Meta 挑战 OpenAI｜Apple 起诉 OpenAI｜SK Hynix 265亿美元 IPO

Jason Calacanis

As 3 inveterate Twitter junkies here, we know where the action is, and it ain't on Threads.

Okay, let's be honest, people. If there was action on Threads, Harry, you'd be there.

I would.

Jason Calacanis

If there was action on Mars, Harry, you'd be there this week.

Number 1, Apple sues OpenAI for trade secret theft. Number 2, Meta fights back on the coding model front, and Zuck is back on X. And then number 3, SK Hynix prices a $26.5 billion Nasdaq listing, bringing the compute and infrastructure layer to the public domain.

Jason Calacanis

To some extent, it may have been a distraction for OpenAI that they were so successful in consumer. If this was a mercy killing, Apple may have done them a favor. In the early stage, when you're on the board, you can't invest in 2 competitors. In the late stage, structurally, you have to.

If you don't want to be worth 1x, do something before it's too late, man. Ready to get him?

Guys, it is so good to be back. As always, there is a lot for us to discuss. Why don't we start with the trillion-dollar company that is Apple, as it brings in another trillion-dollar company in the form of OpenAI, where Apple sues OpenAI for trade secret theft?

Essentially, one of Apple's long-term, 24-year vets was taking secrets to OpenAI and sharing information that he shouldn't have been. Just to be clear, this is a really serious thing, and there are actually some conclusions that we can share with listeners in a second.

But on the facts, the 24-year veteran is Tang Tan, and he's named in the suit. The person who allegedly took the secrets was the 6-year Apple employee Chenling Tang. Tang Tan is in charge of the hardware division at OpenAI after having been a senior executive at Apple. The allegation against him is that he encouraged hires to bring the stuff over, right? Which is just slightly different, right?

Jason Calacanis

Absolutely. And he encouraged them to bring actual prototypes for show-and-tell.

Exactly. I think the big picture here is—and let's start with the advice to anyone listening—don't ever do this. The individual who brought that stuff across, I read the lawsuit this morning—41 pages. I didn't read it all, but I got through a good slug of it. There's a lot of fact-based evidence that the person who took this stuff did it, and he's screwed.

The first thing he's going to discover is that there's no gratitude in litigation, right? You think you're helping your new employer, or trying to get a job, and they're going to burn you so fast your head's going to turn. They're going to say, like Claude Rains in Casablanca, “I'm shocked and appalled to find that there's gambling going on here,” and he's going to be left high and dry. Don't ever, as someone interviewing for a job, do this, because you're going to be screwed. That guy's toast, right?

Jason Calacanis

I hadn't thought about that. You think you're helping your company. He's already lost his job. Forget what his compensation or exit package is—he's already gone, right? You're not helping your company.

Agreed. So here's the next level up: the gentleman you mentioned, Tang Tang. He also has some issues. It's not as clear because there's no fact-based evidence, but it's not just hearsay, because a bunch of folks have said, “He did it. He encouraged me to do this.”

Even in the email traffic with Changlu, there was a lot of, “Hey, we were asked to do this.” So the risk he's up against is that there's now going to be a ton of depositions and a ton of discovery. If there's any smoking gun that has him emailing or encouraging people to say, “Hey, bring in information when you're coming in for your next interview here at OpenAI,” then he's screwed, too.

The same thing will happen, because now you get to the highest level. Look, Apple, in one sense, is happy because they're really pissed that 400 people from Apple have gone to move to OpenAI. Traditionally, over the years, Apple has always hated, in my view almost to the point of not being appropriate, people leaving their place and going to someone else, and has really leaned on other companies not to hire them.

At some point, that gets inappropriate, and you see this whole Steve Jobs kind of issue from 15 years ago. But the bottom line is, they're not looking at this and going, you know, the way when you want to get something from someone and then the other guy makes a faux pas, and now they have leverage. They know they have the junior guy, they probably have the VP of engineering, and they're going to twist the knife, depose everyone, and see how high up it goes.

They're pissed, and they've got leverage. It's a tough place to be, right? Because what they really want isn't—I mean, they care about these secrets, but what they're really dealing with is the big-picture comment that OpenAI has been talking about building a hardware device. That's obviously why they're hiring 400 people from Apple, and Apple is pissed. Now they've got some leverage, and they're going to find out what's going on here.

Jason Calacanis

That's too much, right? But still, you don't always get sued, and it's tough. But if we can't hire domain experts, sometimes we can't all invent LLMs ourselves. Sometimes we need a little help.

I think, Jason, you're exactly right, but I actually think it makes the point that you didn't need to do this. Stepping back, California is one of the most liberal places in the country in terms of employment, which is that there are no non-competes. Essentially, non-competes are typically unenforceable.

There's also this doctrine called inevitable disclosure, which is: if you acquire knowledge in one company, go to another company, and use that knowledge, the California courts are pretty friendly and they'll say it's inevitable because it's part of your trade. They'll even say, “We can't stop you from using the tricks you learned at Company A at Company B.” Let me give you a very practical example of that.

Jason Calacanis

Anthropic is the biggest beneficiary of the inability to enforce non-competes and the doctrine of inevitable disclosure. Those 7 employees who left basically walked out on them because they're way smarter than this guy.

They probably didn't bring anything: no pieces of paper, no stolen code. But the next morning, they start up and say, “We were just thinking about this last week, and we can keep thinking about this. Everything in our brain is owned by us.” And there are lots of states in the Union where you can't do that.

I wonder if Anthropic could have been started in Massachusetts. Probably not.

Jason Calacanis

No, you're exactly right. California's—even though it sucks as the employer—I think California's been right. Oddly enough, it's been a rare policy triumph for the great state of California. It's been really good for innovation in the Bay Area.

You already have that very employee-friendly environment, and given that inevitability, why put yourself at risk by stealing something that you just don't need?

