Coinbase 将 AI 支出砍半｜Kalshi 400亿美元估值与IPO在即｜SaaS并购整合之年

Boys, it is the holiday edition, baby, coming to you from the British Riviera. I always say this is like the UK's Saint-Tropez, and then people look up Frinton-on-Sea and they're like, “Huh.” Americans don't get the British sarcasm. It's my last name.

Jason Lemkin

Yeah, I get the joke, Harry. I—

1. Coinbase Cuts AI Spending

We have a lot of news today. I wanted to start on what I think is probably one of the biggest topics of the day, which is more a macro or meta topic, but it was brought to light by Brian Armstrong and Coinbase. Coinbase's spend is down 50% this quarter, but usage is up with regard to how they're utilizing open source. That's taking away from their frontier model usage. How did we read this? Is this the new normal? Is this a frontier company with a frontier founder doing frontier things in terms of switching so efficiently? How do you read this?

Jason Lemkin

I have a lot of thoughts, and this got so much traction, right? I think it's an important topic, but—

I agree.

Jason Lemkin

I'm just getting burned out on performative social media from struggling CEOs that aren't in AI companies trying to act as if they're at the bleeding edge. I really don't care what a lot of these CEOs think about the performance of the latest LLM on their boring, old, pre-AI SaaS or crypto company. But, man, enough of the crap from Brian Chesky and Brian Armstrong. I mean, these are generational founders, to use Harry's words, but it's just performative. Put the numbers up, boys. Sorry we're in a crypto winter. Sorry Airbnb is still below its IPO price. Go do something about it. Go fucking do something about it. I've had enough.

Rory O'Driscoll

I'm in a totally different place because I actually thought it was a really great piece. It's precisely because, in the continuum from frontier, amazing AI company to boring, stodgy corporate America, Coinbase is now more or less in the middle. This is your point. It's not a 4- or 5-year-old company that's been public. It's not a bright, shining frontier AI company, which is what makes the piece so good.

To me, it was like a common-or-garden tech CEO saying, “I was spending X a year ago. It exploded in the last 5 or 6 months, primarily because of the ability to do code generation. I got to grips with this thing, and I reduced my spend by 50% in the last 2 months. Here's how I did it.”

It's precisely because it wasn't some frontier-leading company that it makes it more relevant, because I think every single company spending $50 million or $10 million on Claude is going to look at this and say, “If we haven't done all these 3 things in the next month, someone's head is going to roll.” It's cost management 101, but what it said is, in the space of 2 months, by getting to grips with your spend, you can continue to innovate, continue to generate more tokens, but cut your spend by half. So I thought it was a really great piece.

Jason Lemkin

It is. It's just Coinbase fell minus 30% in its last quarter.

Rory O'Driscoll

Yeah, but if those things are on par with—

Jason Lemkin

No, I'm not disagreeing with you. I'm just—

Rory O'Driscoll

Sure. Ah, no. Those things are on par with—

Jason Lemkin

And I think the graph and the chart were great. I'm just burned out. I want a leader doing this.

Rory O'Driscoll

What about an even shittier company? We have lots of shitty companies. If one of the automotive manufacturing companies, which are facing fundamental business challenges, also did the same thing and said, “We were spending $100 million on AI, and now we're spending $50 million,” it would be just as relevant.

In fact, I would argue it's precisely those companies that are under cash pressure and earnings pressure that are going to get their shit together on not spending too much money on AI, which is why it's more important. Frankly, that's most companies. Very few companies have the luxury of a venture-backed, AI-forward startup that can say, “Just get this done no matter what.” Most companies have cost discipline. And I—

Jason Lemkin

Is optimizing your LLM spend really going to help if your revenue is shrinking by 20% to 30%? I just don't see how it's going to reignite growth in your crypto user base.

Rory O'Driscoll

But he didn't say it's going to— Again, you're just being a grump, guy. He didn't say it was going to reignite growth.

Jason Lemkin

No, I think it's a sign. I think the data is valuable. Let's move on, right? I think it's great, but I am getting burned out on struggling CEOs on Twitter sharing performative AI data when they're not AI companies. Show me the money.

Rory O'Driscoll

I don't think it was performative. I'm going to argue again: I think it was a fact-based piece. I actually really liked it because it was a fact-based piece. By the way, it just showed, “Here's our spend by month, here's our tokens generated by month, here's our peak, and here's how we were able to clip 50% off it.”

Let me put it this way. Rather than arguing with you, Jason, which is usually not productive for either of us, I'm willing to bet that every single CFO in the Fortune 500 sent some version of that article to their CIO and said, “Dude, look what this smart guy in the Valley is doing. Figure your shit out.”

So I think, actually, let's leave Coinbase out of it. Let's pretend we don't even know the name of the person who wrote that article, right? Let's ignore it. I think the really interesting question is, what does this mean for revenue traction for the hot, sexy foundation models?

I was looking at the numbers. In one sense, oh, my God, you reduced your spend by 50%. That's terrifying if you're the company getting that revenue. On the other hand, the positive spin is that it only went back to the spend they were doing in November. In other words, basically all that happened here was all these companies were spending and growing pretty aggressively in terms of their spend with, probably, Anthropic and OpenAI, and then in November and December, with coding, it exploded.

It took about 5 months for everyone to get their shit together and say, “We can't be doing this.”

Rory O'Driscoll

“Let’s cut the burn,” and then they figured it out and reduced it by 50% back to roughly the spend in November. Now, the interesting question is: Does that imply that Anthropic, which last year exploded from a $1 billion run rate at the start to $9 billion at the end, to $44 billion mid this year—I’m not saying this, but you could say, “Oh, are you saying that their revenue’s going to go down by 50%, so they’re going to be at a $22 billion run rate?” I’m not.

But it’s going to have some impact on the growth rate.

Jason Lemkin

I think so.

Rory O'Driscoll

Yeah.

Jason Lemkin

I think there are 2 issues here. One is that it sharpens the question of whether the rise of open source and others is actually going to impact the growth of the frontier models more than we ever predicted. That’s a big question, and I’m going to say I don’t know on June 30. It’s easy to say that. We could point to a lot of data. I think maybe that’s the topic here.

I think there’s a second point in the post, where Coinbase was really helpful. It’s the less dramatic version of what I was saying. I do think there’s a second point that the post made that people maybe missed, which is that as we round into the second half of ’26, folks are realizing they radically ramped up their AI spend on product. It seems to have worked subjectively, qualitatively, but the productivity isn’t there to justify it.

I think that’s what Brian was really saying. The data didn’t say it, but he’s like, “Listen, if we had shipped so many new products to Coinbase, if our product velocity had quintupled because our token spend quintupled, I’m all in.” If that flipped him around from minus 30% growth to plus 30% growth, I don’t think we’d be arguing. He might have still done what he’s doing, token-routing a model. He wouldn’t be making this point.

I think if you look across many of our portfolio companies that are doing well and are not purely reselling tokens, they’re coming to the conclusion that, “I’m not quite sure what the hell to do. I know I want us to do this. I can’t put the genie back in the bottle. AI is great, but it’s not lines of code. What the hell?” So many folks are not seeing the lift in net revenue and net productivity they thought they would get from agentic coding.

It’s almost a conflict, but it’s something we’re all going to have to deal with in the second half. But it’s not the same as cutting costs. It’s saying, “Jesus, I spent an extra $10 million in the first half of the year, and we grew the same as we did the prior 2 quarters. Where’s the lift, boys? Show me, show me the lift.” CFOs are struggling with that, too. Even if the business is doing well, they’re struggling with it now.

Rory O'Driscoll

Yeah, depending on the company, this AI spend on engineering, if you’re a software company, it should credibly give you a revenue lift, because you’re making more software. You should get more revenue lift, or if you’re a digital goods company. So you’re right, Jason. If you’re Coinbase, you’re like, “I’d like to have seen revenue lift here,” right?

I think it’s even applicable for companies the further you are from a digital good. Silly example: If you have a tech team and you’re a car manufacturer, going back to it again, you’re not going to get a whole ton of lift from your extra software unless you’re Tesla with FSD. But at a minimum, you should be seeing savings. If you were spending $100 million on software and now you’re spending $10 million on tokens, at a minimum you should be seeing savings.

And if you’re not seeing either, you’re going to be looking at this with a pretty jaundiced eye. I think that’s what’s happening.

