Jason.

It's pretty rough on everybody. He's forgetting about the cash. He's like, "I gotta do half a percent or more to get the people I need today." And he's like, "I got no choice," right? Coming up on $200 million, that's the time when engineers start to get 0.00001% of the company.

3. SBC Dilution Gets Expensive

That's partly why this wave, I think, will be highly damaging to venture returns, because the employee stock-based compensation dilution is going to be so significant, and so much more significant than in prior generations of venture. If you're Anthropic or OpenAI, you're just continuously having to top people up every year.

The Information said OpenAI has more SBC this year than revenue. They came out with a piece this year: more SBC than revenue.

And intellectually, you're correct. I'm not going to lead the "SBC doesn't matter" comment because—

If it's more than revenue, it might matter.

But it's more than revenue only because there's this spurious way of accounting for it that's cash-based, which is quite misleading. It's more clear to think about it as a dilution percentage versus cash, and I understand there's an element of both. When you're sitting there allocating the money, the more it's freely tradable stock, the more you should think of it as cash and a direct replacement for cold, hard cash. In which case, if it was a 100% replacement for cold, hard cash, you could argue their loss is approximately double what it's stated to be.

If you end up building a big-ass company and you look back and go, "This is horrific—you had 20% or 30% more dilution than you'd expected," you still got a pretty big company. I'm not advocating mass dilution, but the truth is, it's not going to be the fatal error. Not getting it is the fatal error. It's like a line I often use: no one ever said to Winston Churchill, "Did you bring World War II in on budget?" They just said, "Did you win World War II?"

Yeah.

Right? When it really matters, the truth is this: when it becomes existential, you do what you have to do to win. This is where, again, we said it before, you give Meta huge credit. I don't know if they're right in making the assumption that this is existential, but once you've decided it's existential, you just do what you have to do.

Listen to this statistic. It's pretty funny. Last year, according to The Information, OpenAI had $4.4 billion in stock compensation. That was 119% of its GAAP revenue, okay?

Yeah.

Now, that's a lot, but it projected it would fall to 45% this year. It's not going to happen. Now it's going to still be in the triple digits. That's to Harry's point. I would assume if your SBC exceeds your revenue at scale, it's pretty dilutive, right? But the fact that it's 119% instead of 45%, that's a lot of dilution, isn't it?

Oddly, it might not be. I'm not trying to really get down in the weeds here. I'd like to know the percentage dilution, because one of the weird things about SBC compensation is it's priced when you issue it, and it's an accounting entry based on the 409A valuation of the stock, which has been motoring up enormously. So they may well be recording a large, very large, quote-unquote, "GAAP SBC number," but the dilution, while still big, may not be nearly as catastrophic as the GAAP figure.

Oh, for sure. I'm being too simplistic. If they're worth $300 billion and they're issuing—

Yeah.

—$10 billion of stock, it's 3%, right?

That's exactly right. I couldn't articulate it because I'm a bit jet-lagged, but that's exactly it. You see this a lot, where you have a company that has an enormously high mark from the VCs off in the public market. They issue stock, and then you see this a lot in the public market: 2 years later, the stock has returned to a much lower level, but the SBC still rolls through the books as if it was all priced against the $300 billion value. Then you have those absurd public companies where, oh my God, the SBC is 3x the revenue. It's all about this nominal accounting charge, which is meaningless.

Now, in this case, it is interesting because to the extent that the $300 billion is a money-good valuation, if you give someone 3% of the company, they are in fact getting $9 billion or $10 billion in value. If they can realize it via a series of tenders, then good on them. If the market cap of that company stays at $300 billion or greater for the next 4 years, the people who got $10 billion today will have $10 billion of cash value. Good for them.

4. CoreWeave Uses Its Stock

Speaking of capitalizing on appreciation of stock price and using it strategically, there was one that I thought was really interesting: CoreWeave, which bought Core Scientific for $9 billion. This is after they bought Weights & Biases for $1 billion. I'm intrigued. Is this just an incredibly strategic use of an appreciating stock price? How did you think about this?

I thought it was a great deal. Quick and simple. I remember the hosting business in '98 and '99, and I looked at Exodus way back in the dawn and couldn't get people there—a tiny number. It went public. It was basically offering the first-ever hosting facilities, went public, had a huge valuation, and then in the dot-com bust, it went bust. Absolutely bankrupt, because it had leases and debt, and it killed it.

These guys have a mark, have a currency. They bought—was it Core Scientific? Apparently, they lease a lot of data centers from them. It effectively de-leverages them somewhat. It uses equity to take out rent expense, and it means that over the next 5 or 6 years, if the demand for data centers doesn't continue to grow to the sky and they have to have a more robust balance sheet—I mean, shock horror for this next sentence—if the demand for data centers doesn't continue to grow to the sky, they will look back on this and say, "Great move. We took a bunch of fixed costs out of our P&L in return for a small amount of dilution."

I'm sorry, who would take the bet that the demand for data centers would go down?

I might consider it, Harry. I can see it in your face. I can see it—everyone's like... I would put that bet at least on the table as being plausible at some point in the next 3 or 4 years, that people find they're ahead of their investment schedules and just want to slow down. I can see the pained look in your face, but if I was the CFO of a company like CoreWeave, what they effectively did was take themselves from 100% leveraged on the upside of infinite data center demand to 80% leveraged on the upside of infinite data center demand. That's a smart, slight de-risking.