My question was: these individuals are screwed. To be blunt, what does it mean in terms of timeline for OpenAI? Does it cause meaningful delays? Does it impact their ability to ship their hardware product when they wanted to?

1. Is OpenAI's Hardware Dream Already Dead?

Listen, if I had to read the tea leaves, hardware is probably on the bubble. This was one of Sam's initiatives that made sense when everything was working and OpenAI had an unassailable lead over its competitors.

Right now, everything's—I mean, we've got Sora. Why the hell did we buy io, as great as it seemed 6 months ago? Hardware seemed like a great deal for $6 billion when I could get the legendary Jony Ive, but his team stole everything. What's the point? This is just an effing distraction for something that's going to hemorrhage cash.

Jason Calacanis

I'm all in. If we go back 60 weeks ago, I'm all in on persistent AI around us. Zuck, with his new eyewear and all that, is onto something. But this has got to be almost on the cutting board. It's like the Apple Car. It's got to be close to being cut.

I think the reason this lawsuit may have happened is to push it over the line of getting cut. Just kill it, guys. It's not directly cause and effect, but it really could push it over the line to say, “Let's put a pause on this for a year, boys. Let's put a pause on it for a year.”

I didn't pile in on that initially because I felt I would pile in on this point, but I totally agree. I remember when they did the individual deal, I was like, “This is stupid.”

Sometimes, you know, the cliché is to keep the main thing the main thing. The main thing here is that LLMs are really amazing at code, and the amount of economic value that can be created from that dwarfs everything else.

Right? I've seen the columnist Ben Thompson articulate the view that, to some extent, it may have been a distraction for OpenAI that they were so successful in consumer, right? Because where the value has clearly been created is not just within enterprise, but within enterprise for coding. And I'm with you, Jason: if this was a mercy killing, Apple may have done them a favor. Move on, slow it down, stop doing this.

Jason Calacanis

And just to your point, if you think about a year ago, when OpenAI was a consumer company, I know you thought it didn't make sense, but step back: it actually all makes sense. If you're a consumer company, should we go into hardware? At some point, you do think about it. Should we go into media as a consumer company? Yeah, I mean, we could buy the top podcast for $400-something million. Worst case, it doesn't work.

If you're a pure consumer play, this is how you keep building. You add to this; you add to the layers of the concentric circle, right? An enterprise death-march LLM makes no sense. It's baffling, right? Let someone else build the pin, the pin that clips to your shirt.

2. Zuckerberg Returns to X to Take on OpenAI

You mentioned that one. You mentioned Zuck. Zuck returns to X for Muse Spark 1.1, and it's the first time ever that Meta has started charging developers to use their own models. It was a very meaningful release. Some people think Alexander Wang called the pricing of it, in particular, very aggressive versus OpenAI and Anthropic. As I said, Zuck broke a 3-year silence on X to launch it. How should we think about the release of Spark, and was it a meaningful progression for Meta in the fight against OpenAI and Anthropic?

Jason Calacanis

Yes. I'd like to come back to how it actually works, but let's do a tangent first. I wonder what the internal Zuck algorithm was: how good does something have to be before I, spitting and cursing, put something on the competitor's social media product? Maybe the comms person goes in and says, “I think the new Llama release is awesome. Should we use X?” He's like, “Fuck no,” right? And then finally he's like, “All the developers are on X. If we're going to launch Spark, we just look like idiots.” And then he's like, “Okay, I'll type something on the clown car company.” Right? Anyway, that's fine.

But before you move off that, Alexander Wang has done 27 posts on X to 3 posts on Threads, and it actually got people quite pissed off online.

Jason Calacanis

They probably forgot they had Threads, as I did until the other day.

It's got 400 million users, apparently.

Jason Calacanis

Sort of. Because it's integrated into Instagram, you can't help but see it. It doesn't mean there's any real engagement.

Activity is the word. Engagement. Let's be honest.

Jason Calacanis

We're 3 inveterate Twitter junkies here, guys. We know where the action is, and it ain't on Threads, okay? Let's be honest, people. If there was action on Threads, J. Harry, you'd be there. If there was action on Mars, Harry, you'd be there.

So let's focus on the real point. I thought there were a ton of things that happened here, right? You're right: they shipped Spark 1.1. They moved away from—it's free. As I always remind myself when I forget it, Llama wasn't open source; it was open weight. But they've moved away from that to charging on an API. They've embraced the same business model as the other frontier labs, right?

It's a decent product. I've used it as an end consumer. I compared and contrasted it to Claude, and in some answers it was pretty good. But, of course, the real test isn't the end user; the real test is coding. They reported good benchmarks, but in the past their benchmarks haven't been as good, and we'll talk in a second about benchmarks.

Fundamentally, the big-picture point is that they're in the game, using the same business model as everyone else, entering the marketplace with the low-price product. It's a big hit to competition. They're going toe-to-toe with the balance sheet to afford it, right? Will it be ROI-positive over 5 or 10 years to be the fourth player in the market, selling at that price? I don't know. But if you're OpenAI and Anthropic, you'd have preferred this not to happen.

It seems to me like, look, it's the tweets that's fun. Clearly, they want to win the eval game and the eval-tweet game that's out there. It gets attention; it clearly gets attention, right? Where you are on the evals gets attention. It's very valuable. It's worth going on Twitter for Zuck.

My Captain Obvious view is that everyone's going to be token-maxing if they aren't already. Everyone's going to hit the limits. I finally hit my Claude limits yesterday for the first time on my Max plan, and I had to decide what to do, right? So, as every organization goes from “Burn all the tokens you want, baby” to budgets—however you do it, whether it's automated, whether it's manual, whether it's picking a dropdown—everyone's going to have to have a cheaper model. It may be in your product for cheaper workflows; it may be for coding. So, at the moment, this seems like a battle for the cheap seats.

Jason Calacanis

It's a big battle for the cheap seats, and we'll see what happens. We don't talk about it, but I think Haiku from Anthropic is a pretty damn good product, and in 1 day they could make it better just by adjusting the slider of how much Opus you get in Haiku. I literally use a massive amount of Haiku, and for cheap, simple stuff, it's really good, and it's a tenth of a cent, right?