Jason Lemkin

I have a portfolio company that is a software company. Every number is green, right? Way overloaded with investors, way overloaded with everything, hit the first half of the plan. Everything’s great. You would love all the numbers.

But at the last board meeting, they came in and wanted to double their token spend, which was massive in the first half of the year. It was enough to move the burn from no big deal to—

Rory O'Driscoll

Big deal.

Jason Lemkin

—even for a company in the top half percent, even folks were like, “That’s a lot of extra budget.” And the feedback, for the first time, the board was like, “Okay, but if you want our approval, tie it to ROI,” and this amazing team couldn’t. Everyone wants to invest, but it didn’t directly tilt the curve.

So there is a point for even the highest flyers where you’re going to say, “Jesus, I’ve got to see the ROI.” And I think that’s the big issue. We went into token maxing, where everyone just tries stuff. That made sense, right? And that led to the early folks that got whiplash, like Cursor, having to go open source really early, and that’s an interesting niche issue on X.

But the real issue is that we can’t show enough lift from this spend, and it’s going to stress even the best of us, not just Coinbase. It’s going to stress everybody, and so be it. It’s time for the next mature phase of token spending and software development. It’s just time, boys, to grow up, right?

If you’re an Anthropic shareholder, though, and you see Dario say, “Hey, we need $1 trillion in revenue, or close to, for this business to be viable, or we will be bankrupt,” maybe he says it kind of superciliously or glibly, but he says it, and then you see the dominance of open source now pervading a lot of usage, you have to be concerned that it will cannibalize that pathway to $1 trillion in revenue.

Rory O'Driscoll

You have at least some concerns, which will segue to our next discussion on distillation and Anthropic’s perspective on these open-source companies stealing their IP, as they would say. We will leave the irony for a later discussion.

But yeah, it is plausible that you have a world where, even if the bulk of the tokens are generated using open-source models, the bulk of the revenue will still obviously come from state-of-the-art frontier models, right? And therefore, there’s clearly a very big business here, right? That’s all great.

It’s to your point, Harry: If you’ve constructed your world in such a way that only $1 trillion is good enough, and you end up with the consolation prize of $500 billion, which is still either the 2nd- or 3rd-largest company by revenue on the planet, and it would be the largest digital company on the planet, if you can be the largest tech company on the planet and still not make money, you might have oversized your ambitions a little, and it might pay to come back a bit.

And that’s exactly right: Nothing in this Coinbase memo or 100 Coinbase memos implies anything like, “Oh my God, these are not going to be amazing companies with great products that have differentiation.” It’s just as you say correctly: If you’ve built a cost structure and a CapEx spend that you need it all, then the last thing you need are cheap open-source alternatives at 1/5 the price.

2. Anthropic Faces Distillation Claims

Before we move on to Anthropic’s perspective on distillation, Jason, I love you, my friend, but what do you want from these CEOs, then? Candidly, he’s being very factual and innovative in how he’s presenting what the company’s doing. What do you want from him? To just shut up and do the work in front of you?

Jason Lemkin

Listen, I value the data, so I appreciate that. I’m not being facetious. I do feel this way. I want to see how AI, if at all, can give Coinbase a revenue lift. That’s what I’d like to see.

Coinbase is subject to the whims of the crypto market, okay? And investors should understand that. When crypto roars back, Coinbase has grown at rates that are almost Anthropic levels for brief periods of time, right? So it’s part of being on a non-recurring revenue journey in a very volatile market.

But I’d love to see how growth is 5% higher from AI doing something in crypto. I’d love to see how it’s driving up insurance premiums and insurance margins. I just want to see where this magical boost is from this utility.

The LLMs are a utility, right? They’re tokens. They’re not fungible utilities. We’re kind of teasing at whether they’re becoming fungible utilities, right? Is one token replaceable for another? That’s the meta issue.

I’m tired of folks like Adobe saying, “We have $500 million of agentic revenue,” and missing the quarter. That’s performative, too. And listen, what would I be doing if I was the CEO of a company not accelerating in the age of AI? Man, I might be doing the same, but I don’t respect it.

But I want to see the real boost. Everyone’s faking, everyone’s feeling like they have to be part of the AI age, but they’re not delivering, Harry. They’re not delivering.

Rory O'Driscoll

I want to applaud Jason for his consistency and disagree slightly. Firstly, I give you credit—you have been remarkably consistent on this, and I’m going to paraphrase what you’re saying. Software companies in the age of AI are either accelerating or irrelevant, and you’re exactly right.

I don’t think crypto gets an automatic lift from AI, so I think you’re being a bit harsh on that company, but your Adobe example’s exactly correct. Oh, and it’ll come—actually, it’ll segue to the Microsoft discussion later on.

Jason Lemkin

I do agree, and again, I give you credit for strong principles, strongly held. If you're not accelerating in the age of AI and you're a software company, you've got a problem. That's your point, and I think you're correct. So, if that tweet—let's put it this way—if that tweet had come from the CEO of Adobe, you would be totally correct in saying, “That's great, but dude, you need more.” Exactly.

I do agree with you. I would exempt Coinbase from that because I lump them more in the financial than the software space, but in the software space, you are correct. If you're not getting on board this train, you're getting left behind.

Rory O'Driscoll

Even sometimes, and I love him—we all love him—for a while I would get tired of Aaron Levie's constant AI stuff. But to answer your question, it has led to a boost at Box. It hasn't turned Box into a 100% grower. It's gotten it back to double-digit growth.

Some of Aaron's stuff is a little bit too much for me, but we all learn from it, right? Some of it is a little bit too AI-reflective, but he ties it to his revenue and his plan. He's like, “This is how it worked at Box. We're processing documents. We're processing content. Here's how it literally ties to our business model.”

So it's okay if some of it is a little bit performative. I give Aaron a huge thumbs-up on it. Anyhow, it's all good. It's great to have the data. I've just become a skeptic when it's not tied to their business. That's all. I'm a skeptic of what the goal is here. That's all.

You know, Rory, Jason is not only consistent, but he's also prescient in being ahead of the times, because he's also cited his—I don't know if I can say disliking, but his favoring now Sam and OpenAI over Anthropic, and his kind of boredom with Dario saying that we're all going to lose our jobs.

It seems the world doesn't like Dario right now, and Dario has continued to whine. Sorry, that's unfair, and Rory did not say that. Dario has criticized Chinese models for stealing—for brazen theft of their work through distillation of their models. How do we respond and think about Anthropic's commentary on whether or not Chinese models are stealing their work?

Jason Lemkin

Sure. You have to power through the irony before you can have the discussion, and we all understand the irony, which is that all the foundation models, including Anthropic, were trained on other people's IP, to the point where Anthropic recently settled litigation with a whole bunch of copyright holders over books because they had, quote-unquote, “unfairly”—to use Dario's word applied to the Chinese—leveraged their intellectual property.

I do admire the element of hypocrisy: being appalled when someone else does it to you, or having done it to other people yourselves. Be that as it may, let's move on from that. Just wallow in it for a second, but then move on.

What's happening, again, stepping back for folks, is the allegation—which Anthropic, the company, made in a letter to the Senate Banking Committee recently—is that, basically, the Chinese open-source companies are bootstrapping their development of their state-of-the-art models by effectively breaching the terms of service of Anthropic and sending literally millions of prompts to Anthropic, recording the answers, and using that as training data to start training their models.

It's basically taking—and to some extent they would say it's taking—Anthropic's IP and using it to build open-source models, which then compete against them. So that's the comment here, right? Other than “so what?” and “Ha, that's funny,” what happens after that?

Well, the first thing—the main thing—is: Is it illegal? The interesting thing here is that it's clearly in breach of Anthropic's terms of service, and that's very clearly expressed. But that's not a legal case; that's a contractual problem between Anthropic and the Chinese model companies, and they're more than welcome to sue each other in—you know, knock yourself out in Beijing. Good luck with the lawsuit, dude.

I think the interesting thing, and that's why the Senate Banking Committee thing is interesting, is you could imagine a world where, because as well as just being in breach of terms of service, it is also arguable that you have copyright issues and/or trade secrets acts. They go above the level of contractual issues and then start to get into actual legal issues that the government might take an interest in.

Or maybe, at its most extreme, I could see Anthropic saying to the government, “Hey, these are strategic U.S. assets. We're getting regulated separately on how we produce our products. You've got to stick up for us and say you're not going to let this happen, and put the full weight of the U.S. government behind it.” And that takes what was a contractual dispute between 2 parties and makes it the U.S. government putting the thumb on the scales. Clearly, that's what they're angling for.