Wow.

It's not like they're shorting Nvidia stock here, dude, which is the bet you take if you really believe. There are a whole series of bets you can take depending on how much you think the hyperscalers are going to spend to build these models in the next 4 or 5 years. It can go all the way from, "I'm so leveraged to the upside that I'm going to buy Nvidia out-of-the-money calls." That's if you really believe, and you can just go in a descending order of risk reduction from there.

These guys are very much leaning in on data center demand. They're not walking away from that, but what they're doing is replacing—because it's effectively a financing company, CoreWeave—they're replacing fixed charges of debt and leases with equity. It's a good move.

Yeah, it would help us all if there were 10 CoreWeaves. If you have a public company trading at a crazy-high multiple, say, 20x revenues, that has existential losses, you've got to use your stock as your currency ASAP. You've got to buy everything you can that can address the bottom line or other challenges you have, right?

These are fun times if you're a target. You're going to get bought in an hour because the CoreWeave corp dev team is going to be looking for 20 assets that they can buy that can fix some of the structural challenges in their models. We haven't had an IPO like this in a while. We haven't had someone searching with high-priced equity, and it's a great deal for everyone to turn high-priced equity into cash.

Yeah. And I don't know if they even say the challenge is in the model as much. I mean, you're right. It's not the negative in the model—it's just the nature of the model. The nature of the business is, you buy these, fix that. They have long-term contracts against them, so they're a lot less exposed than some of these guys that are doing short-term contracts.

But it's just a fixed-cost business. You own this great big fricking data center, all these machines, and as long as there's money coming in the top, it's fine. But if not, it's brutal.

It makes sense.

All I was thinking was, Jesus, look at the other meme stock being Circle. Why are they not leveraging their stock appreciation to do the same?

Probably because they’re still not out of the first 90-day period, when it’s a pain in the ass to do it. And believe me, because CoreWeave, remember, has been out there a couple of months longer, at least.

Generally, there’s a period of time where you want to put your first—I remember asking this, actually, when I was on a public board. You probably want to put one or two, at least one, earnings release on the board as the asset that you have sold to your investors before you start doing crazy deals. I’m sure that stock is burning a hole in their pockets too.

Yeah, they’ll do something.

But Circle had net income of $156 million last year. So CoreWeave and Circle both have inflated stock prices. They both should use that to enhance their business models, and they will. But it is different if you have net positive.

When you’re profitable, you think about these deals differently, right?

Yeah.

You do.

If you’re burning cash, no one sweats the dilution from an existential deal. You just do it. It’s like, “Let’s get going, guys,” right? When you’re profitable, it just confuses a lot of things, right? It impacts your earnings. It impacts—you care more about dilution. There are a lot of things going on when you’re profitable.

I think if Circle does something, it will be about buying distribution, because obviously their biggest cost is the money they have to pay Coinbase. So finding other ways to get their coin into the hands of users will be their thing, whereas for CoreWeave it’s just addressing their cost structure over time.

It’s a weird thing. It’s like SBC. If you’re losing money as a public company, for real, you don’t care so much about dilution. It’s just more dilution to achieve your goal.

You only really start to care when you cross the curve to profitability, right?

You’re exactly right, and I would just add one note. I think it’s not that you don’t care. It’s the Maslow hierarchy of caring, right? If you don’t make profits, then it doesn’t matter, because you’re screwed. So you are—and that’s why early on, VCs are right to keep an eye on dilution but not get all antsy about the economics of it. Just keep an eye on dilution.

The way I think of it is, we’re all in this together. If we create something of value, everyone gets a share. Once you’ve created something of value, then a decision to allocate some of that is a much more meaningful economic term. I could not agree more. The way to think about SBC as a venture firm or a pre-profit company, and the way to think about any kind of dilution, is just very different than when you have a cash-flow-positive asset. It’s a different thing.

5. Private Equity Goes Shopping

Speaking of different things, a different kind of asset but one that I found very encouraging was Thoma Bravo going shopping with Olo. It’s a business with $320 million of ARR, it’s GAAP profitable, and it’s decent. It’s bought for $2 billion by Thoma Bravo. Is this PE coming in to save the day, as we’ve heard time and time again? Remember, it was public as well, so they’re taking it private.

I have talked about how, personally, I’ve been concerned. I haven’t seen the PE deals that I’d hoped for with these funds raised. It’s a good sign, right? It’s a vertical SaaS player. If you want to compare it to Toast, I know it’s a tough comp, but it’s $2 billion versus $25 billion, and you could see the upside.

There are positives and negatives in that, right? Defensible. It is defensible, right? Sticky. If you squint, maybe there’ll be more of these vertical SaaS deals.

6X ARR, though, for us as investors, this doesn’t buy the biggest house in Atherton. This is not going to buy a fourth house, right? At Yellowstone Club, we’re just going to get a condo. We’re not even going to get something on the hill at 6X ARR.

You’re right. It’s a meat-and-potatoes deal. That was my aha. It’s a 20% grower, profitable, 6.5X—everything’s right with the world, a sensible kind of deal.

I think it went for $10 a share. I think the damn thing went public for $25, so just as a reminder of what’s happened here: it went public in 2021. Hopefully the VCs made money. I can’t remember who did the deal.