So there's a battle for this second bucket. How much of the margin is there? How much of the revenue? Who wants to win in it? I don't know. So I think it's interesting, but it appears to be a battle in the frontier bucket, and I think it's a battle in the B-tier bucket, where there's still a lot of volume, but how much margin or revenue is TBD?

I think that's correct. Not quite. But, on the other hand, first, I agree with you, Jason. You're exactly right. The Databricks paper and a couple of other papers that we'll talk about have discussed this concept of buckets and tiers of models, and I like what you're saying, Jason. You're exactly right: when the pressure on spend is out there, every company—the sentence you said that really resonated with me is, “Every company is going to have an internal tiering.” They might all buy it from OpenAI or Anthropic, but every company with a CIO who's half awake is going to have a cheap-token model to hand out to stop this madness. And to some extent, that means there's a slot for that.

A fun, interesting one would be—you guys know that, obviously, Meta did that contract. They're starting to offer their capacity now, just like SpaceX, on a short-term basis. It'll be fun to figure out at what price. It's not clear to me that they would make more money selling at these token prices. They might actually make more money renting their capacity out. So it'll be interesting to think about compute going to the highest bidder. We'll actually have that discussion again when we come to Grok and SpaceX, but this is an aggressively priced product, and good on them. You're right, Jason. It's pushing at the low end.

We'll probably play it by ear. If this is more successful, then they get more capacity. If it's less, then they give it to their competitors. They'll probably play it by ear.

We mentioned the Databricks paper. It got a lot of attention. Can we actually stay on that? Can you summarize what happened and why it's important?

3. The AI Pricing War Is Just Beginning

Jason Calacanis

Some of the stuff is obvious. I read it all, and the big picture at the start was the basic point that cost per token is really not a useful metric. Cost per completed task was the first thing: how much does it cost to get things done, right? One of the ahas is that some of these models that have a low cost per token have reasoning tokens that are more expensive than their basic tokens, and then you don't know how many they're going to use. So there's a whole bunch of things. The first big-picture point was that cost per completed task is really the thing you have to assess.

Then the second point they made, which was interesting, is that you end up with a Pareto curve for different kinds of tasks: which is the best model for this kind? And this is Jason's point. There will be a big-brainiac, frontier set of tasks for which you use them. Once you quantify the cost per task, then the second thing you're going to have is different tasks. Different models are more cost-effective at different tasks. And a quite cheap model that just requires infinite reasoning might not be as good as paying up for the more expensive model. So you can imagine this Pareto curve of value.

The third point, broadly speaking, was the harness you use. In other words, your infrastructure around that massively impacts how efficient it all is.

Databricks. Yeah, they did. Exactly. Yeah, I mean, I'm shocked—exactly. I'm shocked to discover that a company that effectively helps you manage models says that managing models is important. But, to be fair, they produce the goods, Jason, in terms of facts to back it up.

And it makes sense. I think moving away from this is important because every time someone announces a new model, they kind of give you the cost per million input and output tokens. I think the emphasis on cost per completed task—I can totally see that making sense if I'm the CIO. And again, it's in tune with what Jason was talking about.

Jason Calacanis

I like the post this morning from Aaron Katz at ClickHouse, who said that their AI spend is up 60x since February.

Oh yeah, do you see this? It's a post.

Jason Calacanis

Yeah.

4. AI Is Burning Through Token Budgets

And then he said, “AI is fundamentally changing what is expected from a data platform. We can't build the best data platform for AI if we don't deeply understand it. We're investing in building a team that's truly leading in understanding and innovating with AI—60x.”

Jason Calacanis

Well, I mean, once you do anything complicated, it's easy to go up 60x, right? Instead of doing 1 design for 1 page of your website, have Claude Design redesign your entire website in the background. That'll blow your token budget in 5 minutes for the whole month. It'll be gone.

These really complex workflows consume a lot of tokens, and even as they get cheaper, the outputs are getting richer and more and more complicated. So, listen, I don't know how much time you want to spend on this, but I think not only is token maxing gone from niche to interesting, right? Every developer can consume more tokens. The mediocre developers can do it to be performative and pretend they're working, even though none of their PRs are accepted, and the great ones literally can run 10 or 20 agents 24/7.

I mean, if you talk to the best developers now that are AI-pilled—and I hate that term, okay—they are coding 24 hours a day now, okay? They are, because it's addictive. It's like video games, okay? Now I can build 10 features, 100 features. I can do things I couldn't do so quickly.

It's 10:00 at night. Instead of doomscrolling, I'm doomcoding, right? I'm going all night, and I wake up in the morning and I want to check in on that workflow. Instead of checking in on 1 feature, what if it could have built all my features, right? I don't know any—and again, I hate the term—AI developer that couldn't consume even more frontier-level tokens. We just want more, right?

You can do more prime versions of whatever you're working on. You can just do more versions. It saves you time and it's much better. Why do I have to choose A, B, or C? Just build them all. Then build A-prime, B-prime, and C-prime, right?

Let's actually go further. Let's put them all up in production, on staging, or the dev server. Let's see how they all work. You know what? Let's go further. Let's run 1 million test cases against it, right? “Oh, well, that will cost $20,000 in tokens.” Okay, well, I won't. I'll narrow down the—what if I could? It would be better coverage, right?

Any top-tier developer can consume an order of magnitude more tokens than they are now. The ideas are limitless.

Agreed. I want to stay on this one for a while because I think what's interesting about that, Jason, is that there's never been a product—and this is why this is the management challenge—there's never been this product before which, on a totally on-demand basis, can allow an individual worker, with no cost to them, to basically do their work for them and make them look amazing, right? Yet the company is bearing 100% of the cost.