Going back to the Coinbase comment, what happens if, in return for Mythos kind of complying with U.S. government restrictions on overseas access, the quid pro quo in the next piece of legislation is that no Chinese model that has been proven in a U.S. court of law to distill using U.S. foundation-model technology can be used by a U.S. company? It's not crazy. You can see where they want to go.

They're not just crying because it's unfair. In my view, they're laying the pipe for a set of trades to push back on this open-source stuff if they're, in fact, doing distillation. So that's, I think, what's going on here.

To Harry's point, whether he's no longer a successful communicator at this point, whether the spite-startup vibe of the safety guys has expired and led to conflict with Trump, putting that aside, I think he wants the models banned for use by U.S. companies.

I think he wants Chinese models banned for use by U.S. companies, and I think in an era where SBF may get pardoned and the founder of Binance already got pardoned, this is pretty imaginable. I think it might be perfectly logical for them to get around the table, especially when we're jostling on a geopolitical level, and say, “Listen, we're just going to ban it,” or, “We're going to do some weird tariff on any U.S. startup that uses a Chinese model. They have to pay a 100% tariff tax.”

That's beyond my skill set, but clearly they just want U.S. companies to stop using them. You can't stop China from doing what China's doing. I just got back, and you ain't going to stop China from doing it for a million reasons, but just put an end to what Cursor and Harvey are doing. No more. You guys can't do it anymore.

You guys have to say, on grounds of national security, “This is theft of our IP. This is theft of our data. We can't trust them.” And hey, Cursor and Harvey, your models are just destroyed—your business models. Thank God you sold for $60 billion, because Chinese open source is banned.

I don't think it's implausible. Is it on Kalshi? It seems to be more than a 10% chance it's going to happen.

Would you not say it's actually almost inevitable, not plausible? When you look at both Sam and Dario advocating for it, and the people around the administration advocating for it too, you don't exactly have an opposing side.

Rory O'Driscoll

Well, you should have, because this is—and Bill Gurley's great on this—regulatory capture in the extreme. The truth is, there are 2 separate issues that they're brilliantly conflating.

One is: Should they be, quote-unquote, “banned” because they distilled our prompts and, as such, got a leg up on that? Naughty them if they did, just like naughty Anthropic and naughty OpenAI for stealing your 1.5 million books to leverage your property 3 years ago. So pay the naughty fine and move on. No one's banning OpenAI and Anthropic because they were naughty, so logically the distillation thing shouldn't result in banning.

The separate thing, and they're conflating the 2, is whether you're really trying to find a motivation from the government that says, “Ban—not because of the naughtiness, but because Chinese open-source AI running in the U.S. sounds scary.” I can imagine that. There are products that are in the interest of national security, like Huawei, that are banned in the U.S. So it's not crazy that, if we believe frontier models are equivalent, you could see that kind of thing.

Now, the argument that every tech CEO will and should be making is that these are open-source models that are sourced and running in the U.S. on U.S. infrastructure. There's literally nothing going on. The code is open to inspection. There's no backdoor here. There's zero risk involved.

But to make that argument implies that you have, on the other side of the table, a government entity willing to listen and do nuance, and that's hard. So I think that you could imagine, in the absence of a sensible regulatory function, that you conflate the naughty tax for stealing the intellectual property, the security risk because it's China, and then the deep, dark secret is the frontier models are actually just trying to defend their vast CapEx spend by eliminating a low-cost competitor.

And it all comes together in a big kind of policy mismatch in return for some of these restrictions on security and/or usage.

Jason Lemkin

There's also a middle ground, or a partial win for Anthropic and OpenAI: just, listen, we can't stop Cursor, which they could. We can't stop Cursor and Harvey, but you know what we can do? We can make every single Fortune 500 company uncomfortable using open source.

They’re not comfortable from a security perspective. There’s enough ambiguity out there that it’s just not worth it, right? Startups can cut corners that enterprises are not comfortable cutting. All you have to do is make it look dangerous to enterprises, and they can just ban any open-source use in their company.

Rory O'Driscoll

And again, to be clear, I have to distinguish something there. They wouldn’t say, “Ban open source.” They would say, “Ban non-U.S., China-based companies distilled on U.S. models.” They’re not going to ban Reflection AI or Poolside. At some point, there’s going to be an interesting opportunity for a U.S. open-source company like Reflection AI or Poolside to take some of this revenue.

But you’re right, Jason. Some version of banning these companies that have been found, quote-unquote, “guilty of distillation,” and that are based outside the U.S., in China, is plausible. Stepping back, one thing I noticed here is that Anthropic wrote to the Senate Banking Committee. Obviously, the minority head of the Senate Banking Committee is Elizabeth Warren.

You really have to want to get something done when, as a tech company, you say, “What I’m going to do is pull Elizabeth Warren onto my side.” Very few tech companies say, “That’s what I really need here today.” It just shows that you’re trying to get something interesting done here.

3. Chinese Models Face A Ban

Guys, personally—and you don’t have to answer—but I’m just intrigued: do you think we should ban them?

Rory O'Driscoll

No, because I think you have to be very pragmatic. And again, I’m willing to be corrected, but there are 2 separate issues. If they’ve done the naughty, then they should pay the naughty tax. They should pay exactly as much as Anthropic had to pay to the book guys, and they should have to pay it to Anthropic. It’s easily priced, and that would make things, quote-unquote, “fair.”

So from a distillation perspective, no. Separately, from a national security perspective, as I understand it, if you are downloading the model, the code is open for inspection, the weights are available, and there’s nothing running and no telemetry back to China, then I don’t think there’s any danger. So I think you probably shouldn’t ban them. That would be my take.

Jason Lemkin

I think AI is going to be like the oil situation in the Persian Gulf today. As an economy in the U.S., we are so addicted to this: our 401(k)s, our stock market, everything we do, every bet, every re-election for every politician. We are so tied to the AI boom.

I’m not going to debate whether it’s a bubble, right? At some level, it has to be a bubble. We’re all on this journey together. 40% of the S&P 500 is tied to this bubble. I think we have to protect it.

Sarah Friar, or whatever, a year ago saying that the U.S. government should backstop all of our data centers—I think it was her who was flamed for saying that, right? It was probably true, and she probably shouldn’t have said it the way she said it. I think it was an offhand comment.

Everything is going to be circular, because we can say we don’t need that oil because we make plenty of it from fracking, but look at the impacts, right? Even if gas prices go up a dollar a gallon, people are freaking out. If their 401(k)s go down 20%, if Nvidia stops selling chips, I don’t know. I think we’re going to end up doing everything we can as a society to prop this up.

I think we’re all going to be, and it may even be more important than the other issues. We just don’t want unemployment. We don’t want these things to happen, and we’re all in on this, whether we realize it or not. We’re all tied to the AI economy.

It’s all about me. I care about unemployment, I care about everything, but my 401(k)? Don’t touch it, man. Don’t touch it. I feel a little bad about the price of gas, but don’t raise the price of my gas. Don’t touch my 401(k).

Rory O'Driscoll

What you’re saying might be true, Jason. I’m not sure it is, but it might be true. But if it is true, what’s implicit in that statement is really very negative. In other words, what you’re saying is: keep up the price of AI as an input to the rest of the economy by protecting these foundation models’ pricing structure so that they can get $1 trillion in revenue.

The loser, as is always the case with trade restrictions, is the rest of the economy, which won’t get cheap intelligence. They’ll get dear intelligence. What it means is the stocks in your 401(k) that are making AI will continue to make money, and the stocks in your 401(k) that should be getting the lift from AI are going to lose money.

It would be as if, in 1981 or 1982, the governments—I’ll give a classic example: IBM licensed MS-DOS, and they should have protected it, but instead everyone was able to clone it, like Compaq. It would be like the U.S. government coming in and saying, “No Compaq, no Dell. You can’t make clones of the IBM. We’ve got to keep IBM’s stock price up, and therefore we’re going to ban the clones, and we’re going to keep the PC industry a small, tiny, profitable industry for IBM.”

It could happen in this case because the dynamic of overseas national security makes it happen, but it would be, for the record, so fricking dumb.

Jason Lemkin

Well, look, here’s the counterargument. You might be right; it might be dumb, right? I think if you step back for a minute—

Rory O'Driscoll

But to be clear, dumb things happen all the time, especially when governments are involved. I have high confidence in dumb. See Harry for details. I’m in the U.K. Sorry, Harry.