It trades way down, and obviously it gets to a point where, after 3 or 4 years, people are willing to transact, and the buyer’s willing to pay 6.5 times. So, a totally solid meat-and-potatoes deal. And if you’ve got any more 21% growers that are profitable, that have, as Jason said, good defensible vertical niches, I’m sure they’d be glad to give you 6.5 times for that too. It’s a start, but it’s not going to save 500 to 700 unicorns single-handedly.

I think that 20% number—we talk about it a lot in venture at scale, right? But it’s an example of how important it is, right? You’ve got to be—

Oh.

—above that number to be of interest, right? Probably, right?

And profitable. Yes.

Yeah, and profitable, right? Yeah. It’s still probably the Rule of 25. We should look it up.

Yeah.

Right? But you’ve got to be above 20 and profitable, and there are too many unicorns that are profitable but not above 20 or 25. That’s your 6X outcome.

What are they underwriting this to? Is this like a 3X?

I think that business is inherently a 2 to 5X kind of business most of the time. So I’m sure Thoma Bravo is looking at this and going, “There are 5 other pieces of technology you can bolt on here and sell to the same customers.” Do the kind of thing that, frankly, PE does well—to some extent, I would say better than us—which is just figure out what your top customers want, go buy the other little things, bundle it all through, get more share of wallet from the top 500 restaurant chains, and just build a business here.

Jason, does this make you more or less excited to be an owner, seeing this happen?

Neutral. Owner is even smaller than Toast, right? So its core customer today—there’s a lot of overlap, right? But it’s smaller.

It just reinforces the non-obvious thing that when you’re selling to end consumers, the long tail is sometimes where the biggest dollars are. I’ve been on too many boards with SaaS VCs who are always like, “Go more enterprise. Why aren’t you going more enterprise?” They don’t understand certain end-vertical markets, right?

And then sometimes ServiceTitan is very enterprise, right? It’s not obvious. ServiceTitan is not a bunch of plumbers paying $20 a month out of their own pocket. But more of this B2B-to-C that hits consumers—the long tail’s where the money is.

Even today, Shopify is very enterprise. Only 25% of its revenue is in enterprise. So, going to Rory’s point, I don’t know. I’m sure the game is, one way or another, to combine those 2 with 2 or 3 players and build something that’s 3 times the size. So I assume that’s the playbook.

Guys, when I listen to this, and I’m a founder, I hear crazy payments being made for venture funds. I hear PE come in. I hear massive meme stock price rises. And then I look at Carta’s data: VC deals at an 8-year low, and it’s hard to raise money for a lot of founders.

What world are we living in, and how do you think about these 2 paradoxical statements—everything we’ve discussed combined with VC deals at an 8-year low?

I think they go together, and it’s tough. I think what’s happened is there’s a whole bunch of things all pushing. Somebody used a really nice phrase: a flight to consensus or a flight to quality. The truth is, there’s a small number of things that are working really well, and everything else looks dull in comparison and is struggling to get attention.

It’s become a very consensus bet. And there’s a whole bunch of reasons for that. We’ve talked about this before. The whole trend of staying private for longer means a much smaller number of companies are going to get all the way, and they tend to get all the attention.

Second, entirely separate comment: I think the early AI market has been very much attention begets more attention begets more attention. So if you start to pull ahead, provided you continue to execute, it’s very hard to catch up. I mean, coding, for example—Cursor pulled ahead. It is better, but there were 20 people trying to do it.

You get that early lead, and it builds on itself. And then the other people are just left in the dust. And it doesn’t matter that they’re nearly as good, because the sad truth is, no one wants nearly as good. So it’s a very steep fall away from the small number of things that everyone wants to a much wider number of things that might just be okay, but they’re not amazing.

Frankly, as an investor, it’s very hard if you’re not in one of the amazings. You kind of have your sad face on. You’re working hard, you’re doing your job, but nothing’s exploding. That’s when you’ve got to keep your head, keep the companies moving forward, converging on acceptable growth and acceptable profitability.

But it’s a weird time, because you’re reading about all this amazing stuff, and then you’re going back to your day-to-day job, which most of the time is a grind.

Is it really a sad face, though? Before, I’ve said it’s the end of triple, triple, double, double, and I remember you said to me, “I’d take triple, triple, double, double all day long.” But triple, triple, double, double, bluntly, is boring today.

That's not zero to $100 million in a year. That's not lovable. Replit and Macaw, any of them.

But you know what it is? This weekend I was talking with an entrepreneur who was on the triple, triple, double, double path in a space and a part of go-to-market that I know well, and the deal size was right. If this was even early 2023, I probably would've done this deal. But today, there's so much competition, and the differentiation is less clear.

Great founder, but is he generational? I hate this term—I mean, I put “hate” in quotes. Harry, you use it a lot. It's a fair criticism, but are they really a generational founder? I hate this term. I just said to the founder, “I've got nothing to criticize. You have the numbers. You have a great approach. It's interesting.”

In 2025, my brain just says, “I only have so many shots on goal. I don't want to take this shot today.” I think it ties to the Carta data, right? People want to swing more for the fences. Either it's there and it's off to the races, or they've got to believe it. People have just got to swing harder. Nothing negative to say about this triple, triple, double, double. Nothing.

Nobody stays down on the farm when you can play in the gold rush. The attraction of that kind of upside, the truth is, it turns everyone's head. If there's 2 games and one of them has that embedded 10% chance of an amazing, $20 billion Cursor outcome, and one of them just doesn't, it's actually a very interesting math question: How much cheaper does that other non-Cursor-embedded-upside deal have to be to cover for the fact that it doesn't have that kind of outsized tail outcome?