If you're a developer and you're sitting there going, “I could crank this thing for the next 2 weeks myself, or I can press the magic button, and somewhere out there a whole bunch of expense would accrue to my company, but it's not being taxed on me, and it can get me a long way down,” I'm going to do that. If I'm an intern—I mean, summarizing references in our shop this morning—I'm going to get the magic Claude summary, right? And why not?

To be clear, sometimes that's going to be wildly value-accretive to the company because you saved an hour of someone's time that you're paying, you know, $500, and you paid $20. You're ecstatic they did that. But as Jason points out, if there's no governor on that, eventually you're going to hit the moment where you spend $600 to save $500. That's what's going on right now: how do people get to grips with that?

Jason Calacanis

Yeah. Yeah. Let me just give you a simple example, even if it's very basic. I know some folks will make fun of me who watch this, right, or listen to it, but it took me a while. Now I'm all in on Claude Design. It got really good, okay? When it launched, I wasn't sure if it was a toy feature. I love it now. I think it's gotten better.

But as folks said—and it took me a while to get it—it consumes more tokens than anything else I do. It is massive, okay? Yesterday, for my first message, I'm like 5 minutes in and it's like, “You have to wait till 3 p.m., Jason. You have to wait till 3 p.m. to use your credits.” And that's to basically redesign 1 page because it uses so many.

What I would love to do instead is, imagine this: here's my entire website for any product. It's 100 pages, whatever it is. Every night, come up with better versions. Every night, so I'll wake up in the morning and I'll doomclick through your ideas for my website, and I'll pick the 2 or 3 of the 100 that are better.

5. Will AI Replace Designers and Figma?

Think about it: I burned through my tokens in 20 minutes building 1 page. How many tokens would it consume every night to redo 100, right? Help me with the math, Rory. It's a lot. It's orders of magnitude more tokens, right?

So I think literally every product person, every designer could do that and consume 100x the tokens they do now easily, without even blinking.

Does it change your Figma usage?

Jason Calacanis

Listen, clearly, I don't think so. But what I do think—not to go down a rabbit hole—is that if Claude Design keeps investing in it like they have, and we can figure out the massive token consumption, you can get further without a designer, or you can get further with half a designer, or an outsourced designer, or something.

You can build your whole design schema and build it. People are going to make fun of it. You can still smell a Claude Design app. You can see the elements, but you can get pretty damn far. It's pretty smart on the product side about how to do it.

It doesn't mean you don't need a designer, but it may mean that we don't need that crappy designer anymore to get us to a V1, right? There's just no need to have a—

Does that harm Figma? No. I think in the age of AI, every time you start losing the bottom of your market as a company, I'd be nervous. It won't show up in that quarter's numbers because it's the bottom of the market. It's the ankle-biters. Figma has already gone upmarket. Figma's tracking million-dollar deals, right? Who the hell cares about the low-end, 1-seat deal that it lost to Claude Code?

But over time, those ankle-biters move up the leg. They're the piranhas. Then they take over the knees and the thighs, and pretty soon you retain the big customers but you've lost a portion of your funnel, right? If you lose the bottom, you've lost a portion of your funnel.

They don't all wait and say, “You know what? I'll graduate to Figma when I'm bigger.” This is where Salesforce is at risk. People say, “I love Salesforce. Run it headless,” but the next generation of folks may not graduate to Salesforce because they never started there. They started agentically. They may never graduate there.

If I'm a CEO, I'm at risk that folks won't graduate into workflow tools anymore. Even if just 10% or 20% fewer folks graduate into them, it leads to a death spiral of slow growth. That's why, for public companies, the only thing that really matters to me is whether they're growing net-new logos 15% or more a year. If they are, they'll figure it out, right?

If you're growing your net-new customers 15% or even 20% a year, but as soon as it falls, man, you could raise prices every year, but it only takes you so far. So the rambling answer is, it does worry me, but it's not a threat to this sort of enterprise Figma play at all. That's if I were building a startup, say, right at the start.

Rory, what point did you think was more interesting that I didn't mention or glossed over?

Jason Calacanis

No, actually, it wasn't the point they made. It's just things you find yourself puzzling over as you think about the market. I'm just looking at the level of traction on these companies right now, which has been astonishing, right? I mean, the Anthropic explosion to what now might be $50 billion from $9 billion at the start of the year.

I saw the SemiAnalysis breakdown, and, yeah, $2 billion or $3 billion of that is quote-unquote Claude Code. But the real truth is, I think a huge portion of the API part is code as well, right? I've just been wondering: when something's growing that quickly, what could cause the momentum to change?

I think there's really only 2 things. One is mass competition, which we've been talking about. If everybody dive-bombed the price and it went down by a factor of 2 or 3—yeah, 70%—the kind of low-end token-max pricing you're seeing here, is that likely? I don't know. I don't know if that's going to happen.

So then, yeah, if the momentum continues, what's the constraint? One of the things you figure out in all these kinds of very late-stage growth investing is that, in the end, the TAM is the criterion.

What happens is you start to look at the dollars being spent here relative to the size of software engineering, and it's pretty huge. Remember we talked about whether it's going to be 20% of software engineering, right? And, kind of to Jason's point, everyone goes, “Oh, 30 million developers. It's all fine.”

If you go to the BLS, which I did last weekend, and see there were only 1.88 million developers in the United States, of which only 200,000 of them work in software companies.

Another 600,000 work in tech companies—you know, HP, IBM—and a million of them work in literally JPMorgan, BofA, right? That’s a total spend. The median wage is about $140K for that role. It’s a total spend of around $250 billion, right? And if most of this OpenAI and Anthropic enterprise revenue—if a large percentage of that is software—they really are close to 20% already.

They already hit it.

Part of me says, “Oh my God, what’s happening?” Because there haven’t been 20% layoffs for these companies. It’s just—it’s a huge—my point is this: it’s a huge number relative to the total software coding spend in the US.

