Chime IPO：IPO市场比以往更热吗？

There’s no way that Scale can recover from losing its founders. I love you, but it’s a dead man walking instantly. I could imagine an entire IPO cycle where, once again, neither of the 2 names decided to go out, and we get to the end of the year. I’m starting to lose confidence in the old guard.

In situations of ambiguity, the person who has the most leverage has the best chance to win. I would argue Microsoft has quite a lot of leverage because it’s not existential for them, especially now. I think they have no leverage. Actually, honestly, I’m not trying to be a chaos agent like our prior guest.

1. Meta’s $14.8B Deal for Scale: The Analysis

I’m so excited for this. It is my favorite time of the week, and what a week of news we have to go through. We’re also thrilled to be joined by Garrett from Handshake. I wanted to start with the news of Scale AI and the acquisition there: Scale AI, $14.83 billion to Meta. Garrett, how should we read this? You’re the expert in the room. How did you take the news?

Guest

Well, I think it’s pretty unprecedented to have one of the leading players completely bow out. What we’re seeing is many of the labs having to reallocate their spend and trying to diversify away from some of the leading players that they no longer can trust to do work.

Right now, I have a bunch of questions I want to ask Garrett. At a deal level, it seems pretty clear: it’s a weird deal. They’ve put almost $15 billion into Scale. Rory may have the details, but most of it comes out as a dividend, and the CEO leaves to go run a seemingly much broader portfolio at Meta. He’s not just running training, right?

The whole mystery is that Facebook, or Meta, doesn’t even seem to care about the revenue. If only OpenAI is committed to maintaining that as a partial presence, I assume the revenue will decline rapidly. I don’t know, though. You’ve already benefited from that, right? Almost overnight, you’ve benefited from that. In a way, you are the mole here rather than us.

Guest

Agreed. We have seen a huge surge in demand. Our primary concern right now is hiring on our team. Demand tripled over the course of the week. I’m running on an average of 3.5 hours of sleep for the last 10 days.

I think I’m going to zoom out even 1 level, just to restate it, because we’ve dived right into the details. Zooming out a million miles, Scale AI is an amazing company founded 10 years ago to help companies build great AI models. They do that by rounding up human experts and initially just doing very simple data labeling.

Increasingly, over the last 4 or 5 years, as the tasks that the model builders have to attempt to solve get more complex, they’ve been hiring experts at the PhD level to help with post-training reasoning questions around making the models better. Zooming out, it’s an amazing business. It grew to $800–900 million, at kind of 1 level below the model providers themselves.

This is a company selling to those 5 or 6 amazing model-provider companies a vital service that’s probably a pain in the ass. If you’re OpenAI or Anthropic, you don’t want to spend your life rounding up literally thousands of people to answer what, 10 years ago, were very mundane questions like, “Is this a stop sign?” Now, as the AI has gotten smarter, they’re much more advanced questions, at the level of PhD knowledge, as Garrett knows better than me.

That’s the business Scale AI was in, and it was doing $800 million. Its customers were mainly the 5 or 6 large model providers. That’s where we were a week ago. It was a highly valued company that had just raised at a $14 billion valuation, and probably the leader in the space in terms of size—not making any comments on quality.

Then, as you say, a week ago, Meta announced a fascinating transaction, and one that definitely is a bit of a head-scratcher. They invested $14 billion into Scale AI, took 49% nonvoting control, and allowed the other investors to literally take that money back out as a special dividend. All those other investors got $14 billion in cold, hard cash, and some of the key executives at Scale AI moved over to effectively work with Meta.

Those are the facts on what happened. The open questions we’re starting to think about are: What’s the impact on Meta? What’s the impact on the remaining Scale AI business? And, as Garrett says, what’s the impact on all the other providers?

If you think about it, Facebook is another contender in the AI model wars. If you’re OpenAI, Anthropic, or any of the other guys, you now have a key supplier selling you a pretty important subsystem of what it takes to build your model, and that supplier is now owned, controlled, dominated, infiltrated—pick your word—by one of your direct competitors. It’s got to make everyone pause, which is exactly what you’ll see happen with Windsurf being turned off by what was likely Anthropic.

Back to Garrett: what you’re saying is that, in the last week, the phone’s been ringing off the hook?

Guest

Absolutely. As you talk about the shift from generalists to experts, we haven’t even talked about the future. The future is going to involve more audio. It’s going to involve more tool use. It’s going to involve more trajectories as some of these agentic systems and step-by-step problem-solving are improved by all the frontier labs in their pursuit of AGI.

What that really means is that you need experts in domains. Right now, we’re focused on core STEM skills and also skills you’d imagine—finance, law, and medicine—the large markets that these frontier labs are chasing after.

One key question for you, though, Garrett, as you’re having these conversations post-announcement: if I was running procurement or was a VP for one of the other AI companies—and this has happened to me once—how do you think it will impact how they contract with people like you, companies like you, and how much of the process they’ll let you have visibility into?

One could argue that the asset Meta may think it has bought is that Scale AI has a lot of knowledge, just by virtue of the questions they’re being asked and the kind of expertise that people are seeking from them, about where the most advanced LLM companies are going. If I was procuring Handshake after that experience, I might have some different perspectives on what I can let you see or not see. Is there anything you can comment on there?

Guest

When we’re talking to our customers, we’re talking about the absolute frontier tier of what’s happening. Our customers really care about 3 things that are always in balance.

They care about quality first and foremost. You have to have high-quality training data and high-quality evaluation sets. Then they care about volume. It’s really hard to get to scale on volume if you don’t have an audience. I would say the only durable moat in the entire human-data business is access to an audience.

Then they care about speed: how fast can you turn them around? If you’re another company in the space, 1 of our advantages is the ability to activate volume and quality quickly. We don’t have to run month-long advertising campaigns to make that happen.

2. Will Scale Lose Their $800M ARR? Will All Customers Leave?

In a year, Garrett, over or under, 1 word: does Scale have over $100 million or under $100 million of revenue?

Guest

Over $100 million.

That’s a great call, Garrett. The reason it’s a great call is it was such a savvy thing, because you’re probably correct: it takes a long time to go from $800 million to $100 million. So you didn’t diss anyone, and you also gave a probably factually correct answer.

If I asked a different question, which I’m now going to ask, do you think the revenue goes down? Do you think customers will reallocate significant spend away from Scale AI to other providers of data in light of this acquisition? How would you answer that question?

Guest

What we’re seeing right now in the market is that there are hundreds and hundreds of millions of dollars of spend trying to be reallocated to leading providers. The primary constraint is the ability to deliver volume and scale. We can deliver quality, and we can do it really fast.

After you’re wildly successful in this entrepreneurial endeavor, there is a career in politics ahead of you. The way you answer those questions without avoiding the pitfalls is excellent.

But you’re right. What you’re saying, basically, is yes, there’s a massive spend reallocation away from Scale, which gets to the interesting question about the deal: what did they get for their $14 billion?

I know what the investors got for their $14 billion: they got $14 billion, and it’s just interesting to speculate. I think everyone’s really confused by this in a way that they shouldn’t be. Zuckerberg was behind with Llama. He needed to show the public markets that they were still a frontrunner with an AI slant. Bluntly, that was the play.