Jason Lemkin

They do.

Rory O'Driscoll

I have high confidence in dumb.

I thought you were going to stop at “See Harry for details,” and I was really hurt there, Rory.

Rory O'Driscoll

No, no, no. No.

Jason Lemkin

If you look at the history of LLMs on this show since it started, what has really changed—and Coinbase is just one example of this—we are probably leaving the oligopolistic age, right?

When we started the show, you could argue there were 2 monopolies. Anthropic owned coding. Elon Musk, when we started this, said, “Those Anthropic guys have something special. I can’t compete in coding.” OpenAI owned consumer.

They both had sort of weird pricing they made up and were losing lots of money, but they had these sort of dominant positions. As we began this year, we entered an oligopolistic era where we had 2 leaders. For a while, what happens in oligopolies traditionally is you compete aggressively on features and not on price. You essentially soft-collude on price, and the models were somewhat similarly priced.

OpenAI would be like, “Oh, we’re half the price for Codex,” but not really—we’re as competitive, right? Typically, when you leave oligopolistic stages, there’s massive price erosion because you’re no longer competing on features; you’re competing on price. That’s exactly what we’re seeing.

And governments do support oligopolies all the time when it’s—

Rory O'Driscoll

No, you fix pricing because of regulatory capture. I agree with what you’re saying: oligopolies are great for the people involved because, yes, you can have a price monopoly.

Jason Lemkin

And sometimes for innovation. If you look at the study, when you have folks that are brutally competing—we’re still competing—but if we agree that, basically, we’re $200 each for our Max program and we’re going to charge this much per token, how do you win?

You win because Anthropic comes in and builds this disruptive coding model and takes all that revenue, right? That’s how you win in an oligopolistic market when pricing doesn’t matter and features do. It can actually be great for innovation in the short term, but maybe not in the long term, right?

Rory O'Driscoll

I mean, include me out on that. Competition works. Again, going back, I think any number of examples says you’ve just got to let competition rip. You’ve got to let the open-source providers rip.

It’s the Coinbase example. The counterfactual is that if those open-source products didn’t exist, Brian Armstrong’s tweet would have been, “We were paying $10 million 6 months ago for our AI intelligence. Now we’re paying $60 million. What the frick do I do?” It wouldn’t have been as good a tweet.

We need competition to make this shit happen. I hear you, Jason. It might happen, by the way. You might see some pressure, but I think it would be a bummer. Move on.

Jason Lemkin

Hard to predict, at least, right? Hard—

Rory O'Driscoll

Yeah.

Jason Lemkin

Hard as of this date, I think it’s hard to predict. Hard to predict.

4. Microsoft Faces An AI Reckoning

We’re going to jump around a little bit here, but few companies have had as much of a competitive tussle when it comes to regulation and government intervention—or tried to—as Microsoft. Microsoft is in a pretty poor state. It’s their worst month since 2000, I found, and they’re down 16%, 16.5% as of today.

I’m a holder of Microsoft, and just when I think it can’t go lower, it does. What is going on?

Guys, can you help me out here?

Rory O'Driscoll

So Microsoft, I don't have a theory of the case for why, this month, it suddenly dawned on everyone that its strategic position was weak. Ironically, just when they had one of their big announcements and Satya made all this conversation about their direction on AI, maybe it suddenly opened people's eyes. Because we've been saying for the last year, “Look, they don't have the standalone model. Their AI story really is: We spent a lot of money on CapEx, which we do to support OpenAI. We have a good investment in OpenAI, which is looking pretty good, though perhaps not as good as it was a year ago. But our core software business doesn't have a compelling AI product.”

Let's get real: Cowork and Claude Code are eating the 2 core parts of the Microsoft franchise back in the day, which was Office for the individual knowledge worker, which is what Cowork is becoming, and developers, developers, as Steve Ballmer used to say, which is what Claude Code is becoming. So when you zoom out a million miles, Microsoft, which is in the software business unlike Apple and therefore can't afford to stay apart from the great AI wars, other than its equity ownership in OpenAI, doesn't have anything compelling to bring to the table.

I think the market's looking at that and going, “You're not going to zero, Harry. Don't panic. You've had your hit.” You don't get the lift that you get from the growth stories. They'll say they have an AI growth story, but if you break it out in the numbers, which they don't do, a lot of that is just, “Hey, we're selling inference to OpenAI, so we have a growth story.” But what they don't have is a compelling end-customer growth story here that they own themselves. So I think the market finally caught up with the bullshit.

Jason Lemkin

Look, I don't know the answer either, right? It could be... It's a very complex business model. Microsoft just has—it's a conglomerate with Xbox, which I'm not a total expert on, and a whole bunch of pieces.

Stepping back for a minute, I think what is troubling is Microsoft guiding that Azure growth is decelerating. I think any time you see deceleration at all, Rory can say, “Jesus Christ, how can you keep growing 40% at this size?” Law of gargantuan numbers, right? It's massive.

But I think, just as a student, the only way your stock price is going to trade up is if you beat, raise, and grow. You have to accelerate in this market, and you can be an oldie like Okta and Twilio and just your stock can blow up, or even Navan if you reaccelerate. But the market's expectations are so high that guiding to 37% from 40% growth is a fail.

It's a fail for Azure. This is why maybe that regulation is coming for those Chinese open-source models, because we just can't afford the deceleration. But Azure is not the only factor, right? It's so important, and it's almost a canary in the coal mine that if there's any issue in Azure, we should at least reflect on why it is.

If there's any slowdown in this era where we're claiming we're all going to be running 20 agents 24/7, Azure should be accelerating, shouldn't it, even at this scale?

Rory O'Driscoll

Yes, because, again, Azure—and I cited developers, end users, and knowledge workers—but you're talking about the separate aspect of the business, which is the cloud business. And you're right, the Azure cloud business decelerates. But I think more fundamentally, a huge slug of the Azure cloud business is simply inference for other providers like OpenAI. So yes, I agree.

Jason Lemkin

But it should accelerate, right? That should accelerate.

Rory O'Driscoll

Yes. If you recollect—and this is a harsh comment—3 years ago, Satya did the “We're going to make Google dance” comment, and the truth is, since then, Google has massively outperformed. For all their faults, and I still think they have significant issues and risks, they at least have their standalone model and a product to sell.

The truth is, today, Microsoft has a 30% ownership interest in OpenAI and doesn't have a state-of-the-art frontier model themselves. That's a big difference.

You know, it'll be interesting if Anthropic really does IPO in the coming months, if it stays on track. I have to imagine it's going to be one of those volatile stocks out there, right? Even a hint of news will send it up and down. I mean, even Cerebras is massively volatile, right?

But if Microsoft has this level of volatility, we're going to be reading every tea leaf about Anthropic. Anthropic looks perfectly linear. Well, I guess it's exponential; it all looks perfect and up and to the right today while it's private, right? All these rounds. But, man, I wonder how the volatility in that stock rises. A hint of bad news—boom.

5. Kalshi Chases Forty Billion

Jason Lemkin

Well, the experts in volatility are Kalshi, and Kalshi are apparently raising a new round at $40 billion. They raised their last round in May at $22 billion. Is this just the ultimate sign of the casinoization of society and risk-on mentality from consumers? They recently announced being at $2 billion in revenue. How do we feel about this?

Rory O'Driscoll

People like to bet, and the US had a prohibition on gambling on sports for the longest time. We talked about this last week. Poor Pete Rose never got into the Hall of Fame because he bet on himself. And now we have a couple-hundred-billion-dollar-a-year industry doing exactly the same.

So they're just riding that trend. They've got other products. It's not all sports betting, but it's about 70%-plus sports betting. It's killing it in revenue, so it totally makes sense.

Will Kalshi be a $100 billion company in 12 months' time?

Rory O'Driscoll

I doubt it. Look, you get there 1 of 2 ways. Either sports betting continues to expand and they can take disproportionate market share. That's 1 option. Or the other is the whole crypto-perpetuals business, which I just started to learn about, which is effectively kind of futures on crypto, which is gambling to the end, right?

Either that business turns out to be much bigger than we realize, which is not impossible. I mean, look, ICE—not the ice, but ICE, as in the owners of the NYSE, Intercontinental Exchange—has a big ownership stake in Polymarket. Why that's relevant is that ICE runs stock trading and real financial markets, and it's taking a significant—I think about a 20%—ownership interest in the other big kind of online betting prediction market player, Polymarket. They're clearly saying this is a huge thing that can have a $50 billion or $100 billion kind of valuation.