And I freely admit I don’t have a conclusion from that. I’m looking at it going, “Hmm, right, is it US revenue?” I read the SemiAnalysis stuff. I don’t think it is. I think a lot of it is in the US. Now, there’s some purchasing in the US for use overseas, but it’s just an astonishingly high number relative even to the biggest number you can think of, which is the wages being spent on software engineering. And it’ll be interesting to see how that shakes out in the next 12 to 20 months.

It is a great analysis because it sort of—you know, it is definitely a bear case that, no matter how great these tools are, we may soon hit the ceiling in terms of how much companies, just on an absolute sense, are willing to spend here. There is a ceiling. You can’t spend more revenue than you take in—and that’s if you have 100% gross margins, right? There are just physical limits. Even if you reallocate budget to it, there are just limits.

You’re right, Jason, and I almost don’t like to posit it because one of our rules of thumb is this. I always joke, simplistically put, Newton’s laws of motion are true: things in motion stay in motion. Things growing at 10x year on year might decline to 8x or 6x or 4x, but they don’t come to a grinding stop, right? And the amazing thing is, when you’re growing at 10x, even 6x next year and 4x the year after that is a huge number.

6. The Hidden Cost of AI for Every Business

I mean, if you’re at $50 billion to $60 billion this year in GAAP revenue and you 6x next year, that’s $300 billion, right? That’s more than the total wage spent on software development, right? So you look at it and go, one of 2 things is going to happen in the next year. Either this thing decelerates faster than anything you’ve ever seen decelerate—and I don’t think that’s option A. Option B, to Jason’s point, a whole bunch of CFOs are literally going, “We used to spend $4 million on tech, all of it in salaries, and now we spend $8 million on tech, $4 million of it on salaries. WTF? We’ve got to do something.” Right?

So it’s just, as I say, no conclusion. I’m just thinking about things on July 15th. It’ll be interesting to see. And you’re right, Jason. There are finite limits. I mean, it may well be that the prize for becoming the fastest-growing company in human history is you may hit TAM faster than any other company in recorded human history. And it’ll be interesting to see. You may have hit the limits of how much money there is in the till.

You know, there’s just one other vector, just to add it up for fun. And again, don’t mock me if folks want to. So there’s how much can we spend on engineering, right? And then there’s another math you can do: okay, every piece of software is becoming agentic very, very rapidly. Okay, so what’s software spend? A trillion and a half. You’re better at the numbers.

Plus or minus. Yeah.

A trillion. So I think we’re coming to the rough conclusion that folks may tolerate, say, a 10% gross-margin spend on AI. Let’s just assume that for a minute, right? That’s another $140 billion that can go to Anthropic and friends, right? So that’s significantly more than the math you just did, but it itself has a ceiling, right? It can’t be all of the $1.4 trillion. Ten percent feels about right today. There’s $140 billion. All the software spend is going to put 10% of their topline into tokens.

Agreed. First of all, I totally agree, because you asked that question a while back and I got the answer wrong because I was an idiot, and then I processed it, right? Which was, how much does it cost to run my little sales and marketing agent? And it was way less than you thought, right?

And the reason this is relevant here is this: it turns out that outside of coding, most of the token costs for agentic software outside of coding are manageable. To your point, they’re 10%. They’re not 50%. Right? So you’re right, Jason. If the software industry is doing a trillion in other kinds of non-coding stuff, that’s another $100 billion right there. You’re right. Salesforce could comfortably pay 10% of its revenue for tokens from someone. They’re not going to pay 40% because they only have 22% operating margins.

That’s $100 billion of revenue that’s almost accessible now. And it’s all going to agentic, right? So that—we need someone smarter than me, but that could be a huge—I don’t know what percentage of Anthropic’s revenue that is, but it’s a lot. That $100 billion.

No, you’re right. It’s really 3 buckets. There’s the coding bucket. You’re right. Then there’s the agentic 10% of software tax, just like Amazon took a tax on software 10 years ago. Everyone’s going to run on Amazon; give me 7%. Now you’re right, everyone’s going to run using LLMs; give me 10%. That’s $100 billion more.

And then the last thing you start talking about is, does Claude work replace the knowledge worker at $20, like Microsoft Office? But you’re right. I mean, the big picture—and I’m still wrestling through and trying to look at those numbers. I’ll report back. But you’re definitely at the stage where you’re talking big-ass numbers relative to everyone else’s numbers.

Is there movement into legal, biosciences—not just an appreciation of the fact that they are potentially hitting TAM limits in where they are already?

Jason Calacanis

Potentially—I want to say something: potentially. There’s no evidence that they are, even though logically, again, I want to say it because I don’t want someone to say, “Hey, the hidden TAM…” All the traction and momentum seems to say they’re not. I just look at the actual market and I go, “Huh?” Right?

I personally think, on the medical side, oddly enough, I don’t know if that’s as much as I think the founders of Anthropic—just like the founder of DeepMind—are very much motivated on the medical side by the desire to—everyone almost references the world where cancer is cured when they’re talking about their LLM.

So I think, actually, people’s interest in doing life sciences is 1/3 TAM expansion and 2/3, “I’m putting everyone out of a job. The least I can do is keep them alive. I want to do good because I want to have meaning.” It’s a search for meaning, and, you know, starting with DeepMind and going from there.

I mean, if you don’t think the other founders wish they had a Nobel Prize like DeepMind, like Demis, you’re crazy. That must be good. That must feel good. Jason’s going to get a Nobel Peace Prize for—

Jason Calacanis

Everyone’s not going to get a Nobel Prize with their companies, okay?

But Jason should.

Jason Calacanis

Yeah, exactly. Impact.

I’m waiting just to get an offer—a decent offer, any offer—from Anthropic. I haven’t gotten a single one. No one’s offered to bring me in as a chief anything.

It’s a joke. What would your price be, Jason?

Jason Calacanis

$200 million.

Wow.

Jason Calacanis

After tax, it’s not worth it otherwise. Plus, I don’t really want to work for the man. So that’d be the—if I really wanted to do it, $200 million would be the minimum entry price.