I fear the answer is either it has to be a lot cheaper or, even worse, there's no price at which you'll do it.

The triple, triple, double, double, less-AI-centric companies are not realizing that they will have a discounted price because they are not this new wave of company. It's hard, and I'm sorry for them, but it's just the truth.

Agreed. There's a price at which you can make your target return. I remember looking at a deal with a strong founder, good economics, good growth, and a mid-sized market, with no amazing outcome. There's a price at which you'd say, “You want your base-case return to work there,” right? But it's quite a big disconnect.

It's interesting, though, Rory. If you and I were partners at Scale—and I hope I'm not overstepping here—I would be like, “1,000%, let's stick with triple, triple, double, double,” because we have a much higher certainty of winning there versus going for the AI-halo moonshots, competing against Thrive, Founders Fund, and Andreessen, where the Scale enterprise brand doesn't carry the same weight that their glossy brands do. You have a tiny chance of winning versus a massive chance of winning in the triple, triple, double, double space, where most people have left.

I will absolutely look at deals where you have that profile, but it has to have that profile, that valuation, and that upside, with pricing such that you can pencil out the return.

One of the problems with the private-for-longer dynamic is that you can do those deals when they're already at scale, if they're at $50 million or $60 million and they're on that trajectory, right? The problem is when you go super early and predicate it on follow-on rounds. Some of your destiny is outside your control, and if people aren't willing to fund it, you have to have existing scale, sufficient traction, or confident profitability. It's not enough that you like it. Enough people have to like it along the way to be able to raise the money and get there.

I'm more intimidated by the deceleration at scale of folks I did not expect to decelerate.

Yeah.

I want to know earlier that you are clearly differentiated in a way that can win, which we gave up on in 2020, 2021, and 2022. We didn't care what the difference was between a lot of B2B players. I'll do the triple, triple, double, double, but it's got to be durable for real.

I have to believe that it's just going to keep going. You have to dominate some segment of your market for a real reason that is enduring.

Everyone who's been in the business 5 or 7 years did a whole bunch of deals that were growing like crazy, man, in 2017, 2018, 2019, 2020, and 2021. They've all decelerated at scale, and they now have a bunch of deals at $100 million to $200 million with 10% to 30% growth rates. Most of their waking hours are spent figuring out what the frick they should do with those deals.

It's human nature. The one thing you don't say to yourself is, “This is so much fun. Let me add to that collection.”

The other comment I'll make is this: The industry as a whole is probably making the same mistake with AI deals that it made in 2021 with a whole range of SaaS deals, which is extrapolating the current growth rate to the sky. You have to have some theory of the case on how it all shapes out.

That's why I deliberately mentioned data center spend slowing down. You at least have to contemplate that and say to yourself, “In the 20% chance where that happens, am I done?”

The scar tissue from those 2020 and 2021 deals may be a small part of this Carta slowdown.

6. Vanguard Enters Private Equity

Speaking of riding momentum, we saw Vanguard adding PE exposure. Is this the top?

I think it's a terrible sign.

I mean, look, it's definitely a sign. We discussed this before: The industry keeps looking for new sources of capital, in part because some of its standard sources of capital have issues to deal with, most obviously endowments. That's probably not as true for PE; they're just bigger. But pension funds.

People continue to look at a lot of stuff. It's high-fee-bearing, which breaks your heart as a Vanguard ETF and mutual fund investor.

So what does it actually mean for venture? From my venture seat, with Vanguard adding PE exposure, how much more money is going to come into the industry?

All the PE shops have been doing some version of this. We talked about Coatue doing it in the crossover space. Blackstone and all those guys have been doing it. This is a category, a trend that's going to happen.

I don't think it's going to be the same wall of money as sovereigns or pension funds. It's just another source of capital—another, pick a number, 20% more. What the heck?

Structuring it for venture will be harder because you have to make all these partial-liquidity assumptions, which will get tricky for venture. I don't know if you saw it, but Elsie Stefanik, who continues her crusade to make life miserable for Harvard, is indirectly going to make PE's life miserable because she's been saying, “Hey, Harvard, your accounting is incorrect because half your assets are PE and other private assets, and they're not marked correctly.”

That process is ongoing. The whole process of private marks is challenging, and the less sophisticated the investor—or the less able that investor is to take the long view—the more challenging it becomes. My guess is that the retail investor is least in a position to do that.

Listen, here's a bad sign: They're partnering with Blackstone on this, right? They want to offer PE access in target-date funds. We can talk about the pros and cons of target-date funds. They have their place in a non-taxable account.

What's a target-date fund?

Target-date funds—you don't even need to know anything about investing. When do you want to retire, Harry? I know you're starting—what year would you like to retire?

At 80 years old, so 2074.

Well, we have a 2075 fund for you.

Yeah.

It will start off today at 95% equity and 5% bonds, and each year it evolves. When you hit 78, you'll be—I mean, Rory, you can correct me if I'm wrong—95% bonds and liquid and 5% equity, right?

It's clear that you're not spending any time thinking about retirement, Harry, which is entirely sensible at your age. But yes, Jason's exactly right. These are structured products for the mainstream market. If it wasn't for the fees, they're a broadly good idea: You don't know, Mr. or Mrs. Whatever, how to think about equity versus other things, so we'll just make 1 big decision and land the plane for you.