The price is less than 1% of the market cap. Honestly, he doesn’t give a shit, and he gets good talent with some people that he likes. Roll the dice.

Guest

I agree. That is the analysis. I was asking the question, but you’re right, Harry: implicitly, what you’re saying when you make that answer—and, to be clear, I 100% agree with it—is that the deal was about all of those things.

Guest 2

What you're saying is this is not a corporate. There's no internal DCF that says why this is a good idea. There's just, “I want to be relevant. It's less than 1% of market cap. I'm doing it. Somebody pay for the bill.”

If you piece it up logically, you're right. You gave a company $14 billion for half of the company. The $14 billion has moved out of the company, so that's now gone, right? It didn't move out as a repurchase of shares, so you still only own 49% of the company. The cash is gone, and therefore you now have half ownership in a business that obviously doesn't have that $14 billion and just has whatever revenue it's had, and we just agreed that's declining.

The value of that asset—it's not nothing, but it's nowhere near $28 billion post, right? We now have an $800 million declining-revenue business with so-so gross margins. The amazing thing you've got is, you're exactly right: the talent of those people that you brought across, and the talent, and the messaging for a company, and the messaging, and the knowledge.

3. Is Alex Wand Better than Bret Taylor?

Well, I'll tell you, if I had to simplify it, I think the question is: Is Alexandr Wang as good or better than Brett Taylor? Because, look, here's my view. Salesforce spent $750 million to buy Quip in 2016, a word processor that barely worked, for 1 dude, right? One of the greatest of all time—the CTO of Facebook, right? And then almost CEO of Salesforce, except he bailed, right?

Rory's really good at this math. What is $750 million in 2016, when great exits and IPOs were lucky to be $1 billion, compared with 2026, 10 years later? It might be the same as Scale, like the multiple. It might be 20× bigger. So 20× $750 million is almost exactly the deal size.

I don't think he's saying that. I think you're placing emphasis on, bluntly, the talent acquisition. I think $27 billion of the $28 billion is on messaging. It is about showing the public markets and the world he is still a front-runner.

Important, in fairness, to say it's only $14 billion, not $28 billion.

Guest 2

Yeah, $14 billion. So I think the math to Quip is actually pretty good. Rory, you're better at math than me: $750 million in 2016, $14 billion today, for 1 dude, basically.

I mean, who uses it? Just you? Raise your hand, Garrett. Help me. How often do you use Quip for your word processing?

Guest

Pretty often. Every day.

Don't be mean. How are the LLMs?

Guest 2

No, what I'm saying is, it seemed crazy, right? It's got to be. But if it's the same deal as Google buying Bebop to get Diane Greene for $400 million, right? That was the generation before Quip. Everything's bigger. Everything is bigger now.

There are 7 companies who can write stunning checks, and it only costs them a quarter's cash—literally 1 quarter of free cash flow. Therefore, if you're in the path of the corporate imperative for 1 of the top 7 companies, you can just get huge amounts of money. And you're thinking, “Oh my God, this is the most amazing thing of all time,” and they're literally thinking, “Tick, that's Q2 done. What am I doing in Q3?” Right?

Harry, broadly speaking, we're all saying the same thing. Facebook's market cap is $1.7 trillion, and its free cash flow for a quarter is $15 billion. Even if it's a total write-off, Mr. Zuckerberg literally goes away and, in 90 days, comes back and says, “Whoops, that was an error, but we earned it back. Keep moving here, people.”

Garrett, what if Microsoft put down an offer tomorrow for $4 billion to acquire Handshake? Would you say yes?

Guest

What? Absolutely not.

I love it. Jeff at GGV is just calling me. He says, “Fucking sell at $4 billion.” I tell all founders to sell now, by the way.

Guest 2

No, just as a challenge. I'm not saying you shouldn't take this fake offer, but this is my life lesson: I tell you to take it, and then if you come back the next day and say what you just said, which is, “Fuck no,” it's the right answer, right? But I've decided I want to be the one—as crazy as it sounds—I want to be the 1 guy to tell you to sell, as a challenge.

My observation is no one's opinion on what they would do at a hypothetical offer is worth a damn, and what people actually do when the money comes into view tends to be very different on both sides.

Listen, Garrett, you're going to get a term sheet from a $4 billion price on the back of that. Congratulations. This was a very well-spent—

Guest

I'll put in $100,000, Harry.

Guest 2

I'll put in $100,000. I believe in Garrett. Especially if I get 10% off.

Can I ask Garrett 1 question, Harry, before—just, if we have time? Just because it's a meta question for founders out there and VCs struggling with all this demand that you have now. Are you an AI company? Is this a 2-product company? Because I feel like you, in a way—listen, you've worked so hard, right? But you also have a little bit of luck. This has bounced for you, right? You've become an AI company. When did you found the company?

Guest

9 years ago.

Guest 2

Okay, so you weren't AI-first. You were a network, when you were sort of the network for jobs to university students. How do you think about it yourself? How do you think about taking advantage of these opportunities? What's the meta lesson?

Guest

Harry, you're the boss. I just think it's very interesting because so many founders are trying to become AI startup companies, right? But they're not quite hitting it, right? They're launching features and products, but the growth isn't there.

I think this just accelerates our mission of democratizing access to opportunity and becoming the number 1 job platform on the internet.

Guest 2

No, that's not the takeaway, Garrett. Sorry. No, the takeaway is: stay in the game long enough. Don't burn too much cash and, bluntly, get out of business too soon.

There was a business aggregating students, and they just spent the money badly, and they ultimately didn't survive. The business is about being ready to play when you're called on the pitch.

4. LPs Just Got $14B Back. Are They Reinvesting?

Garrett, we're going to let you go, my man. Great to meet you, Garrett, and congrats. Rock on, man. Great job.

There's a man who's going straight back to the phone to start returning customer calls. Good for him.

5. Jason @ Discord Being Removed

Now, the thing that I do want to discuss on the back of that, guys, is the $14.8 billion back to LPs as well. Is that a lot of money? Do we see a resurgence in LP activity, reinvesting that money, given what has been a period of illiquidity and people struggling with liquidity? Do we see LPs return to the game at speed with Chime, with Circle, with Scale coming back? I mean, it's got to help.

Guest 2

Look, there's no doubt that liquidity this year is going to be way up on the last couple of years. It's got to help. I think, when you're dealing with a couple of trillion of NAV, terrifyingly, $14 billion—let's call it $20 billion, round up—is 1%. It's a good start, right? But that's the terrifying thing about big numbers: they're big.

LPs, at least the LPs I interact with, don't get ahead of the distributions. What I mean is, yeah, Scale, as near as I can tell, has already distributed cash as a dividend, right? It's out, and folks may have already wired the money out to their LPs, right? Have some fun, right?

6. S&P Prediction, iPhone Assembly in the US, and Rory’s Rants

The Chimes, the Circles, the rest—the 6 or 7 great IPOs—a lot of the top VCs are going to be managing those distributions out over 36 months following a lockup, right? So my LPs don't—they're looking at dollars out now, this quarter, before they get excited about putting money back in, even though they should get excited because they can see it, right? The money's coming back over 36 months, and you can do a model and model whether it's going to be more or less than its current price. But I haven't seen that excitement until the cash is actually back.