I strongly need to say to you, Jason, that based on my understanding of the Anthropic org structure, you, in fact, don’t work for the man; you work for the woman, right? Daniela runs everything.

It’s a gender-neutral term, I would say, but, for me, take the job and take—yeah, $200 million.

There we go. We have—

Jason Calacanis

I don’t think they actually can risk it. I might not work out for 5 years there. There’s no chance I don’t take it.

There’s no other way of saying this without being hurtful.

It is $200 million. That’s my number.

I don’t think you’ll pass the personality test. I mean, I know I wouldn’t, right? But they’re doing good, caring, all that stuff. I think you just flunk out.

Yeah, but some people think I’m a little driven, Rory, even at this point in my career.

Yeah, you’re driven, but they filter out for cynicism. You just won’t make it through. You’re just not idealistic enough. Look, I haven’t gotten the offer, so empirically, you’re correct. I do want to make sure we—

I’ve done 63 of these podcasts. I’ve written 10,000 blog posts. I haven’t gotten one offer from Anthropic. It’s just not coming, is it?

7. The $26.5 Billion AI Chip IPO Explained

It’s age discrimination. You’re right. The infrastructure that powers foundation models—SK Hynix, Nasdaq listing—you want to go there, Rory?

Jason Calacanis

Great. I mean—

Yeah, great. I’m suggesting topics these days, Jason, because he just scowls and I’m like, “Oh, fuck.”

Jason Calacanis

No, no, no. I thought that—

It’s a winner.

Jason Calacanis

I mean, look, obviously, the stuff everyone knows: there are 3 memory companies. They have been a huge beneficiary of the AI capex boom in the last year. Those stocks—SK Hynix is up 6x, right? 2 of the 3 are based in Korea: Samsung and SK Hynix. Micron is based in Idaho—go America.

And the truth is, it’s an oligopoly. They’ve made out like bandits, and it’s actually very hard to buy stocks in Korea, which is why even today, interestingly enough, after the ADR is traded on the US stock exchange, it’s still trading at a 20% premium to the same stock in Korea. It’s just hard for an American to open a brokerage account in Korea and buy the Korean thing. So I think it’s just good that these guys are accessing the capital markets.

The next Nasdaq listing, largest ever by a foreign company.

Jason Calacanis

It had a great reaction and popped 13% at the close of the day. Then it was back down. I mean, look, it had a very difficult day yesterday and over the last couple of days, right? It would be tempting to go to some kind of, “Oh my God, the AI capex bubble is busting,” and at some point it will, but I actually think there’s a lot of weird technical stuff going on there.

The 2 memory companies, no surprise, dominate the Korean stock market. Samsung and SK Hynix are 60% of the stock market and roughly 120% of GDP in terms of market cap. They’re just huge, right? They’re highly volatile, there’s a lot of retail action in Korea, and there’s a lot of intraday volatility in the Korean stock exchange. It’s all kind of crazy, casino-level wealth there, right?

So even though the stocks have moved down in the last couple of days, it’s not as clear whether it’s fundamentally a DRAM correction or just weird trading. On the DRAM thing, the fundamental bet is—I don’t have an opinion on it—but those 3 companies are all trading at 5 to 8 PEs. In other words, they’re dirt cheap on any kind of screen, and that’s the bull case: this is going to last.

The bear case is that these have traditionally been capital-cyclical businesses. The minute more capacity comes online, prices have gone up 6× in the last year, which is why operating margins have gone from negative in 2023 to around 70%. They’re the most profitable companies on the planet. Samsung is the most profitable company on the planet right now, and those margins are going to go back down.

So that’s the pro and con. I think it’ll last a little longer than the typical cycle, but in the end, it does correct. You’re not going to have the memory makers earning 70% net margins—not gross margins—forever.

One thing—maybe you’ll challenge the connection. I thought it was interesting that IBM this week, a pretty old tech company, right? I think it was founded in the 1900s or something like that. They had a huge miss, and the stock crashed 20%.

Some of it, listen, is probably an excuse. We have to make up excuses as companies when we have a huge miss. But they blamed memory. They said memory is taking so much of the CIO’s budget, and there’s such a scramble to buy memory before it gets even more expensive. They weren’t buying our mainframes, our servers, or anything from us. CIOs went into a panic to buy memory before it would get even more expensive.

I’m sure it’s an excuse, but I’m also sure there’s a germ of truth—that money’s got to come from somewhere, right? IBM said it came out of their business.

Rory O’Driscoll

Yeah, I think you’re right, Jason. I actually don’t think there’s more than a germ of truth. Look, all money has to come from somewhere, right? If you’re spending a lot on LLMs, and then you’re having to spend even more on your memory needs because the LLMs themselves are bidding up the price of memory, you don’t want to be the last item on the purchase-order list on the last day of the quarter, because I’m sure there’s a CFO sitting there going, “What can we cut? We only have a finite CapEx budget this quarter.”

It’s what you said also about what’s going to happen to the modular SaaS vendor. It’s not that they’re going to vibe-code it away. It’s as much as, “You know, guys, we have a tech budget of $10 million. Last year we spent $100,000 on AI, and this year we’re spending $3 million, so we need $2.9 million in savings.”

Yeah.

Rory O’Driscoll

And it turns out we don’t need 5 productivity apps. You’ll love Teams when you get to know it.

And maybe we’ll stick with what we have, too, rather than buy that module from you.

Rory O’Driscoll

Yeah.

8. Has Venture Capital Changed Forever?

So IBM did tumble pretty drastically today. I’m going to go off on a bit of a tangent, Rory, because you said, “Oh, there’s a world outside of AI.” Jason and I were messaging earlier this week about Jason Calacanis’s syndicate or investing activity, where he basically said, “Hey, due to popular demand, I’m moving where we’ll be investing from very early to later-stage growth opportunities in some of the big names.” Is this symbolic of where the venture market is at today, and a sign of the times? Jason, given it was us talking about it, why don’t we start with you?