So it's not crazy. It happens 1 of 2 ways. Either, A, sports betting gets ultra-huge, and they don't get tied up by regulation and all the issues around that, or, B, the non-sports-betting side of prediction markets, in particular the financial products, becomes huge.

I don't think predicting who's going to win the general election in the UK or who's going to be the next president in the US is a huge, enormous, ginormous business. I love it. I find it fun. It's actually the most interesting part of Polymarket and Kalshi, but that's not going to get you to $100 billion, because the number of people who actually really want to bet on the next president is pretty low.

It's either sports betting, which works because we all love sports. Good luck tomorrow, Harry. And then it's financial betting, which we all love betting because we all love money. I'm sure if you could bet on sex, you'd have the trifecta. It's the human desires.

To get a big-ass company, you need to cater to big markets, and predictions on politicians is a small market. Betting is a big market in sports. People love sports. Betting on money is a big market. That's my point.

Dude, I totally agree. I did a deal into FOMO, which has got their new round on my index in USV, and their new product is Perps, which obviously allows you to do much simpler up-or-down bets on stock prices for consumers.

Rory O'Driscoll

Yeah, I actually went to try that product. That looks super handy.

It's incredible. It's amazing, dude.

Rory O'Driscoll

No, I mean, poor old Warren Buffett is like, “It's time for me to die because you people have lost the plot.” But if you want action on the table, then Perps is your boy, right? If you have a strong feeling on where the market's going in the next hour, then go for it.

Also, when you look at TAM expansion, Perps is a brilliant way to expand a TAM to a mega, mega market. Very exciting. Okay, SpaceX. Has SpaceX's IPO frozen the AI IPO market for now, given the size, magnitude, and weight of it? Does it put barriers up to subsequent IPOs?

6. Bending Spoons Reinvents Roll Ups

Rory O'Driscoll

I mean, it's worth pointing out that the anti-AI IPO is about to happen today. Literally, we're recording this on Tuesday, I think June 30. Bending Spoons is going public tomorrow, July 1, and this will come out on July 2.

Literally, Bending Spoons is the anti-IPO. It's a buy-and-build company, and they own AOL, for God's sake, right? Which, by the way, I remember, has killed its AOL email product but still is an ad network. They own Evernote. They own a bunch of stuff like that.

So that's a company that's literally a 20-year-old software company going public at $20 billion. It's not like the IPO market is shut, right? Which is what I thought you were originally saying.

Jason Lemkin

What you're basically saying is: Is the hidden question here, Harry, has the volatility around SpaceX made OpenAI and Anthropic nervous about going out? Is that really your question?

Yeah.

Jason Lemkin

I thought this was more interesting when I added it to the list, but as I reflect on it, I think as long as the IPO remains up, I think they'll be fine. The volatility has been high and extreme. As we record this, greed will still trump fear, right?

But the volatility has been sufficiently high that you could imagine that changing before now and the day Anthropic IPOs. So I'm sure the bankers who have less to do are monitoring this every day, and I think the board and Dario will think about it weekly, just making sure it is the right time. It definitely shows there's a risk it isn't the right time. It shows there is some risk that Anthropic might delay its IPO. The volatility says it's not a slam dunk, but up is up. Like, everyone's—except for the poor Korean bank that forgot to put in their orders—most folks that at least bought in the IPO are up.

Rory, you brought up Bending Spoons. When we look at multiples attached to IPOs, Bending Spoons is going to go out at $20 billion, which will be a pretty hefty multiple. I think it's about a 12 to 14x multiple, given revenues of 1.5 billion.

Rory O'Driscoll

Yeah, 1.5 billion in trailing revenue, but $600 million in Q1. So probably 8 or 9x forward revenue—a healthy multiple.

For the antithesis of an AI company, is that not an extremely juiced-up multiple?

Rory O'Driscoll

It is funny because you're exactly right. It's like you have a whole bunch of single-product crap, B2B SaaS companies that have slowed to 10% growth and are trading at 3 times revenue, and then you have this company, which is an agglomeration of a whole bunch of tired consumer products that were growing at 10% until these guys took them over.

And when you read the S-1, you realize they don't get organic growth, they don't get new users. They just raise prices, cut costs, and suddenly that's being valued at 8, 9, 10 times revenue. So it definitely feels kind of a little weird.

It would be as if all the SaaS companies that were slow growth were trading at 2 times revenue, and Constellation Software, which is a roll-up of SaaS companies, was trading at 9 times. It definitely feels like, wow, that was a clever way to make money. You do wonder, is it a little faulty? But good luck to them.

Jason Lemkin

I think it's going to do well, over the medium term. Here's why. Good God, don't take the risk factors in the prospectus seriously, right? Or whatever the prospectus said. But there's a grain of truth. They said there's at least 1,000 businesses they've already identified that are material and can move the needle.

I think if they're this good at buying and repackaging these companies and there are 1,000 targets, they can maintain outlier growth rates for longer than we would be in the stock, right? So can they—I guess the meta-question is, going back to the Microsoft issue, note Rory's going to think it's an odd tie—can they maintain this outlier growth for 5-plus years to justify some sort of premium?

We can debate whether the premium's too high, but outlier growth gets a premium. I think without question, if they can execute at the level they have, it's justified for 5 years because there are 1,000 targets. And I think there's also going to be another Bending Spoons in B2B that's going to nail this.

They're going to buy these horrific products like Marketo and others, put a few smart people in them, and just boost the NRR. They're just going to boost the NRR and a few other things. And so I think there's a chance for several Bending Spoons to take the struggling software companies and do a revenue arbitrage because they can package them together into something high-growth. I think there's 5 good years here.

Rory O'Driscoll

I think that's smart, Jason, because you're exactly right. If you're one of those 1,000 privately held companies, and let's assume they're all consumer, there's nowhere else to go. I mean, that's the big attraction. These guys are the only way out, so they can lock—at least until someone else comes along—they can lock in some kind of revenue arbitrage.

Yeah, that's interesting. And yeah, you should be looking at it for B2B.

Jason, you are the new CEO of Bending Spoons, it goes B2B, and you are able to choose 3 targets for your opening quarter. Which 3 targets would you choose?

Jason Lemkin

I mean, Bending Spoons has one trick, which is it uses a lot of folks in Italy who are lower-cost for engineering, right? But I assume the other trick it has is that it's able to incentivize GMs to do a decent job on these. Find kids—and these kids may be 60 and not 16, well, 26—but find kids to run these Eventbrites and AOLs and...

Let's assume you have a steady enough flow of these kids who give a shit, okay? Then I would start to buy up almost any B2B company with 9 figures in revenue and a sticky customer base because, literally, the worst product that we use today is our most expensive product. It's Marketo. Rory will remember it.

They threaten us. The API doesn't work. It breaks every day. They just told us they're raising prices 20% next year, like Bending Spoons, but without any features or functionality. The site went down for a full day the other day.

You put a kid in here who gives a crap, okay? And you say, “I just want you to take whatever Marketo still has at Adobe—$300 million of revenue left, right? It's probably decaying. Take that $300 million and give a crap. Don't threaten your customers. Actually launch some features. Remove the rate limits on the API. Do something so that, to retain your base, it would work.”

Even with our worst software, we would stay. It's not that hard. So I would take all these companies that have a sticky base and buy them. Then you take 10 companies at $200 million, you've got a $2 billion business growing 30% to 40%, and stack it. You've got something pretty nice.

But the problem is they're just putting mediocre people in charge of these. The non-Bending Spoons companies are having PE firms put these recycled mediocre executives in charge of them, who are going on learning tours for 90 days and coming up with no ideas. That's just not going to work.

Can you buy these assets for a reasonable price?

Rory O'Driscoll

I think you can buy some, yeah. What's funny about this is the minute you ask this question, you can tell Jason that I'm going to do it too. You have this hot button of these markets that have been bugging you and companies where you're like, “For God's sake, will you do something?”

Yeah, I think the Marketo marketing automation space is one we both know, and there's something to be done there. I'll give you 2 examples, 1 of which is traded already. And I know Jason hates the first market—I said it before: Semrush.

Semrush is the old SEO optimization. It's so obvious that every one of those customers is going to need GEO, AI optimization. You should buy Semrush, buy some other little tool, and just bundle them together and sell them.