I think that's what's so unbelievable about this deal, though: how rare do you get cash back this quick? No lockup, no delays.

Guest 2

Yeah. It's actually totally weird. And I think it's even more unusual. In the sense of the deal, it's already closed, right? And it's such a workaround on the Department of Justice that, I'm sure, somewhere in the DOJ, someone's head is pounding. Under the prior administration, they'd be trying to do some kind of retroactive restraining order.

It must make everyone's head hurt because it's such an obvious workaround. If the real asset is some combination of the people you hired at Facebook—I'll call it Meta—and/or, as Harry said, some kind of market perception, it's very hard to imagine a DOJ rescission that says you can't work at Facebook anymore. You have to go back to where you were.

I think it would be very hard. This is such a weird transaction. It would be very hard for any kind of judicial review to unwind it, in part because it's so weird.

So I think: clever, clever, clever, and amazing. Clever, but you have to realize you're getting—it's clever if you want nothing. They've bought no revenue, no assets. I mean, if Handshake has already gotten, say, $50 million ARR from Scale already, this is the worst—as a revenue asset, they don't even own it. It's the worst purchase ever of all time for assets, right?

And that, by the way—your first sentence, I can't remember what it was—was a very succinct summary here. I thought that should be one of the sound bites because you nailed it exactly, Jason.

Guest 2

And you're right. The funny thing is Scale AI is still the same amazing company it was, right? 2 or 3—I don't know how many—talented people have left, and Alexandr is obviously wildly smart, but you've still got hundreds of really talented folks there.

But the issue, as we said, is that the impact on your customers is so traumatic that it may be very hard to get back from it.

Guest 2

No, I mean, I love you, but it’s a dead man walking instantly. There’s no way that Scale can recover from losing its founders. It’s too dynamic a business, and it can manage 200 of its 800 employees for 24 months, but it’s a dead company.

Rory O’Driscoll

By the way, first of all, I agree with your conclusion. I’m just being precise. It’s not because, quote, “You’ve lost your founder,” because a company can survive the loss of 1 person. It’s the way it happened.

It sends such a signal about your ability, as a customer of Scale AI, to continue doing business with them when 49% of the company is owned by 1 of your competitors in the LLM space. So I agree with you 100%. Practically speaking—I’m trying to say it nicer. I’m working on being nicer—it’s hard to imagine an independent, viable business selling to the LLMs who are competitors to your 49% owner, where your founder and charismatic CEO is currently working.

When you put it that way, it’ll be fun. In 3 or 4 years, we can do an analysis, or you can help. It’ll be the AI startups. What happened to all these companies, like Scale, that were sort of bought?

We got a big one called OpenAI, which was sort of bought by Microsoft. They’re trying to undo that one right now. But the rest of these little startups, what will they look like in 4 years? That’s a fun fact.

Rory O’Driscoll

Yeah. They’ll pay $14 billion for Scale, but they’ll leave $200 million in the bank. That’s what they’re doing. They’re leaving a little pot of honey in the bank. As folks leave, because they’re all cashed out, when we’re left with 18 employees, the $200 million can last a long time.

I think the interesting thing for me as an M&A investor is: How is the spend distributed across the other players? And is there an additional acquisition that the others feel they have to make as a result of this acquisition to compete with Meta acquiring Scale?

Guest 2

Yeah, agreed. And I could be wrong on this. I agree that that’s the question because, as a good investor, Harry, you’re right: the only response to any great outcome is, “That’s wonderful, but what does it mean for me?”

Harry, you know me so well at this point. I think this is going to be a unique thing. I could be wrong on this, but I don’t think everyone else is going to wake up and say, “OpenAI is going to say, ‘I need to—’” Well, maybe they might, but I don’t think they’ll buy a person like this happened.

You might see some of them say, “Hmm, I would prefer to take some of this work in-house, or parts of this work in-house, because I’m giving too much information to a third party.” Therefore, maybe I continue to do business with Scale AI, but perhaps more of the process is handled by me. Or maybe I buy a small one of these.

But I don’t think you’ll see someone else do a crazy, “Let’s hire the CEO, buy 49%, give Harry the money, and let Harry divvy it out and pretend it’s still a company.”

I mean, the 1 thing I just want to highlight—I always want to do this—is that I think it’s nice. We hear enough about downtreading on people; in terms of bringing people up, Accel made $2.5 billion. Well done, Daniel Lavine. Amazing seed bet early. Paige Craig led a pre-seed—a 1,000x return.

The dude is well known for passing on Airbnb, which he very humbly always talks about.

Rory O’Driscoll

No, he humbly talks about screwing it up. It’s an even better story for him.

Yes, about screwing the deal up, right? I think that’s awesome news there.

In terms of CEOs moving around, as we heard with Alexandr at Scale, Delian tweeted about Jason Citron at Discord being removed, potentially by Benchmark, due to a delayed IPO of Discord. I wanted to hear what you guys thought of that.

If you think about it, I’ve always been interested in the threads on X, everyone hating Bill Gurley for replacing Travis Kalanick. I always thought you want to strongly bias toward backing the founder the whole way. I think that’s been a really good move in venture in the last 30 years—not because it’s morally right, though maybe it is, but because it’s savvy.

Whenever you have to change out your founder, by definition you’re going to lose a couple of years, and it’s a pain in the ass. So, even apart from the moral issues, it’s just not good for investing. But I also feel sometimes maybe you do have to make a change. Rory, have you ever fired a founder?

Rory O’Driscoll

I’ve done it, yes, mainly consensually, and I’m proud of the fact that a couple of the founders I’ve said needed to be replaced have actually been references for me thereafter.

I believe almost all of the time, even when a founder is struggling, you do a lot better by working with them and saying, “What does success look like? Can you get there?” Agree on what success is, and if you’re not getting there, then most of the time people say, “Maybe you’re right. Maybe I can get someone else to do this.”

I very much believe the following statement: If you walk in and fire a founder or any CEO and they’re surprised, you have massively failed as a board member, because you didn’t have the guts to tell them in advance that you were worried.

Personally, I don’t have the capability—or, forget about interest, I could say I have no interest, which is true. I just literally don’t have the skill. For me, if a founder came to me and said, “I’m out,” I don’t have a guy.

Guest 2

One of the deep reasons for not making a change is, “Oh my God, it’s such a lot of work,” and you’re right: You rarely have the guy. I always tell people this is open-heart surgery, even if the founding CEO wants to make the change. This is open-heart surgery, and you’ve got a 1-in-3 chance of dying. It’s really brutal.

Rory O’Driscoll

Just a little while ago—I haven’t updated it, but I don’t think it’s going to change—I did an analysis of all the B2B IPOs. Of the IPOs, 90% still had the founder CEO as CEO. And as near as I can tell, all but 1 had an elective step-down: “We just don’t want to do it anymore,” as with the PagerDuty story and a few others.

So, let me contrast this a little with Discord, because Discord is not a B2B company. It has elements of it. If we’re aiming for these great outcomes and nothing else matters today in the B2B world, then if the founder CEO is out, I’m out, because there’s no hope in B2B. This is just my view: If the founder CEO is out, I’m out.