Jason Calacanis

I just thought it was interesting for him to say that because, actually, you know, Jason Calacanis has some pretty good investments. He does a lot of investments, right? So he’s got some pretty good ones.

I thought the message would be, “We’re adding an annex fund,” right, or, “We’re going to do a little bit more here.” But to say that he has put so much energy, over decades, into creating an alternative path to do a massive number of syndicated investments, and is apparently tilting all into growth, is, in a niche, a huge sign of the times.

Rather than easing into it, completely switching when you’ve got multiple billion-dollar winners that you have material ownership in is, to me, a big deal, right? I mean, seeds for suckers, as we’ve known for 60-something weeks. Other than YC dumping early stage, that would be a real sign, right? YC going on Threads and telling everybody, “We’re only doing growth rounds”—that would rock the world.

But I thought it was at least a sign of the times. Everything’s up, and when everything can exit north of $20 billion, it’s a big deal. Even I have an investment with a $100 million position that took me years and years and years to get to. Then you watch a growth round where someone comes in and invests $100 million, and you think, “Why did I bother?”

All the drama, all the years, all the being the only guy at the board meeting, when someone just drops a $100 million check. If that can grow an order of magnitude, what’s the point of being the guy any earlier than there? No existential point.

Because you love the craft.

Jason Calacanis

I don’t love the craft. Who the hell loves the craft?

David Friedberg. I interviewed him yesterday.

Jason Calacanis

He does. That man loves the craft—he loves the craft so much.

Yeah. Replace me with him on this pod. That man will be doing deals from the old folks’ home, right? And he’ll be getting good ones. I love David.

Jason Calacanis

So do I. Please. That was me defending his love of the craft. I was with him and I was like, “Loves the craft.”

But to give it good framing, I hear you. I think this whole idea of venture early versus venture late—how do you think about that, right? The trite answer I could give, and I always like to lead with the snarky, trite answer, is generally, when something looks easy in investing, that’s precisely the time you shouldn’t do it, and late stage looks very easy right now. That’s the natural contrarian in me.

But I think what’s really going on, Harry, is there are almost 3 different things you have to think about. The first is this big secular trend in private-company financings. As we said, the early-stage business hasn’t changed all that much and hasn’t gotten that much bigger in the last 2 decades. But on top of that, a 5× bigger late-stage business has emerged as companies haven’t gone public as much, and, in the case of something like OpenAI and Anthropic, just became bigger quicker.

There’s been this new business for late stage on top, and there’s no doubt that, looking back 15 years, if you said, “You can enter the one business that’s been around for 30 years and has 200 very good competitors, or you can enter this new business, which didn’t exist before and has relatively few competitors,” all other things being equal, the late-stage opportunity from 2005 onward looked, in retrospect, like the easiest place to play.

So there’s definitely this kind of secular trend, right? And you’re not even getting onto 2 or 3, but you’re shaking your head already, Harry.

No, no, I totally get it. I’ve had a lot of cynical, respectfully older people talk to me about this, and I’m not suggesting that’s you, by the way.

Jason Calacanis

I’m more cynical than you, Harry.

So you are, but it’s okay.

Jason Calacanis

And they’re always like, “Oh, I’ve seen this every cycle. I see this every cycle.” I’m going to push right back in your face, because I actually made a secular comment. I’m going to make a cyclical comment in a second.

I made a secular comment, which is really precise here. What I’m actually agreeing with you, Harry, is that the secular trend is this category didn’t exist, and now it will exist across cycles. This new business—no one in venture when I started, forget 1994, even in 2004, was writing $100 million late-stage checks—and now it’s a thing.

It is a new class of financial product: private $100 million checks in companies already worth $1 billion. It’s a new thing. So I’m not doing the old-guy, “It’s just cyclical.” I’m actually going to do that in a second. But right now, I’m saying, in retrospect, it actually is a whole new category of venture.

It’s not replacing baby venture. It’s adding a whole new category on top, right? And the world needs that category because these companies aren’t public. The Altimeters, the Thrives, the late-stage parts of their businesses—all those guys have replaced what would have been public companies. So that’s kind of not cynical.

Now, the cynical comment I'll make is that, on top of a secular change, you also have a cyclical thing. There's no doubt that the closer you are to the public markets, the more the cyclical thing happens, which is that there are times when the late-stage business looks amazing and there are times when it looks like crap, right? It's really very much dictated by change plus the public markets.

I actually think the secular part is more important than the cyclical part. You just have to be aware of the cyclical stuff. It turns out 2022 was a great time to write checks; 2020, not so much. I mean, the people who did that kind of cheap round and ramp in 2022 are feeling pretty damn smart, right?

So, no, it's just—but there's no doubt that at the height, it gets tough. The thing that I do see is that I don't think we've ever seen the secondary market be as liquid and mature as it is today in terms of the ability to get out of great-name, high-growth companies like we can today. In all of my top names, I can get out of them today with ease, and every single day I have buyers for them. I've never seen such a liquid secondary market.

So the “why now?” is always really important. Outcome scenarios are bigger than ever. The speed of those scenarios happening is faster than ever in a lot of cases. Cursor is a $60 billion exit in 4 years. Along the way, you have the chance to have secondaries that are far more liquid than ever. There are meaningful changes to the ecosystem that make today significantly better than prior cycles.

I agree. It’s worth pointing out, just to be precise, that that's not a “stay private for longer” comment, right? Sometimes people say that because, in fact, these companies haven't been around a long time. You're right. It's a different phenomenon: you're now having companies go from zero to billions of dollars in value creation in 5 years.

9. Why the AI Secondary Market Is Booming

So it's not that they stayed private for longer. I mean, people say, look, SpaceX stayed private a long time. OpenAI hasn't had time to get its systems together yet, right? They're not public because they're staying private for longer. They're not public because—Jesus—5 years ago they were doing nothing, right? So I agree with you. There's this new class of businesses, and there's a small number of winners. If there are 4,000 early-stage companies and only 40 of them matter, and you want to matter but you miss them at the early stage, there's no doubt that putting some money in the later stage is one way to play the game.