You can't do that now because Adobe bought Semrush for under 2 times revenue. Great deal, right? You could have turned that thing into something. And then another example, which the world continues to poke at, is PagerDuty, which is a company that's been out there. It has commanding market share, and it's never added AI-enabled incident resolution.

Jason Lemkin

Yeah, put a kid on it. If you can buy PagerDuty for what it's worth, put a kid on it who's really motivated. It could work.

Rory O'Driscoll

I totally agree. A business-oriented executive who maybe picks up a failed YC software incident-automation product and puts it together and, as Jason says, if you're not reaccelerating, you're dying.

But if you can get reacceleration to 20% by just upselling a slug of the install base, at 2x revenue, you could be Jason Lemkin's Bending Spoons B2B—BS B2B.

Jason Lemkin

I think what folks don't realize is that so many of these companies we're talking about—I can tell you because we interact with these people—they've given up. The entire team, their customer success team, has become a force of evil, right? They threaten you with lawsuits. They threaten you with everything. You can smell that their culture is not that we're not in terminal decline.

I used to have a guy on my sales team whose last job was working for the Yellow Pages. His job was that he got a huge bonus if his patch shrank less than 20% each year. He was one of the top performers. His patch shrank by 16% every year. I mean, at some point you've got to move on from that job.

I feel like this is working with the Marketo team and other teams we work with: the knives are out, okay? So my point is, it is not hard to turn around a team that has completely given up if they have a sticky customer base. This is not...

But you've got to find people who want it, and the cultures are just broken. It's not just Marketo. They're broken everywhere, right? These cultures have given up. So I like this model and I like the 1,000 targets. I think there's 1,000 targets for Bending Spoons, and I almost wonder if Constellation needs...

Listen, they've been wildly successful. I wonder if their model needs to be rebooted so they can get it up. I don't know if these kids want to work for Constellation or not. In the portfolio company where I've watched PE take them over, they're not running the right model.

I've sold a lot of companies to PE. I'm not close to them, but I watch them. They're still bringing in 2021 managers. They're bringing in folks who have never been a CEO before. They were never really a great success on their own, but they have a good set of logos on their account and they're a people person.

Jason's really a people person who's going on a speaking tour, who's going around the world for 90 days to just meet with the team, and then, “I want to meet back with the board in 3 or 4 months and come up with my ideas.” That doesn't work today, boys. That was great when Thoma Bravo bought you in 2021, right? We need the Bending Spoons—I bet at Bending Spoons, when you buy that thing, man, crap happens in the first 30 days. People are moved out. People are moved in. Products are shut down. We need that level of action, man.

Rory O'Driscoll

Pushing on that, it may be in part because—and I could be wrong—but what Bending Spoons is trying to do is similar to what Constellation and PE were doing 5 or 10 years ago in B2B SaaS: in other words, don't change the business all that much, but just optimize it. And you're right. So that's an easier task.

I think what you're saying, Jason, is true: if you buy a B2B software company today that's pre-AI, it is highly unlikely that simply optimizing and pressing the buttons will be enough. In other words, you can play the Vista moves from 2021. You can cut the costs. You can move headcount overseas, et cetera, et cetera. But I think what you're saying is correct. Unless you're generating new revenue from AI and significantly re-engineering the company—which is a bigger ask than just optimizing the existing thing—it won't be successful.

To that extent, I think that Jason's B2B—B2B dot, sorry, Bending Spoons B2B run by Jason—will actually be a harder managerial task than Bending Spoons, because I think all Bending Spoons had to do was take Evernote, take AOL, and just ruthlessly raise prices and optimize. There was a little less innovation required than I think would be required in B2B now. Just thinking aloud here. I hadn't thought about that until now. But I think the upside could be bigger.

You also need a monster checkbook. PagerDuty's market cap says $750 million. You buy it at 2×, you're $1.5 billion on 1 asset. We times that up by 5—

Jason Lemkin

No, no, you buy it at $760 million. No one's buying PagerDuty for 2×. The board would have to take that deal in 6 seconds from a fiduciary obligation, right?

You'd need to sell it at $760 million.

Jason Lemkin

I think any public company today—I mean, Rory's lived this more than I have—I think any public company in decline today that gets an offer at a premium of even 15% has a fiduciary duty to take that very seriously. They have to come back and say, “We genuinely believe this thing. We've got to go hire an investment bank and say it's underpriced at a 15% premium to market.” And then management, with their earn-out, they're going to take that deal in a heartbeat.

“I get to leave? I get to bail out of this sinking rat hole? I'll take this.” Management is going to be so aligned to take any deal where—some of them just get fired. They're not going to be excited about that, right?

Rory O'Driscoll

They just had a good day and are up a few bucks, but yes. I mean, it's $700 or $800 million. You're right. But that's because Jason wanted to start at the $100 million level. You could do the Bending Spoons thing and start with smaller deals and roll up to it.

Well, that's what they did. It's a very long journey. This company is, I think, 15 years old.

Rory O'Driscoll

Yeah, but the point is, I think the meta point Jason's making is correct, especially in a world where standalone IPOs need $500 million and 30% growth. There are a whole lot of companies that aren't that, right? That are below that scale, that—this is going to sound stupid when I say it—but aren't family businesses you leave to your kids. They're venture-backed things with a CEO, and at some point everyone gets old, everyone wants to do something else, and all those businesses have to find a home.

So yeah, I don't think it's crazy. I'd love to see the Bending Spoons consumer list of 1,000 names, but I believe it would be there. Good digital assets that are just like, “Eh, don't matter enough.”

Jason, I want to push you. You've got Marketo, one. What are the other 2 targets we're going for?

Jason Lemkin

PagerDuty was a good one, right? I mean, you need to—Rory's right. That's a very good one because you've got the right customer base, right? Datadog is extremely expensive. They have cheaper competitors. PagerDuty got crushed from all sides, across its whole suite, right? But its customer count is flat, and it's real. It's still got 15,000 customers or so paying for this product constantly. That was a good one.

Asana?

Jason Lemkin

Boy, I have too much scar tissue from attempting to use that product. But probably, yeah. Listen, you have a company where the billionaire founder just up and quit a year and a half ago. That's one where there's got to be some way to turn this thing around in a space that, listen, has existential challenges, right?

Do agents need Asana? Agents don't need Asana. But I think you can make it more agent-friendly. I think you can do better. Probably do better—this is easy for me to say. I don't want to be too much of an armchair quarterback from a product that feels terminal, but it's probably a good one. I think it's a good candidate, right? It is a good candidate. Without all the pressure, too, right?

7. Chamath Builds An AI Startup

Okay, I feel like there's a private company fundraising story that we need to touch on before we do a rage-bait reel. Chamath Palihapitiya has raised $135 million for his AI startup, 8090, or 8090, whatever he calls it, but he is also now CEO. For people who don't know, what do they do? It's a software factory platform that lets teams collaborate with AI to handle the full software-development cycle, from new builds to code refactoring, complete with governance. What do we think about this one, team?

Rory O'Driscoll

I just love the fact that you had Chamath now in the rage-bait category. That in and of itself was progress, right? All credit to him. At the risk of making the cliché, he used that quote a long time ago, but give him credit. He gave the “Man in the Arena” quote, and he got slammed for it. Well, now he is the man in the arena, and all credit to him for trying. Good luck, right?

It's a super-interesting market. There's obviously a ton, a ton of competition, but as he said in his tweet, it's the most exciting space you've seen in decades. How is all of software going to be remade? So I just give him huge credit for going for it, and good luck.

One of the things I'm trying to do, Harry, is I don't have to be snarky all the time. It's easy to be snarky at someone like Chamath, who just lends himself so well to snarkiness, given his stuff. But let's just take the high ground and say, “Well done. Good luck.” Go team.

Jason Lemkin

I don't mean to be snarky. I will say 1 thing in general, right? This is abstracting away from Chamath. There are counterexamples. I can give a few counterexamples. But my scar tissue is that I don't believe he's working 100% on this. I don't believe every waking hour is on this. I believe he's got a team at this point in life.

When wealthy folks—especially VCs—want to be a CEO but they're not working at the insane rate of a traditional founder-CEO, I find those run out of energy. I'm not saying he's not the exception to the rule, right? I know the Spotify guy that Harry's close with runs this scanning company too. That one. We have a lot of CEOs running side companies that are very successful, right?

Listen, anyone that's started a startup, it's all fun and games in the early days if you have any money. You pull together a team. There's a lot of whiteboard talking. You use your brand to get Accenture, whoever's backing you. It's all kind of fun until the shit hits the fan.