If the founder doesn’t want to go the distance and works hard to hire someone great, it’s a little riskier, but you can have amazing outcomes, because that’s what the facts say. Violent change is very hard. I could come up with some examples of where violent change worked, but it’s just so hard, typically.

I mean, the 1 canonical example, as you say, Jason, in consumer is Uber, which is why it attracts such attention. It was a violent change. It was very controversial.

But if a board member feels it’s his fiduciary obligation—and this is where I do give Bill Gurley credit—if you’re on the board of a company that’s raised $10 billion in money from third-party investors, I don’t think you can just say, “I religiously don’t fire founders.”

If you come to believe it’s your fiduciary obligation to make a change—and again, we can’t run an alternative history. You don’t know what Uber would be with or without Travis Kalanick versus with or without Dara—but in particular, in light of the kind of backlash he got from it afterward, I do give credit to anyone who says, “I’m in the boardroom. There are lots of people’s money on the line. I’m going to be a good fiduciary and do the hard thing, not the aesthetically pleasing, soft thing. I’m going to step up to my duty.”

Guest 2

I don’t know the exact story with Discord, but if I had the choice as a seed investor, you have 2 choices. Jason Citron—he’s clearly 1 of the best founders. He seems like the—I followed him since the beginning. I’m a superfan from a distance.

If I was the seed investor and you said, “The company’s plateauing. Jason doesn’t seem to want to go public. We’re worried about him,” and you have 2 choices—we’ve done the math. The last round was at $15 billion. We can either bring someone in and try to go for a $20 billion-plus IPO, or stick with Jason. We’re probably going to have a $10 billion IPO. I’d stick with Jason in a heartbeat.

I’d stick with the risks of the founder that, okay, we’re going to have a somewhat worse outcome. I’ll take my—if I invested at $20 billion pre, what’s the multiple with dilution? I’m there for it. I don’t want to do that. I don’t want to. And I think what Bill Gurley was doing at Uber—if they did it with Discord, I’m sure this is what happened—I don’t think—he’s clearly a generational founder, right? But he’s not the guy running Starbucks from Newport Beach, whatever that guy is, right?

I’d rather make less money. I’m okay just making $1 billion, Rory, in carry—or $200 million in carry—and sticking with the founder. In all seriousness, I’ll take it.

Rory O’Driscoll

I’ve genuinely had those conversations. I’m on the same side as you, and it’s not just because less is—I want less. It’s often because it’s less but with a higher certainty, because I go back to my comment: Whenever you make a change, you’ve got a 1-in-3 chance that the person you hire is an empty suit and a freaking disaster.

So, I hate making change, but I’m going to say it again just because I can’t stop: I do think there are still sometimes when you say the combination of reasons, including for fiduciary reasons, means you do it. Dear God, I hope that doesn’t happen a lot.

If I'm a founder, okay, what do I do if Benchmark did push them out? Right? And Sequoia used to be clear, like, “We'll find the right CEO, whether it's you or not, but you're going to make a lot of money.” That used to be part of Sequoia's pitch, right? It may not be you. I think founders should weigh that in; it is an important factor for founders to consider.

Rory O’Driscoll

It is. And when you sell 3% of your company at Demo Day at a $60 million post-money valuation, it's still your company.

Totally. It's still your company.

Rory O’Driscoll

It is. More and more founders do ask about it. They should. First of all, I totally agree with that. It's relevant, and I think, actually, as you'll talk about founders for a while, it's pretty clear that, early on, part of the reason Sequoia weren't in Facebook is residual frustration with Mike Moritz over prior situations where they had, in fact, made a CEO change. That's why I think the pendulum, which was way too far over on the “always make a change” side, has massively moved the other way.

So if you had to pick—maybe the clear statement is this—if you had to pick 1 default mode, and you could only have 1 button that you always press, the same button, which would you go for?

Rory O’Driscoll

The “never change” button, because it's statistically the right outcome.

7. Ramp Hits $16B Valuation: Are We Back in 2021?

Okay, we're going to move on. There are a couple of other rounds that I do just want to cover. I want your wisdom on this, guys. Ramp announced today that they've raised at a $16 billion valuation. These guys just never stop fundraising; the price just goes through the roof. Amazing—and this is not criticism, it's astonishing.

So, 1: how did you think about raising at $16 billion? And 2: guys, I don't get it. Brex is like 101, and Mercury's at $3.5 billion with $500 million in revenue and owning the banking relationship, which is more valuable. How do you guys think about this?

Rory O’Driscoll

Just to frame it like this: it's 1% dilution. Here's the thing: it's $200 million at $16 billion. Some of these rounds you look at them, they're great, but they're 20% dilution, 30% dilution, right? That's an expensive unicorn, right? 1%—I mean, I know it's a lot, but we're not really going to notice the dilution on our cap tables, are we?

I think Harry's not asking why Ramp did it. I think Harry's asking how it pencils out for investors. Am I correct, Harry? Is that a good question?

Yeah. Why would you pay $16 billion? I mean, I think it's a better comp—

Rory O’Driscoll

You made 2 statements: you compared it to Brex and then Mercury, and you were implicitly saying, “Harry, why would you do that at $16 billion?” My sense of it is that they have a very nice business, which has both card and software, and has headed more upmarket—has also moved upmarket—and the quote-unquote competitor here is a combination of the card business from Amex and the accounts-payable business from the accounts-payable suite of SAP and Oracle.

So they have a software-plus-kind-of-transaction business, which is pretty nice. They've got hypergrowth, so I get it in terms of just the explosive growth they're seeing in that space, though I think it's slowed down. The million-dollar question on all these fintech companies is how ultimately do they trade if and when the growth slows down.

That's probably a question you'd be asking yourself if you were paying $16 billion for that business, for what looks like a $700–$800 million-revenue, pretty reasonably high-growth business. It feels lofty, but as we're going to discuss, if we get to our agenda—which I doubt—Founders Fund have shown an uncanny knack for getting these things right, so I'm not going to bet against it.

You're right. Ramp's probably at basically the same amount of revenue, right? Mercury's approaching.

Rory O’Driscoll

I don't think Mercury's bank relationships are worth anything, Harry. I think they're worth zero. I could tell you why. But putting that aside, here's the theme that I see in the private markets, which I think is weird: let's assume Ramp is the fastest-growing of the 3, for the sake of argument, because it has the highest valuation.

It should be.

Rory O’Driscoll

I still feel like, for startups, we're giving revenue valuations that don't have an adjustment for the different ways their comps trade in the private markets. And if Ramp is in a space that's 2–4× revenue when it's public—or whatever, you could look at the comps, 3–4× revenue—see, that's the part I don't get.

A lot of fintech revenue is lower margin, or the public markets value it lower, and that's the part I don't get, even in my own portfolio, when I see valuations that value 80% gross-margin products the same as 20%, 30%, or 40%. That remains a mystery to me in the private markets, where everyone gets the same ARR valuations, even if the quality of revenue—or even the public comps—the public comps are different, right?