Sometimes part of the issue is that the fees are high relative to what you get, but it totally makes sense. But, Jason, to your point, you're putting these in there, and 1 thing we know about venture is you can't target your return date, right? If you could, it would be easier. So it's a hard asset to fit into a liquid individual portfolio. I think they'll try. I think it'll be hard.

To Harry's point, if Vanguard has $10 trillion, putting 10% of that into venture would move the needle. But I just don't think ordinary investors should be doing this stuff.

It's chasing return, and that's always the nature of it, but yeah.

The hunt for alpha continues.

The hunt for alpha always continues because that's what we're all paid to do.

7. Universities Face Funding Pressure

You mentioned Harvard there and life being hard for them.

There was a $1 billion funding gap, and Stanford are doing layoffs. They weren't huge layoffs, but layoffs are still layoffs. How do we feel about the health of where they're at, how concerned one should be for them, and how we should think about it?

There are 3 levels to the question here. If you're asking about the impact on venture investing as a source of capital, that's an easy question to answer. They're obviously going to be a lot less active.

Next level down, are you asking about the impact on the institution and the wider societal things? It looks tough. I mean, it's a terribly unfortunate way of making policy because a whole bunch of humanities kids yelled in the press about political issues. You're firing the poor guy who's been in his lab for 10 years trying to cure cancer. It's a very awkward way to make public policy and punish the wrong people to punish the institution.

I don't envy anyone running one of those institutions, and I hope they can figure it out because a lot of what they do is really good. Not all of it, and some of the most visible things are quite annoying. But when you look at the things that are being cut, as distinct from the things where you go, "Ooh, that's a waste of money," the administrators ain't getting whacked at scale.

These are the programs getting whacked: the ones cutting science and small grants to graduate students to do amazing fricking things, just when they're most productive and useful. So I think it kind of sucks as an outcome.

I just thought it was interesting—not in a good way—but Stanford blamed its $140 million in budget cuts on federal research funding, right?

Yeah.

That's the issue you're talking about. But they also said there could be a potential increase in endowment taxes. They had to lay off people and cut the budget due to endowment taxes. There is a cost to these endowment taxes, right? Taxing venture, in a sense, by taxing endowments, isn't free.

It will lead to a decline in investment and a decline in human capital. We did get a QSBS break in the new tax bill. I do like that.

We did. And I would like to know—

I do like that.

I do want to ask that question.

Everyone got a little piece. The deficits may ruin us, as Elon said. We all have to move to Mars. But I got my piece. I got my QSBS to $15 million, and I'm excited for it. I got bribed in the tax deal.

What I'd really love to know, and I just don't, is—I have this vision of this crazy sausage-making machine. There's a whole bunch of agendas to pound it to the solar guys, to pound it to the universities. This is an administration that has its hate list and works down through it pretty methodically.

Somewhere, someone crops up in a meeting. I can see it: "But despite all this, guys, let's cut a deal for the QSBS guys." Whoever that person is—

Who knew that was coming?

Whoever that person is, next time he's running through San Francisco, all he has to do is email me and I got him dinner. You're sitting there, and the way it works is they have a target to make in terms of money. Every time they stick it to someone and charge them money—endowments, another $20 billion; getting rid of the EV subsidy, another $20 billion—and then someone had to say, "Let's go backwards. $15 billion to give it to these guys on QSBS."

So whoever had the juice to push that through—maybe it's Peter Thiel and the Bilderberg conspiracy, whatever—I love it. I'm in.

I wonder where it came from.

Listen, we can explain it too it for folks, but Rory, I'm curious. At your fund stage when you write checks, if you had to ba- spitball or ballpark it, how much QSBS benefit do you get in the end at the GP and LP level? Like, 'cause some of it isn't gonna qualify, right? Some of it's gonna be too big, right?

You get some, but it's not as big a thing for us—

50%?

If that. Less than that.

If that. For folks that don't know, and Harry, you might not even know, in the U.S., my taxes in California are 50%. I know all you Brits are complaining about taxes. I think Rory and I pay more than you.

But we get 1 weird, quirky tax break, which is if you invest in startups or small companies below $50 million in assets—now, I think, $75 million—you don't pay any federal taxes on the first $10 million in gains. Now it's $15 million, and every LP, to the extent they're individuals and to the extent they're taxable, gets $15 million on their distribution too. No federal taxes.

California taxes it, which is pretty annoying, but a lot of states don't tax it at all. So you could have no taxes at all on startup gains. For whatever goofy reason, it got increased 50% in this tax bill. So we all got our little taste.

So basically, you have no capital gains on angel investing if you're investing—

No, not on angel investing.

Okay.

Yeah, up to $10 million, now $15 million. And frankly, you can stack it with trusts. For me, it's probably $40 million to $50 million, because I have 5 trusts on this. So that means for each exit I get, $50 million has no taxes. It's possible to do it to $750 million, but it's complicated.

Rory, Rory, he has a big condo in Yellowstone just to add to that.

He clearly has a big condo in Yellowstone.

Yeah, with his trusts. If I have 5 trusts—

I mean, Rory makes the point: you'll take the nickels, and you might as well pay 50% tax in California because it's the best place. But for me, this is my quiet motivator to do early-stage investing. At least I can avoid $15 million of taxes per deal.