But do you want to be running this services AI business forever, 100 hours a week, or do you want to be hosting $25 million fundraisers in your palatial home? I mean, I don't know that you can do both successfully. I know there are examples. I just wouldn't invest.

There are certain things where, for me personally, I have scar tissue. I've written small checks into successful founders doing multiple things just for fun, and they're all zeros, right? I remember 1 of them. I asked this successful founder, “I don't really need to know what—I don't care about the valuation. I don't—just, are you telling me this is the only thing you're going to be doing?”

He's like, “Yeah, this is the only thing I'm going to be doing.” 30 days later, I see him working on another startup, right? So I'm out.

If literally Chamath drops everything but 8090—everything—and I see the sweat from his brow, and I see that—because the dude's fit and he's looking good, right? He's just had a massive win, the largest win of his career as an investor.

Jason Lemkin

I want to see the paunchy middle, the hair loss, his right-hand person quitting on him. Then I'll invest. Not until then. Too much scar tissue here.

It's too easy to start up today. Whether you launch into YC and raise at 30 pre when you started 3 weeks ago, it's not easy, but it's too easy. It's just too easy to start up today. Seeds for suckers, boys. It's for suckers.

We're making the T-shirts. I really raged the internet again this week.

8. The New Series A Bar

I turned down a founder this week. Why? They were finishing the year at $1.5 million ARR, finishing next year at $5 million ARR. Today, brutal as it is, this isn't good enough to raise a good Series A. The opportunity cost of cash is real.

I deleted it because, honestly, no one engaged. It got 30 likes after an hour, which is not very much for our tweets. So I took it down because no one cared and it looked bad on my timeline. And then they put it back up and it became a thing. Do you think I was wrong?

Rory O'Driscoll

Okay, I can answer: yes. I think the factual statement you made is correct. In the kind of growth rates you're seeing now, the bar is higher, and there's an opportunity cost to doing something with a lower growth rate. It's not impossible.

We have done deals with those kinds of growth rates, which would have been very top-quartile in the age of SaaS but aren't in the age of AI. I can imagine doing some of those deals, but it's the exception, and you'd need some other extenuating factor. As a matter of pure truth, it was a correct reflection of the current venture market.

Were you correct to put it up? If you're not careful, as a VC, when you're saying your $1.5 million-to-$5 million deal isn't, quote, “good enough,” you really have to phrase it carefully so you don't sound like an obnoxious prick telling people their life's work is not good enough. That's tricky, Harry.

Well, I'm really sorry. If that's going to ruin your day, don't be a fucking founder. Life's harder than a VC tweeting and ruining your day.

Rory O'Driscoll

Agreed.

Right?

Rory O'Driscoll

I think life is hard, but it's precisely because the rest of your life is so hard that a little bit of compassion from the capital wouldn't hurt. But your message is correct.

I had loads of founders and VCs saying, “Oh, classic Harry.” I was like, “Dude, Legora, Lovable, Mercor, Fireflies.”

Rory O'Driscoll

No, I, look, you could've—

Come on.

Rory O'Driscoll

I'm just pushing it. Now, look, yes. A lot of people say you're mean, but what you're saying is correct. Again, we've discussed this before. You think rage is still engagement, so you're all happy. I definitely don't think you should have taken it down, because that looked like you were blinking. Yeah, that was a mistake. You blinked, dude. That's the bad part.

Jason Lemkin

Can I break it down just a little bit on the tweet? First of all, listen, I'm supportive of it. I would have retweeted it, and I'll also be supportive here. I think the problem with the tweet is that there are 2 things going on. The first part of the tweet is the state of the Series A market. The second part, which is more triggering, although people might not miss it, is that the opportunity cost of cash is real. Those are different points.

I'll tell you something that I think is subtly toxic that all these nice VCs are doing. I've watched 2 portfolio companies recently that were growing at great rates and are going to compound to huge winners. They're capital-efficient, but they're not quite at Harry's level—not at the A, at the B or the C.

I've watched all the VCs say, “Good luck, guys. Go do your round.” I built an AI pitch deck generator that uses all the benchmarks from ICONIQ and Bessemer. It tells you honestly what your odds are. It told all of them that, for a round, they're a B. Okay, it will tell you. Go to Saastr.ai, use pitch, just upload your pitch deck, it will tell you.

Not a single VC in either of these companies would be honest with the founders. I've tried in the past. I get my head cut off. In both of these cases, I said, “Here's my new thing. The hot companies in your space are basically later-stage, right? They're getting funded at, like, 30X ARR. Back-solve into what numbers you would need to get to raise at 30X ARR.”

This is me trying to guide founders there. It's too subtle. They don't listen. Then 2 months later, they're like, “Hey, I'm growing at this still top-10% rate, but not enough.” No one's honest. So the honesty of Harry's thing was very helpful. The opportunity cost of cash is a little more for people to process, right?

Rory O'Driscoll

For what it's worth, I do agree. I don't think I'm changing my mind, but I like what you said, Jason, because I'm playing the pattern back and I've seen the same thing.

You look at these companies and realize that everyone around the table knows that's not a deal they'd do, but they say to the company, “Have a go,” and they're not being honest. They're looking at the growth rate and saying, “That's not compelling compared to the other things I'm seeing. Just have a go,” when perhaps the better advice might be, “If you're only growing at 50%, should you converge on profitability? Should you raise a lower amount?”

You're right. Sending people out to get a harsh message from the market just because you're too big a wuss to give it in the boardroom is actually a pretty pathetic act, and I do agree with you, Jason. To some extent, I'm backing into Harry's point.

Do you know what? I told this founder exactly this, which is why I actually tweeted. No bullshit. He said, “Do you know what? That's really helpful. I had no idea that wasn't good enough.”

He was super receptive, and he was like, “Honestly, dude, I didn't know that. That's really helpful, and I'll change how I project future revenues.”

Jason Lemkin

To dig deeper—maybe Harry doesn't want to go this deep—I do think there's a logic here. Here's what's wrong in Harry's tweet. Harry's tweet is, “Turned down a founder at $1.5 million, finishing next year at $5 million. That's not good enough to raise a Series A.”

If this was a long tweet, it might be, “Or maybe it is, but you might have to meet 150 VCs.” That's what I would add. If people are honest, I think they should be like, “Listen, you're at the edge.”

Rory O'Driscoll

Agreed.

Jason Lemkin

There's nothing wrong with $1.5 million to $5 million. You did better than I did back in the day. There's nothing wrong with it. If you put it on a spreadsheet, and your burn is low and you don't quit, you can build a generational company with those numbers.

But what it means is that it's literally frigging hard to get attention. Higgsfield, where I invested at seed and Harry invested, is just crossing $500 million in revenue in less than 18 months.

When I thought about that today, I looked back on my email and thought, “Why didn't I invest even more?” Then I get another email from a portfolio company growing at decent rates. It's hard to even pay attention, right? It's hard to even pay attention.

You have to realize that just getting attention is hard in this crazy world, and you've got to hunt Higgsfield or better. That's your job, right? When I started talking about Higgsfield on the show, nobody had even heard of it. You know why that's interesting? That means go find it. Go find it and stop worrying about the $1.5 million to $5 million.

If you talk to 150 investors, you're going to find someone who believes in you. The $1.5 million to $5 million doesn't really matter, does it? It's where it's going to go over the next decade, right? Someone may take that bet.

But don't run a process. Don't build a data room, give people 1 week to look at it, and ask for checks, right? Give it time.

Rory O'Driscoll

Yeah, no fast process. I think some versions of what you're saying are right. If you're at $1.5 million to $5 million, it has to be, “If you want to raise money, just understand the facts.”

The deals that are getting swept up are going from $1.5 million to $15 million. You're not that, so that has consequences. It has consequences in terms of the number of people you'll have to talk to, the range of people you have to talk to, the amount of capital you can realistically raise, et cetera.

I don't discount the fact, but I just want to say this because founders are listening: I do agree with you, Jason. You can be at $1.5 million to $5 million and still end up with an amazing, generational company because, yeah, we've seen that data.

I can't remember when, but there is a correlation—it's modest—between initial growth rate and overall outcomes. Companies that have grown slowly at the start have been huge at the end. CoreWeave was a slow grower and then became a huge outcome.

I do agree with you. It's not that we're saying to a founder, “Your dream is impossible. Go away and die.” That's just not productive, especially when they're growing from $1.5 million to $5 million.