So I totally agree with you there. I just was like, “Wow.” And just the speed at which they're always announcing fundraisers—I mean, it's every 3 months Ramp has a new round—which is, again, great. I'm just like, “Wow.”

Rory O’Driscoll

1% every 6 months with a little bit of tender offer thrown inside is no big deal, but I hear your point. As a reminder, it is probably, at the margin, a cash-consumptive business because it is—I mean, just like Amex, it's giving people credit cards that they pay for things with, and the merchant gets paid the next day, and then someone has to float the card for 15 to 30 days.

It's not a lending business in any long-term sense, but on a short-term basis you are lending money, and you have to fund that capital. So there probably is a capital need. Now, I don't know how much they're doing that with securitization or receivables, but fundamentally, this is not the kind of company you can run on the cheap. It's not like a software company where you can be down to your last dollar and juggling payroll, as I'm sure Jason did many times when he was starting out.

When you're starting out, you have big-ass receivables, right? I remember doing the math, and you have to make a bunch of assumptions. If you have revenue of $1 billion, then if you're getting 2% or 3%, you can work out the gross transaction value per year. And then, if you say to yourself, “Settled over 12 months,” you probably, if you're doing $1 billion in revenue, have roughly $4 billion in receivables you're financing.

In other words, you're advancing money—maybe $2–$4 billion, depending on the credit cycle—to your customers that you have to fund on the balance sheet, and that's probably where some of the money goes.

8. Ramp vs Brex vs Mercury: Who’s the Real Winner?

Totally agree. I actually think it's a really interesting way to do relevance. The most important thing today is to be relevant and to own consumer attention, and I think by having frequent fundraisers and frequent media hits, it's just a great way to continuously keep velocity and momentum ahead of someone like Brex.

It's actually important, Rory—in the nicest way, I would push back on you. You went—

Rory O’Driscoll

Yeah. No, I do think it is, especially because most CFOs—a lot of consumers—don't know which one to pick. You have a brand; it's all brand anyway, right? And Ramp, Brex—I don't even know. But Ramp is always on Twitter, and they've got Saquon Barkley, or whoever this random footballer is, running with Ramp on, and that goes viral. You've got to do it. They're crushing.

You're right, Harry. You did call me on it, correct. I did make a little face, but I'm willing to say you're correct.

Listen, $16 billion at Ramp. Perplexity is raising again. First tranche at $15 billion, second moved to $18 billion because there was so much demand. Rory, is it the same as you said last time, which is just, “Hey, it's an infinite shot at a massive $500 billion outcome. Roll the dice”?

Rory O’Driscoll

The 2-step price-uptick thing is weird, but you're seeing some of that now, where I think the 1st round gives you certainty and then the people who didn't get in say, “I'll just pay 10% more. I just want in.” It's just a sign of deal heat. It's a sign of the market right now, and rational CEOs take advantage of it.

Listen, we have OpenAI creating more news, as always. We have a $200 million defense contract from the Pentagon, which is actually very sizable. It's like the largest the Pentagon has given to a single provider. How did you guys read that? Also, OpenAI's 1st foray into defense in this way?

Rory O’Driscoll

I mean, I didn't—I mean, first of all, stepping back, aggregate comment, it's probably good news.

I just love the fact that the Pentagon is spending money with all these venture-backed, high-growth, very technically savvy, forward-facing companies. I think it speaks to some level of improvement in the procurement process. I think Palantir started that, SpaceX, and all that.

Now, I think it’s great because you look at the news right now, and there are wars in other places, and you go, “I want to make damn sure that we have the best stuff if we ever go to war.” It looks like that’s not being made by the old fuddy-duddy companies. So, big-picture good news. Go team. I don’t know what you guys think.

Guest 2

I just think, related to what you said, Rory, that for OpenAI to achieve its mission, with the market share they have, they have to be friends with everybody. Yep.

Before 2020, there were so many principled CEOs. We were so principled about issues, about not working with defense or the government. There are many things folks are principled in; in fact, many things that are very aligned with my values that folks are no longer principled on. You’ve got to be neutral.

With the limited press I saw on it, Sam said, “This is only $200 million. This is not material to us. This is a small deal.” Right? He probably didn’t do a customer call on this one. I mean, maybe he did, but it is immaterial.

You can’t be Republican or Democrat or anti-this. The threat to OpenAI from being political is so high, right? You’ve got to be friends with everybody. Other than rogue states, I think OpenAI has got to sell to everybody. They’ve got to own it.

For the record, he may want to be friends with everyone, and he doesn’t appear to be doing a great job of staying friends with Microsoft. What happens there with the Microsoft feud?

Rory O’Driscoll

Listen, I think they’re lucky in that Microsoft crossed the line first. Microsoft basically bought OpenAI, and they’re going to end up being a 33% shareholder of a standalone company. This is the opposite of scale in some ways. They’re de-scaling.

They were a company that could not scale before ChatGPT. This was, in some ways, a struggling company right before. I mean, ChatGPT exploded, right? It was great, but the revenue was, in the grand scheme of things, relatively minor. So, they sold 49% to Microsoft. Now they’re going to de-sell it.

Microsoft does use multiple providers, right? It does compete with ChatGPT and with OpenAI. So, even if it’s allowed under the contract, they did cross the line in terms of remaking, I think, the spirit versus the letter of the relationship.

Guest 2

I don’t know if, just because Microsoft crossed the line first—and I’m not sure they did—I’m not sure that will impact how it ends up. I don’t know how it ends up, but I’m not sure that it will be determined by who “crossed the line first.” I think it’s a very weird contract.

It’s right up there. It’s a very different version of weird than Scale AI, but very weird. It’s not just about ownership and preferred stock and common stock. It’s all sorts of different rights around profit-sharing, all sorts of different triggers around AGI.

I would say, in situations of ambiguity, the person who has the most leverage has the best chance to win. I would argue Microsoft has quite a lot of leverage because it’s not existential for them, especially now.

I think they have no leverage. Actually, honestly, I’m not trying to be a chaos agent like our prior guest. But here’s why. Listen, I’m still a student, right? The reason I think they have no leverage is because I think that whatever AGI is, we’re going to be there pretty soon.

So Microsoft loses. What Microsoft wants out of OpenAI is a relationship where as much of it survives—I mean, obviously it’s tied to profit-sharing—this moment in time in the relationship, right? I’m sure they would argue over it, and I don’t know the details of the contract, but the profit-sharing and all the rest ends at AGI, right? Part of this relationship ends at AGI.

That’s a good reminder of that term because, if you recollect, a few weeks ago in one of our quick questions, someone asked me, “When are we going to hit AGI?” My answer—which now I feel smart about because Jason confirmed it—was: we’re going to hit AGI when Microsoft and OpenAI litigate what AGI is, because everyone else is using that term loosey-goosey.

It could be now. It could be 2045. Who the fuck knows? Now half a trillion dollars of value depends on that word. I can tell you, you’re right, Jason: that is the only part of the contract where you go, “Ooh, that is a bit of an issue because it’s ambiguous.” So, you’re right, that will be a fun vector of this discussion.