8. AI Rewrites Enterprise Sales

Okay, we're going to do a final one, which is Microsoft. Microsoft laid off 9,000. I know, in the grand scheme of their workforce, it's not huge. It's still 9,000 people, replacing general salespeople with solutions engineers. I was walking with a Clay founder the other day. They don't have salespeople either. They have pretty much the same approach. Is this the future for everyone? How do you see this?

This bothers a lot of generalist sales folks. We could have a longer discussion about how much of sales will be replaced with AI. My rough sense is 30% to 40% of 1-to-2-call sales reps are going to be replaced by AI. It won't be the same as with support, but it'll approach it. It'll be relatively small in the enterprise, okay?

Everyone's like, "Oh, well, you can't go to Pepsi, and AI's not going to show up." But you know what's going to happen? It's what Clay talked about: forward-deployed engineers or solutions engineers, or what Microsoft's doing.

We're not going to have a guy who doesn't know our product in the age of AI show up to big deals. I would rather have a solution engineer who knows this cold, who partners with somebody who's less good in sales. And so I think you better be worried if you're a generalist sales guy who thinks being a relationship guy wins today.

That's Microsoft's point. We don't need relationship people. AI has raised the bar for customer expectations. Here's the important point: it has raised the bar. Microsoft's doing what everybody wants to do: replace folks who don't know my product with folks who do.

What's interesting is it wasn't couched as a "replace with AI" story. It was couched as a "replace with better people" story, and it's hard to argue with that. As a random comment, it's always impressive to me that these companies with 40% operating margins are still willing to grind another point out of it. It's just so capitalistic. It's great to see.

I thought it was super interesting. I didn't see many folks talk about it, but Canva's doing an AI Discovery Week this week.

Yeah.

All 5,000 folks are released from their normal jobs to learn about AI because a lot of employees were saying they were too busy to learn AI. They were too busy at their daily work at Canva to learn AI, so there are teach-ins.

The CPO was very clear: "Come on, Cameron, you've had free ChatGPT for 1 year. You've had free Claude. You've had access. You can pick from a bevy of tools for 1 year, but you're still too busy to learn AI. So here's your week, guys. We're going to have classes, teach-ins, sit-ins. We're going to have a hackathon."

And it sounded great. The angst I hear from folks who've been around B2B for a while, Rory—and you see it on LinkedIn and in person—is, "Rory, I need to be reskilled. I'm frustrated. I need to be reskilled." I think reskilling doesn't work, and I think Canva's basically saying, "Shit or get off the can, guys. Here's your week."

I'm really sorry; they're twofold. I think, 1, if you need a discovery week for AI, you have people who aren't curious enough to want to progress in learning their job. No, seriously.

It's a kind way to say a lot of things.

I'm on the board of a public company, and I got asked the other day, "Well, Harry, what do we do with employees who are not embracing it because they are fearful of it and don't want it?" And I said, "You fire them."

Yeah.

I'm sorry. It's super unfortunate. I do not say that happily.

Yeah.

But if you don't want to embrace it, you are going to make this ship sink. If you need a week for AI Discovery Week, I promise you, you're not going to work, and you're going to take the kids to the playground and go eat chocolate in the cinema. And you know what? You've got to be curious yourself. This is not the way to do it.

Well, it might be. I think a lot of leaders are trying to do gentle messaging and stage it, and I think it may work for Canva, or it may just be their warmer way of doing it.

Yeah.

Right? Here's your week. It's also notice, right? It's notice to those folks too, isn't it?

In many respects, you're both correct. To me, this was performative: “We're letting you all know expectations.” That way, 2 months from now, to Harry's point, if you're deciding as managers that some people are surplus to requirements, we're all humans, and you can sit down with them in HR and say, “Look, you had your opportunity. You didn't take advantage of it. We're going in a different direction, and here's your package.”

Big picture, zoom out. There are lots of different ways to do it. Let's just say we're not talking about the new crop of startups where it's all people who are AI-native. But all these large tech companies who have 10, 20 years of employees—you know you've got to get on the AI journey.

One of my bigger ahas is, it doesn't matter how you do it. Don't over-agonize it. Some people might just fire the people out of the gate. Some people might do a training week. Some people might do a training month. We know the direction of travel. 2, 3, 4 years from now, you're not going to have people who say, “I'm too busy to use AI.” They're going to be long since gone.

You're going to have people who have been automated away, and then you're going to have people who are using the products and are doing the automation. How you do it, maybe Harry's just meaner than the nice people who run Canva, who seem to have done very well, by the way, by being nice. So let's not have any complaints about that.

And Cliff is amazing. He's the co-founder and CEO. He's super nice. Lovely dude. Always been great to me. So that wasn't a shit on him.

No.

It was a shit on his people.

So let me just say, how much time do we all get to get with the program?

1 thing that we did here at 20VC is that every Friday, everyone has an hour from 4:00 to 5:00 where they get to try new tools. It's much nicer to do it in a continuous way where it becomes a habit and becomes fun.

Yeah.

We also do it in the office together, and then we do a show-and-tell from 5:00 to 6:00 on what we've learned and what's cool. Really cool way to do it.

Agreed.

9. Kalshi's Market Forecasts

Okay, so we're going to do a Cow Sheet quick fire. As you know, this is like a speculative marketplace. There are bets. You always want me to say the exact odds, Rory, quite rightly, so I will, exactly for you, this time.