What you do want to say is, “If this is your reality, you better think about how to cut your cloth accordingly and how you plan your raises.”

And maybe you're not a venture asset anymore in a new world of venture, and that's totally fine. That's okay, too.

Jason Lemkin

You might be. I still think—listen, I just think there are 2 different tweets in your tweet, right? To Rory’s point, the reality is 95% of investors you’re going to meet today are going to say the opportunity cost of cash is too high here, right? They’re going to believe that. And even if they don’t believe it, their job’s on the line. They have to find a high flyer.

Even if they’d be happy to do this deal, they might get fired, right, if they don’t run the place, right? So they’ve got to find—everyone’s got to have 1 of these lighthouse investments in their portfolio, or you just might not be part of the next fund, right? This is a reality issue, right? That’s a different tweet from, “What are the odds, if you’re at 1.5 going to 5, that you’re going to raise funding today?” They’re just different tweets.

And listen, I’ve got your back, but you gave people 2 different reasons to get triggered, and you saw their reaction, right? It doesn’t mean any of it was—it was all correct, though. It’s all correct.

Rory O'Driscoll

Yeah. You should just run the Jack Nicholson “You can’t handle the truth” quote—that little clip of Jack doing that in the movie. That’s what you’re saying, Harry: They just can’t handle the truth.

But you are a little bit punky, though.

It’s going to work in my favor, isn’t it? Really win them back with that one, Rory.

Rory O'Driscoll

Yeah, you really—

Jason Lemkin

You might—

Rory O'Driscoll

…win them back. You’re really going to win them back.

Is there anything else?

9. Claude Tag Enters Slack

Rory O'Driscoll

Yeah, 1 thing. I’d be curious to get people’s thoughts on the whole—you’ve mentioned here Claude Tag and Claude Tag in Slack. Jason, I’d love to get your thoughts on that.

Can we just provide some context? What is Claude Tag, for those that don’t know? Just some context here.

Rory O'Driscoll

It’s basically the ability to have Claude as a fully present member of a Slack channel focused just on whatever that is. If you look at the announcement, it could be your legal team. It’s a Claude agent for legal that’s just focused on legal, that just has access to that sort of information, but is a fully present member of your channel.

Jason Lemkin

Yeah, and in theory, it’s autonomous.

Rory O'Driscoll

Yes.

Jason Lemkin

Right? That’s the—that’s the point. It’s not just that it has—well, listen, I don’t know, because I tried to deploy Claude Tag, right? But you have to—I'm not—

Rory O'Driscoll

You have to be an administrator.

Jason Lemkin

You might have to. I think we’re just—I’m just not on the right enterprise plan, and Claude is the biggest issue. So I haven’t used it. I tried to use it for the show, right?

So listen, a lot of things like Claude Design—before the internet and information says the world has ended, let’s actually see how important this product is to Anthropic. If this product is existentially important to Anthropic, this could be the biggest deal for traditional software there ever is, right? It runs cross-platform. It runs on Salesforce, it runs on HubSpot, it runs on all the other things.

If the agent can run 24/7 autonomously, take all your data in, build all the analytics, build all the dashboards, and run autonomously out of it, then your data can flow between apps and you won’t even care where it lives. All the fears about headless become true because Claude is your head. Salesforce and HubSpot really do become dumb databases.

There is a version of this where Anthropic puts its best people on it and doesn’t quit, where it is existential to everything. Let’s give it a week, right? Or is this Zapier on steroids? Or is it even very good? Because Slack has a Slackbot, which is pretty good, right?

One of the things the press was asking was, “Why is Salesforce supporting this when they launched their own version of this a couple of months ago?” Well, what choice do you have, right, at some level? But I wouldn’t be surprised if this vector doesn’t maintain so much energy and Slackbot is better, right? That’d be the most logical thing, but we could be wrong.

Enterprise is the big battlefield, and as much as Claude Tag created some anxiety at Salesforce, it might be the Trojan horse and, in 6 months, it’s like it’s a big effing deal. But we have no evidence of that, right? We have no evidence Claude Design is going to kill Figma in any way, shape, or form at this point. We have no evidence there’s a long-term commitment to that, so I’m skeptical. But existentially, man, it could disrupt everything in software—in business software.

Rory O'Driscoll

That was a good summary, Jason. I agree. Someone asked, “Why did Salesforce let it happen?” They have no choice. They own Slack, they have their Slackbot, but you can’t be the cross-platform comms—you know, communications platform—for your company and then not allow access to an agent that’s enabling you to do better work, because that just pushes people away from you.

So I think the interesting thing here is, if you’re lurking on a Slack channel as an AI, you really just get a very good handle on how people do the context part of work—that lovely post from J.R. Storment and Foundation Capital talking about capturing context. Basically, that’s a fancy word for capturing all the weird shit people do on top of the actual apps, which reflects how they actually do their work and how they configure their work to suit the SaaS apps, which is what happened in the prior generation of software.

Capturing that context is really useful because it allows an AI to automate that work. The truth is, a lot of that context exists in Slack. So if you’re watching people interact on Slack, and you watch it autonomously for weeks and months on end, you probably will get a pretty good handle on how people do work—how Jason and Rory handle whatever exception we’re dealing with and when we’re talking about it on Slack.

So I think it’s an interesting entry point. You’re right, it’s only an entry point. It’s not the end of the world. But I think Salesforce is right to say, “Okay, you’re in there now, and we’re going to make sure that the Slackbot is better and remains better.” So, yeah, I agree with your assessment. Super interesting. Watch this space. But definitely an interesting entry point into capturing what’s going on at the context graph level.

Jason Lemkin

By the end of this year, Anthropic will have more revenue than every public software company combined.

Rory O'Driscoll

Unless Brian Armstrong has his way and cuts it in half.

Jason Lemkin

Yeah, maybe, but for perspective, we just have to wonder: Is some of the stuff that we think is very important, or that the media acts like is very important, just not material to Anthropic and OpenAI? It’s just not.

As these guys cross 100 billion in revenue, 200 billion in revenue, they may be like the early days, when I started in B2B as a founder and when I first met Rory. Most folks thought it was just too small. It just wasn’t worth anybody’s time. These markets were just too small. Now they’ve got big, but AI got so much bigger, and they may not—it just may not be worth Dario’s time to worry about whether he’s disrupting Salesforce.

It’s not even Salesforce’s 42 billion. He may be looking at the net new bookings. What is Salesforce adding at 10%? 8 billion? He’s like—because, you know, materiality has always been 10%, right? So if I can’t make 10 billion-plus, I don’t know if Anthropic can get out of bed for something that doesn’t generate 10 billion of revenue.

Maybe that’s always been the definition of materiality in my experience. I think it was even the SEC’s, right? 10%—you had to disclose it in the old days. I don’t know if Anthropic can get out of bed for less than 10 billion of revenue by the end of the year. It’s just not enough. It’s just not enough.

Now, it’s one thing to just do an experiment or build something that makes Claude better, right? That’s an integration. Like, we’ll make Claude better. We’ll integrate more with every single app and take in the data, but I think they may not care about that revenue that the leaders are terrified of losing.

That’s probably why there was that crazy disconnect with the guy leaving the Figma board, right? There was such drama for Dylan. To Anthropic, they’re like, “Oh, we didn’t know you’d care.”

Rory O'Driscoll

Gotcha.

Jason Lemkin

This isn’t even important.

Rory O'Driscoll

Yeah, sorry—

Jason Lemkin

Sorry.

Rory O'Driscoll

What? How bad?

Jason Lemkin

Genuinely sorry. We didn’t even—this wasn’t even—we don’t even talk about this each week at the—

Rory O'Driscoll

It’s the—you know, when elephants dance, the little people get trampled. Exactly.

Jason Lemkin

Yeah, we didn’t even know.

Rory O'Driscoll

Didn’t even know.

Jason Lemkin

Sorry—genuinely sorry.

Rory O'Driscoll

Yeah, whoopsie. Next time we’ll be more careful.

Boys, it’s a wrap from the British beach. Thank you so much for this, Rory. Always a killer line at the end there. What is it? “When elephants dance, the little people get trampled.”

Rory O'Driscoll

I think it’s something about mice getting trampled or something. Whatever. Yeah, no. Well, Harry, by the time this comes out, we’ll know how the US and England have done, and by the time I see you next week, hopefully we’ll both be progressing to, let me see, the round of—

I’m going down the pub for the game.