9. Chime IPO

Guest 2

Well, I’ve talked—listen, I’m not smart enough to do a lot of stuff in AI. I’ve talked with several of the smartest folks I know in AI, okay? What they think about this, what they think about AGI—and they basically all say very much the same thing—is that, at this point in time, it’s something likely Sam Altman made up to simplify a lot of concepts. He’s the one pushing the narrative, and none of us really believe there’s a magic point for AGI, right?

At a lay level, I think we’re close to being there already. That’s the risk to Microsoft. My only point is, since they both have something to lose in this, and Sam has signaled early—he always does; the guy signals early—and he’s saying, “Microsoft will end up with 33% of a new entity,” he’s just trying to create the outline of a deal so that it lands there, right?

They own a third. He owns 6% or 7%. It gets reorganized in some weird new way, right? What does Microsoft get out of it? They get a longer tail. They’re going to get a longer tail in this agreement, but it’s going to be a different one because the cash alone is not important to them, to your point, right? The IP and the relationship are more important than the cash.

What are they going to do with the cash? Dividend it out? Buy Scale AI Prime? They don’t actually need the cash, do they? I think we can agree that they don’t need the cash.

Yeah, agreed. Scale AI Prime coming to you soon, clearly.

Listen, guys, we’ve talked about Chime in 2 separate episodes, and this went out—it went public and it popped 50%. All IPOs are up in 2025 except for likely SailPoint. Is this a sign that IPO markets are roaring? There hasn’t been a better time in years. And how did you analyze Chime specifically?

So yes, they clearly are roaring now, right? There hasn’t been a better time in years, yes, because there’s been none, and some is better than none. It’s not a complex question.

So, you think it’s pent-up demand that’s just latching on to available supply, not necessarily the quality of companies coming out?

No, I did not say that at all. Look, how do windows open? Traditionally, good companies go out at attractive prices. That’s how you—if you’ve had the window shut, the only way you open it is with good-quality assets and at prices where it’s attractive.

The way it manifests is, unfortunately, you get these pops where effectively the company going public has bribed the buying public to like IPOs again, because investors were all pretty Pavlovian. If you were an IPO buyer in 2022, your searing memory is: “I bought them all in 2021. I lost 30% to 40% on average, 90% in some cases. I’m never doing that again.”

So, you take a long time to come back. Now, in the last month, your narrative has changed. It’s, “Oh my God, I piled into the last 5 IPOs. I’m up 70% on average. I’m up 250% in the case of Chime, in the case of Circle. This has been a huge boost to my fund performance. I’ve got to do me some more.”

The pop helps the win. It’s not ideal. There should be a better way, but there’s clearly positive momentum now from good companies, and there’s going to be more of an appetite.

Will this lead to Databricks or Stripe or one of the big boys going out?

Even if they go out, this sentence is wrong, Harry. It doesn’t lead. You see, “lead”—you’re implying that they couldn’t go out and now they can. They can go whenever they want. They could have gone in 2022. They could go in 2023. They’re so big and so good, they can go whenever they want.

That’s an idiosyncratic decision that they’re going to make that I don’t have a ton of visibility into.

No, no, no. They can go out whenever they want, but now they have data to suggest that if they were to go out at this point in time, it is highly likely that they will be priced at a premium with a huge amount of demand. They did not have that historical data beforehand. So, this data is leading to a new decision for them.

I’ll give you that. It’s not clear to me that the only angst was on pricing. Maybe the better statement is this, Harry: I think it’s much more true for the whole flotilla of companies in that $200 million to $500 million range who have been trying to figure out where to go and can now see some kind of exit.

For those folks, everyone is dusting off their plans. For the big 4 or 5 companies, like Databricks and Stripe, you’re right, at the margin, it’s a more favorable time. Why not?

I think their reasons for doing it or not doing it are more about the whole—when do they want to do it? Do they really want to deal with it? So, it’s not—I could imagine an entire IPO cycle where, once again, neither of the 2 names chose to go out, and we get to the end of the year.

Was Chime mispriced, or is that pent-up demand?

Well, by definition, pent-up demand. The whole point of pricing is to pick up pent-up demand. So, by definition, it was mispriced.

My limited understanding from folks who have been involved in some of these recent IPOs—and it’s limited, Harry; you probably have had more—but the advice was just to be conservative. It’s just that there haven’t been many; it’s a slow market. Will the advice be to be less conservative now after these IPOs? Probably, right? I think you’re still going to tell everybody to be conservative, right? And it was to be conservative.

Yes, we could talk about how much money people really left on the table versus, in theory, not being able to sell every share at the highest possible price, right? That’s the fiction in the circular math. But even if the company leaves money on the table, the investors distributing over the coming years may have only lost a little bit from the incremental dilution, right? Dilution aside, the IPO price is almost irrelevant for the VCs. It’s when you get out, and getting out strong may be worth the incremental dilution, right?

That’s the one that Bill Gurley, I guess, disagrees with, right? There’s a trade-off. Going out strong doesn’t matter, right? You’d rather go out and limp along. But I think that’s the advice. The other thing I would say on Databricks—and I know literally nothing about Databricks—but the fact that this last week they had an analyst summit where they went through all their metrics, their revenue, and their growth rate with an entire group of investors and analysts says to me they will decide one day to file, and it will just happen.

They’re not only running their company like they’re going to IPO; they’re actually managing to Wall Street. They’re already at the edge of managing like they’re already public, right? They’re already halfway there. I wouldn’t be shocked if we opened up the news tomorrow and they filed, because they’re already completely ready. They’re already in; they’re already halfway there.

So, I think we won’t be shocked by Databricks. We think we’ll see these signs, but it’s like a Sam Altman narrative: the signs are already there. They literally could file next week, and no one should be shocked.

Totally agree with you on Databricks there. I actually interviewed the head of sales, Ron, who was unbelievable and spoke in detail about how they structure and run the company. He had been there since almost the beginning.

Yeah, really crazy. I was so impressed with him.

10. Gusto Going Public with $900M in ARR???

One of the companies in the portfolio that you, I think, very wisely described, Rory, is Gusto, which announced a tender at $9.3 billion at $900 million in ARR. I was actually surprised they were at $900 million ARR. Amazing. And congrats to Tom and Josh and the team there; it’s awesome. I kind of underestimated them, and I thought that was awesome to see. I don’t know if you guys had a take or a read on that coming out at $9.3 billion.

I mean, yes, you underestimated them. We had talked to them way back when, and obviously you should have done the deal, so I share your pain. Look, payroll is one of the biggest markets out there. Sometimes the obvious [__] is the most important [__], right? It’s a huge market because everyone gets paid, and you get $5 a month per U.S. worker that you’re paying.

ADP and Paychex are 2 old-school companies that are worth approximately $100 billion. There are a bunch of other companies at the $10 billion-to-$30 billion level of public comps in the space. So, it’s a huge-ass market with a lot of attach, and it was pretty stodgy.

At one level, I’m not surprised, given the TAM. They’ve clearly executed really well to get to that size and scale, but they have a modern SMB-focused payroll solution. Half of the workers in the United States work in SMBs, so you’re probably at 60 million people on which you can get—I’m just doing the math in my head—$5 a month, $60 bucks a year. Hold on. Yeah, I mean, you add it up and get to a huge market very quickly: a $6 billion market.