Number 1, this one's a good one. I love Sean, and so I'm going to get in the firing line here. Sean Maguire tweeted some controversial things recently again about your New York mayor thing. Sorry about that. And everyone's really quite upset about it. The question is, will Sean Maguire leave Sequoia this year? Yes, 34.8%. No, 11.5%.

In other words, highly unlikely, and the market is saying it's highly unlikely that he will.

Exactly.

I'll take the bet, then, that he will. And I'll take this actual bet with you, whatever the money is. I'll do the bet for real.

I'll 1,000% do that.

Am I understanding this, that there's almost no chance he's going to leave based on this bet, right?

Pretty much, yeah.

I can tell you why. I'll take the bet that he leaves this year. I don't know him. I'll take the bet.

Do you know his investments? Because if you knew his investments—

Yeah.

—

I know that, and there are different ways to leave. Listen, we're in the age—one great thing about X is we see the grouchy billionaires. We see just how unhappy so many billionaires and mid-centimillionaires are. The happy ones go off and leave their venture, multibillion-dollar venture fund, and join Meta, right?

The unhappy ones just use this as an endless megaphone. And listen, I get his points, but—

But—

They're just so amplified, right? They're just so amplified. But you could leave as an investing partner. You could become a venture partner. You could transition. I'm giving myself credit for all of this bet, Harry, that the answer is nuance.

It's most favorable to me that I'm not saying he's expunged from the website. I'm just saying, if this keeps going week after week after week, I just think it's a sign his head is not into the investing.

I think a lot of his returns have come from working with Elon, and Elon loves this shit. So he's actually giving his client what he wants. He might get promoted for this. Congratulations: you pissed off 2 million people, and the owner of X, SpaceX, and Tesla loves you. You win.

So I actually think, on the merits—quote-unquote—of whether or not what he said or did is going to result in anything, I don't see it. I'm on the no side. I mean, the only argument for taking the yes, Jason, is, look, statistically, X percentage of people seem to leave every year, so at some point the bet's not priced in.

That's what I'm saying. I think this is a tell that there's a higher-than-otherwise chance he leaves this year. That's all I'm saying. I'm just looking for the tells, and as we've learned, we know a lot of folks in B2B who have quietly retired over the last year or 2—

I think it's obvious—

It's not clear on Twitter that they have. We know folks that have, right?

State of the economy at the end of 2025: soft landing, 76%; high unemployment, 16%. There's no actual dollars tied to this one for some reason.

This is easy. Roughly 1 year in 7, the economy's in recession. So you should say, if someone asks you, “Is there going to be a recession at the end?” the default probability is about 16% to 20%.

What this is saying is the economy is no more likely than not. It's a normal probability of being in recession, and I'd say that's about priced right. I don't know what would cause it in the short term. So, my perspective on that, it's a roughly correctly priced bet.

You should wake up every day and say, with no new information, the Bayesian prior is there's a 1-in-6, 1-in-7 chance that 12 months from now things will go to shit. And when you start seeing the VIX climb—and that's why you saw it when the tariffs came in—that 1-in-7 chance goes to 1 in 4, 1 in 3, and then, when things come down, it goes back to normal. What this bet is saying is we think things are “roughly normal now.”

End of 2025, are you saying soft landing or high unemployment, then?

I think it's priced about right. I'm actually indifferent on the bet. I would pass on putting money in that bet because I think it's priced. I don't have differential information other than—

Well, then forget the bet. Just tell me which one you think it's going to be, wise one.

I'm not going to tell you, because the whole point—Actually, that's why you're asking a dumb question, Harry. All you can say—

Kalshi's worth $2 billion. There are no dumb questions.

There's no evidence anyone—

There are no dumb questions on Kalshi.

There's no evidence anyone can call it. Actually, I think Kalshi—by the way, this is the real insight here—is calling it correctly. And I agree with Kalshi. Kalshi is saying that there's no more information about the probability of a recession at the end of next year than there is in a normal year, and I'm with them on that.

But if it is 1 in 7, right, the odds have to increase as you haven't been in a recession—

Yeah.

—for multiple years. So how does that factor into the bet?

It's a very interesting question. I'm going to answer that because it—

What the fuck? He gets an interesting question, and I get a stupid question.

No. Okay, well, you said it, Harry. It—

It's like flipping coins. This is the opposite of flipping a coin: it does make the odds higher the longer the boom is here.

Actually, there's an expression: “Booms don't die of old age.” In other words, what you're saying is, just because it's been great for 6 or 7 years doesn't mean that it has to die in the 8th.

Yeah.

Australia had this run of 17 years without a recession.

Logically, my mind says it should mean that, unless what really happens is recessions. Booms can go on quite a long time unless people do dumb shit. Now, because we're humans, we invariably do dumb shit, but if that didn't happen, it could go on for a while.

Yeah, it was a smart question, Jason.

Booms don't die of old age.

Thanks for that one, dude. That was smart, dude.

I can't say it. Never mind.

Good thinking.

Yeah. Oh, wise one. There's no chance we have a recession at the end of the year. There's just no chance.

The 1-in-6 chance. The 1-in-7 chance. It's twice.

I'll take that bet. I'll put a grand in on that bet.

Well, yeah, take your bet.

Jason's putting bets down all day long. Actually, I would probably do 50 grand instead of a grand if you want to do it, Rory, because the technical definition of a recession is pretty tough to meet. True. That's a good point, right?