They’re also lucky. I think all these guys are lucky in that they have a great comp. Paychex, again, I think, was founded in the 1870s. Rory, we could look it up. Paychex is an old one. It’s trading at 10 times revenue, at $55 billion in market cap. It’s not just trading at an all-time high; it’s at an all-time high today.

So, if you’re a VC, it’s very easy to say, “Well, look, okay, Gusto’s at $1 billion. Paychex is at $5 billion. Okay, Gusto’s growing a little bit faster. It’s newer. It’s better. It’s founder-led.” You can back into Gusto. All of the payroll folks—you can, with a $55 billion ancient public company trading at 10 times revenue, justify any of these deals, right?

Some people don’t realize that Paychex, Toast, and a few others have some great comps out there.

Well, ADP as well. I mean, you have ADP. The 2 big dogs in payroll are ADP and Paychex. I know Paychex, I think, was founded in the 1960s or ’70s, because the original founder continued to run for governor of New York as a Republican. All very weird.

ADP is almost at an all-time high too. They’re 2 big businesses.

So, provided you can make the economics of replacement work, you can build a big company.

Actually, I didn’t realize ADP is trading at 15 times revenue. Why didn’t we just put the fund into ADP, Rory? Forget about putting the fund into these newer kids. We should have put the fund into ADP.

Look, if we’re going to go with ADP and Paychex—I mean, that’s what my grandpa’s startup used. They used Paychex, right?

Rory, what was your reflection on turning down an early Gusto round? Did it change your mindset or teach you anything?

Yeah, that I can occasionally be very stupid, which probably didn’t need to be thought. We talked in between rounds, so it wasn’t actually an engagement, but I think I underestimated both the founder and the openness of the market to switch. I’ll say something else: the willingness of investors to pay up quickly for that, such that it got outside my price range very quickly.

Stuff happens, you know, and you were wrong. I mean, I was wrong on that decision. Being wrong sucks.

It’s funny. They’re still niche products, actually. I’ve been a Gusto customer for a decade, I think. I’m a case study on their website. I actually think it’s some of the best software out there. I can tell you why, but I have to do more work than I used to, because when I was a Paychex customer, I just called up my rep: “Hey, Harry, process the payroll. Hey, Harry, take Rory off the insurance.” I didn’t have to do anything. My human did it for their commission.

Now I have to log in. I have to run the workflows. It’s elegant and it’s great, but I actually have to do more work with Gusto than I had to do with Paychex. More work. I don’t care about the money savings; it’s irrelevant on my payroll, right? Which makes me conclude that somewhere out there, someone is building an AI-enabled, agentic payroll where Jason can just talk to the freaking chatbot and it will do all the work.

They have some of it. I’m sure they will. It needs a little bit more time in the oven, but yes.

11. Dropbox vs Glean: Can the Old Guard Survive the AI Wave?

Final one before we do a quick-fire. We mentioned the old guard competing with the new. Glean is a $7.2 billion company today, moving very fast. Dropbox is still working to get Dash out, which I’m sure is a good product, but it’s separate from their core product and core business. Can the old guard compete with a very, very fast, well-funded new guard?

I’m starting to lose confidence in the old guard. I wanted to believe. I think because the LLMs are open, anyone can use the API, and most of the work is done by third parties. Most of it is not. Yes, you can build, you can use an open-source framework for your chat, and then you can use an LLM, and then you can RAG a bunch of data and stick it in. It’s pretty simple.

I just figured that, at the software level, the big guys would catch up, right? But in my ecosystem, they’re still too slow. I just don’t see it. I don’t see it. I’m worried.

First, the minor comment: I do believe the so-called old guard can compete in certain circumstances, if the situation is right and they have leadership to just push it through. But I don’t think that’s the issue here. It’s a different issue.

The real truth about Dropbox shipping or not shipping whatever the product is—Dash—is that it doesn’t matter. Let me tell you what I mean by that. Typically, the old-guard play is some version of, “I’m the system of record. I’m already in there. I can add this stuff on top.” That’s the pitch every time: “I’m ServiceNow; I can add an agent. I’m Zendesk; I can add an agent. I’m Salesforce; I can add an agent,” right?

The new guy’s pitch is, “We have this agent, and we can run on any system of record.” So, you can buy us, and you don’t have to change out your old guy, right? That’s the vector of competition. What it means is you can only be an old-guard system-of-record provider. Your right to win is predominantly in your existing customer base, and ServiceNow has a big one and Salesforce has a big one. Brutally, in the enterprise, Dropbox doesn’t.

Even though we threw out Dropbox and Glean, they’re not in the same quadrant. It’s not a relevant question. With all due respect, Harry, Glean is out there competing with all the people on enterprise-wide deployments for big-ass companies, and they’re going to have a whole bunch of competitors in that space that we could talk about.

Hear me out: even if Dropbox had pitched a perfect, functionally equivalent product to Glean, it would mean that, for their existing business customers—which is a smallish percentage of their total business—they would win all that business. But then, when they moved on to customers who didn't have Dropbox, they would be at ground zero competing against Glean for that.

And that, I think, is the real issue. It's not some kind of, “Are they powerful? Are they good enough engineers? Are they strong enough founders?” I think, situationally, starting where they are, it's just hard to get there.

Rory, the joy of me is that I have very few feelings, so you can always say it's a wrong or moronic question. My retort to you would be, “You're very wise.” What would be the right question to ask with regard to the new versus the old guard?

Whenever you're dealing with the old guard, by definition, they're not going to bring oomph to the table. They're going to bring the assets of what they have. So the interesting questions are things like Salesforce, ServiceNow, and people like that.

We had a mention there of—I will say—the whole Salesforce throttling Slack seems pathetically lame and irritating to customers. I was asking someone about this, and they had a good take on it. They said it probably doesn't survive in that form, but it probably survives because, if I'm a customer and you're telling me I can't access my Salesforce records when I want to run them through Glean, that's not going to—I'm sure that's not going to survive.

12. Is Slack Dead as a Platform? Salesforce Shutdown Slack API?

But the question is: is there some kind of API connectivity fee? Are they going to try and monetize that? It definitely felt a little lame. I mean, you'd really want them to say, “Of course, you shouldn't use Glean because our product is so much better.”

Has Slack, bluntly, had its best days, and is it now in decline?

I mean, almost by definition, your best days are as a startup, going public, and then selling for 27.7 times revenues. I don't think it would get 27 times revenues today. So, yes, by definition, it's on the downward slope of excitement, which isn't to say that, if it was well-run and integrated well, it couldn't be a perfectly good and successful part of the Salesforce ecosystem.

We're small in size but big Salesforce users, so it's not like it's doomed. But they've got to make it happen.

Guest 2

Do you think it will be a good and successful part of the ecosystem if it continues as it is?

I don't have a developed opinion, Jason.

Guest 2

Well, look, first of all, it's probably doing $2.5 billion in revenue today, somewhere around that. Just as an aside, I don't think Salesforce got a terrible deal in the end. You just have to view it by where it ends up as a revenue multiple and remember that Salesforce is growing single digits. So I don't think Mark would want to give up that $2.5 billion today, right?

In fact, interestingly for Salesforce, not all of them, but so many of the acquisitions—Slack, even Tableau, which I think was founded in the 2000s—they're actually all growing faster than the core Salesforce. So we can say, “Oh, Slack isn't what it was,” which it isn't, but growing in the teens when the core is growing 7–8%, and you have billions in revenue—that's material, right? This is hard stuff to do.

So I think Slack today is pretty successful. LinkedIn is not as good, but Teams and Microsoft—it's just a different Slack.

Will it be a hub for developers that people build their ecosystems around? Those days of Slack as our hub are far behind us, right?

Guest 2

But here's the irony of it: my view is that Slack got acquired. It was a better—an even better—deal than it looked. You know why? It only had 1 product.

Stewart is a generational founder, right? He did it multiple times. He's the kind of CEO we'd all love to work for, right? But I think deep down in his heart, he didn't want to do multiproduct, or they would have been multiproduct. It had an existential ticking time bomb because it could only be so big as a single-product company. So it had to evolve.

This is a new enterprise Slack. It is what it is. We still use it, right? Just like Workday and LinkedIn are locked down, it can only be locked down. The 3 of us can complain, but if Allstate and State Farm and those folks don't complain, it ain't going to change.

We don't matter. The small Slack and Salesforce customers—we don't matter. At $30 billion ARR, we just don't matter, right? We don't matter.

Will Apple announce iPhone assembly operations in the US this year? I'm sitting outside the US, and I hear constantly, “Oh, we want to bring back manufacturing, starting with phones.” Will they bring it back? Based on the news today, I don't know what the odds are. I don't have them in front of me. I'm going to say they're going to announce it. Yes.

Guest 2

Because Trump announced the Trump phone at $499 to be built in the US yesterday.

It's gold. It's gold. You guys are so classy. Your president's announcing phone packages. Look at you. I like all gold. You haven't seen our new office; it's all gold as well.

Prime, that's because your prime minister couldn't build one.

Look at you, Rory. You've got a coin and a gold phone.

Not nice. I think, listen, I would have to do more analysis to see if it's not possible, right? If it's close to impossible, but announcing it and beginning a process, you might have to do it. You just might have to do it. So I'm going to say yes—announcing, yes.

Rory, announcing, doing, yes or no?

I think politically it would be smart to announce it even though it's impossible. I think Tim Cook is such a straight shooter that he just couldn't pull it off with a straight face, because everyone would know it's absolutely bullshit.

But, yes, in a cynical world, I would make 1 phone in the United States of America just to confirm that we, too, can do the low-margin, commoditized parts of the electronics value chain and end this question.

Will the S&P finish positive this year, yes or no?

You know, if I knew that, Harry, I wouldn't be talking to you. There's an easier way to make money than venture capital if you know that. So it's just not knowable, right?

That's why it's the bet, Rory.

That's a different question. I made the bet. No, actually, that's exactly where I was going to go. The record of analysts predicting the S&P at the start of the year—this exercise where analysts predict the S&P at the end of the year—is universally wrong. The error rate is huge. There's nothing to it.

A much more interesting question is, okay, now let's talk a bet, because then you have odds and now you can actually make an economic decision. The odds here are that the S&P will finish positive—I'm reading it as 70% likely. Is that correct? So that's what they're saying. Given that, I would take the no.

Guest 2

You would make more money betting no than yes at that kind of ratio, which is not to say I have any insight on the S&P. I merely look at the volatility and say, “On average—well, not on average, in this case—not a bad bet.” But it's a very ill-informed comment.

I'm all in. I have no more cash, so I'm all in on this bet already. I think it's an easy bet. But Rory's right about the 70%. I'm not—when it's my own money, I don't get the benefit of the 70%, right? I get the benefit of the 50%. But personally, I'm 100% in, so I already made the bet.

Yeah, you're right. That's actually a good point, because often I want to build on that. I think when people ask, especially investors, their opinion on bullshit, the correct answer is: don't even answer that question. The real question is, where is your position?

And I'm with Jason. I'm about 75% equity—mildly terrified—almost nothing in bonds, and then invested in short-term and weird stuff.

Guest 2

No, I'm back where I was in 2008: 100% equity in my personal position. 100%.

Back just like 2008. It was great.

Guest 2

What was the white part of 2008? It was great.

Okay, final one. Will a Chinese AI model reach number 1 this year? The odds are pretty low. Obviously, they do evals on effectiveness and performance. Will a Chinese model reach number 1 in evals this year?

Guest 2

My bet is yes, but not published. So, no.

Easy call. In other words, I'm right, but no one will ever know.

I don't have anything to add here. What are the odds?

Why?

It's very low. Again, you look at that and go, it's not a crazy bet. Do you think there's a 1-in-5 chance that they'll?

Guest 2

I do. I'm going to try to get back to China this year. I just—I don't know. I should know. Obviously, we've all been shocked the last whatever 8 months by performance, right? And I don't mean to date myself, but, man, outside of pockets of San Francisco and areas of London, I've never seen folks work remotely like they do in China. I mean, it's just insane.

Maybe it's changed, because I haven't made it back in a while, but I plan to go back this year.

What's that?

Guest 2

I don't think they're working any less. If the government's behind you in China, and there's so much energy in that culture, I don't want to take the under on this bet, because I feel like this is hubris. This is Bay Area hubris. There's a lot of Bay Area hubris, and this is one of them.

We can't access theirs; they can't access our internet. It's a different world over there.

But, man, the power when everyone's aligned. Jason, I'm coming around to you. The bet has to be yes. There's all sorts of noise about how it's moderated, but the idea that there's a less than 1-in-5 chance—which is what the odds are saying—that at some point this year, 1 of 5 or 6 really aggressive, well-run Chinese tech companies can't achieve even temporary parity with one of the incumbents just feels to me like a stupid bet.

So you're exactly right. The bet at these odds has to be a yes, which is different from saying they're, quote unquote, going to displace OpenAI. Probably not. Just like at various times when one of the other models sneaks ahead of OpenAI for a month in the various measures, you kind of go, "That's interesting as a state of the art," but that's not actually how the business value is created at that level.

I come around to—I'm with Jason. The answer is yes. There's no world where we've got all the smart people and they don't. It's silly the minute you say it, so of course they will go.

Well done, Jason. We got that. That was a very definitive answer, Rory.

Yeah, Jason convinced me. He's exactly right.

No, I love it, guys. I hadn't spent a second thinking about it beforehand. Well, now you're a true venture investor, Rory. You should be proud. You say things with little thought but great confidence.

See, that's why I was afraid of what would happen if I did this.

Have you said anything more likely to cause me to fade? Don't worry. Did you have such a visceral reaction to Jason saying that about Tableau in the '50s because you turned that one down at C?

Oh, no, no, no. I didn't, but it was—I want to say the early 2000s. I remember 2003.

2003. Yeah, exactly. And I'm sure the poor Tableau guy is screaming, like, "Please don't say that about me. I'm relevant. I got 40 million. I am somebody." Oh dear. Well, guys, listen. I always love my time with you. Thank you so much for doing this with me. You've been fantastic. Rock and roll.