So I'll put 50 grand on this one if you want to do it, because I think even if we are, I don't think there's any chance we'll be in a recession at the end of the year. But I think even if we are, it's not going to be called a recession yet, right? It's already July.

I'm going in on this bet.

No, I'll pass, thanks. But that's a good point. You're right.

It's always the highlight of my day. I have the most fun on these calls. Honestly, I love doing this with you both, even if it's incredibly humbling being informed that after 10 years of asking questions, I still ask shit questions. I think you're pretty good.

I wouldn't say—

Thanks, Jason. I love you too, dude. I miss you.

Okay, what else? Come on.

Okay, final one. Linda Yaccarino: does she leave X this year? Jason, this is a tough one, right? As an outsider, I don't see any merits in her. Not from day one, right? It's hard to see.

Sorry.

I just don't see it, right? There are so many public CEOs, and the fact that on a social media platform she's chosen not to be insightful in public, I think is a bad tell, right? Having said that, there's just no way he wants to run this thing, right? Especially the core X. She's not working on AI and Grok, right? She's managing advertisers and the P&L.

So I think if it's not broken, he's not going to make a change. This is a murky one, but I say she's here through the end of 2026.

Wow. Through the end of 2026? You're even higher than—

He just fired his head of European sales for Tesla too? I mean, you can't. It sounds fun to fire everybody on the team, and there are moments in time when you need to, but, man, you gotta keep a few pieces going on the board. Even if you're on Adderall or whatever he's on, there's only 24 hours in the day. I think he's going to keep this soldier in place.

Yeah. I think the interesting one, obviously, this week was the market response to the new party and, correctly, the Tesla response to the new party, which was shareholder barf. I was just reading—

Straight away, you guys both said no to that, by the way.

Harry, the definition of a good forecaster—and we've talked about Superforecasting and Philip Tetlock before—is you update your priors with new information. The fact that he went on to form a political party, to just go down that route, would update your priors on whether he would still be CEO at the end of next year. You kind of say to yourself, increasingly, are you the right person to be running that company?

I think I was still right.

I don't think he formed a real political party the way the bet was. That's why. Not to spend too much time on it, I don't think he's going to run candidates in all 50 states. He's not. This is a tactical move that, if it happens, is at the edge of what a political party is.

As much as I love Kalshi, you gotta give me a little credit for definitions here, like recession or political party, right? And Kalshi was a lawyer, wasn't he? Jesus.

Yeah, he was.

That's right, yeah. Come on.

No, what he's apparently doing is cleverer than I thought. You're just going to contest a few elections where money matters, right?

And that's it. If that's what he does, it's very different, but it's smart, right?

Jason, you've said very clearly yes, you think she will. Rory, are you quite as blunt as saying she serves no purpose?

I don't know enough to comment, so I'll pass on that. As I say, I'm more willing to be opinionated on the wider CEO issue, but I'm not opinionated on that. I don't have a clue.

I mean, she has a tough job, but if you run a social media platform, how can you not be out there? It's just weird, isn't it?

Is she coming on 20VC anytime soon? Let's get her on, Harry. I want to see you and her just— Tell her it's a friendly environment. I want this question answered.

Final one, it's not a bet, and it's super quick thoughts. Does CoreWeave and Circle sustain the super-high stock prices? Does the meme-ification of their stock deflate?

If it is a meme, then of course it deflates. The real question is: did investors at the time of the IPO significantly underprice the asset, and have people now realized that, in fact, the correct price is the much higher one?

When we looked at the Circle numbers at the price they're trading at, it felt very lofty, relatively. Traditionally, most of these post-IPO or pre-lockup stocks drift down significantly if the valuations are way above any kind of near-term fundamentals. So you've gotta believe the downward pressure is there.

Right, boys.

Time to wrap. This has been so much fun. I love doing this. Again, Rory, thank you for coaching me on questions. I'm getting better. One day. One day, I know.

I didn't mean to be rude.

No, no, no.

I actually thought Jason's was a really interesting one. Seriously. It's the question of whether the probability of a recession increases every year that goes by without one. It's the classic question in statistics: are they independent or correlated events, or is there some kind of buildup?

There's no reason that there should be a buildup, but in fact, there is, because humans. The reason there is a buildup—the reason Jason is more right than me, actually, upon reflection—is because when times are good, people pile up the dumb, aggressive shit to make money, and the minute something goes wrong, it blows up in their face.

So, at the margin, I would actually revise my opinion and say Jason is right. The longer it goes on, the more likely it is eventually to blow because of—actually, it ties to your CoreWeave comment. Stocks get high, things get expensive, people take on debt, they do aggressive things. You wake up one day to Charles Kindleberger's “Manias, Panics, and Crashes,” and you're done.

Right.

And just to wrap, Harry, the one thing I will say—in all seriousness, I don't know anyone in tech media that can cross as many domains as successfully as you can. A lot of respect to the client.

You can get really good at a narrow thing, but I don't think there's anyone else who can go as deep and thoughtful as you can, especially over the last 2 years. The type of guests, the domains—I can't think of anyone else that can go that deep and thoughtful. Kudos to you for that one. It's easy to go deep on one thing or just be a student, but you're pretty much S-tier.

I super appreciate that. I leverage wise friends like you both, so I appreciate you both. This was good.

Hasta la pasta.

My word, I was almost getting emotional hearing that from Jason.