You can also move more money. So, yes, completely. I thought it was entirely logical from Jason, and I think it's exactly what he should be doing, to be honest.

Jason Calacanis

I think the third comment, though, is—remember I said there are 3 things? There's the structural stuff, and I would put your comment on new, bigger companies in that structural bucket. There's the cyclical risk, because late-stage business is always about valuation risk, because that's the only risk there is.

But I do think you called it the craft, and that's why I think the odd thing about the private markets, unlike the public markets, where hedge funds can be selling tech stocks today and buying the Indonesian rupiah tomorrow, right? They're all liquid. With private markets, if your comparative advantage—if David Frankel's comparative advantage, which I believe it is—is to be an awesome first-check-in investor, it's not clear how you monetize that by putting money in companies at a $20 billion pre-money valuation, right?

If you think your advantage is your stock-picking, then maybe you can do both. But, for example, it would be insane for Y Combinator to say, “We're giving up early stage and only doing late stage,” because they have a position in that market that's irreplaceable. I do think there is an institutional factor to this. If you've spent a lot of time doing X, it's actually quite hard to say, “Now I'm going to switch to something totally different.”

I mean, good luck, Jason, if you can pull it off, but it's not a layup.

Jason Calacanis

And there's a lot of people for whom they're so good at doing X that trying to do Y would be a mistake, even if Y, on average, is a better return profile.

Agreed. A line to that: Paul Graham sparked some fire on Twitter last night because he posted about a Y Combinator company that apologized for only having 36% month-on-month growth because they were fundraising. He got lots of plaudits, and I thought it was symbolic of the fact that it really depends who says it, because when I say something like that, I get in so much trouble for saying it. Then Paul Graham and Y Combinator say it, and everyone's like, “Yeah, marvelous.” Why?

Jason Calacanis

Well, no one's scared of you, Harry. That's the difference.

Oh, you can't say anything about PG, right?

Jason Calacanis

That's it.

Yeah, you can't. And listen, one of the most successful investors of all time, right? He's wildly successful beyond what anybody realizes, directionally correct earlier than everybody else, and his portfolio is unmatched. He has earned the right for everyone to follow what he says. He's earned it, right? He might be the GOAT. Even though he's not described as the GOAT, he probably is the GOAT, right? Who picks up more 9-figure and 10-figure checks off the floor than Paul Graham? Nobody. He deserves it.

Jason Calacanis

You have to be careful if you disagree. You do have to be a little careful. First of all, I have 2 comments. One is on the tweet. I saw that tweet, and I thought it was more of a tongue-in-cheek comment. I think people are overthinking it.

I saw the tweet and thought it was a cute tweet: “My company apologized because they were only going 36% because they were fundraising.” It was a little bit of a humblebrag. It was kind of funny, but it's like boasting about your kids, not boasting about yourself. It was not worth any emotional energy whatsoever.

So the fact that all these people are commenting on it—they just need to get a life, right?

Jason Calacanis

I think, to your point, this actually gets back to the structural comment. I think Paul Graham at Y Combinator is the GOAT, but it's less because he's a great investor. It's actually better than being a good investor, right? He's put himself in a position where he doesn't have to be an amazing investor. He has a machine and a business that makes him win.

He's been able to delegate the entire task of picking to others. He doesn't pick them. He doesn't sit through 5,000 pitches to hear, right? He had an idea, which is to help make more companies possible, and he manifested that in a business that works. He owns a business, not an investing thing. It would be better to own a business than be a great stock-picker.

It's the best business ever, because you can drive around England going to bookshops in small country towns, sending funny tweets while, at the same time, setting yourself up to pick $100 million checks off the floor 10 years later, when the startups that other people have picked on your behalf go public, because you have a lock on that market. It's a work of genius. It's an N-of-1 business.

Well, every 5 or 7 years, you do have to recruit a Sam Altman or Garry Tan to run it. Other than that, yeah.

Jason Calacanis

Yeah, you have 7 to 10 people at the general-partner level picking deals, and then you pick the best one. It's, again, a work of genius. By the way, the carry premiums that they charge are incredible, too. People don't know or discuss that enough.

Unprecedented

40%–50% in some cases. Forty to 50%. Get my LPs. Leave it. You want access.

Jason Calacanis

Congratulations.

And what I like about it is, to be clear, the original idea was to help startups. I know it's one of those things that I believe to be true: oddly enough, as a person who's very investment-minded, some of the hugest fortunes are made by people whose motivation isn't, “I'm going to be a great investor.” It's someone who's just thinking, “I want to do this thing because the world needs this thing, and it'll be fun, and I obviously want to make some money.”

This is an idea, and it turns out to be a great idea. It turns out 9 people in the first class was only the beginning.

Rory's just like me. It was never about money. It was about the craft of the podcast. Yeah.

Jason Calacanis

Yes, Harry, you can't even get through the sentence without smirking.

Anyway, enough about me. Let's talk about me. Let's flip the mic around again. Literally, I'm nervous we'll have a topic because Rory's face is very real-time.

Jason Calacanis

I'm good.

What about the Phoebe Gates–Phia drama? No? Okay, clearly no.

Jason Calacanis

I didn't like the criticism. I might be wrong, but I didn't like it.

Let's just provide some context. Phia—Phoebe Gates, Bill Gates's daughter, and her co-founder—were in the news very critically for, essentially, and you can correct me if I'm wrong, their product injecting its code to take attribution when it probably shouldn't have been attributed to them in terms of online e-commerce sales. That's the story: taking credit when they shouldn't have done. Jason?

Jason Calacanis

Listen, I could be wrong, okay? But doing basically what everyone in your industry does—doing scammy affiliate marketing when it's basically part of your thing. I mean, startups cut corners, right? It's not Delve-esque, but everyone cuts a corner when everyone in your industry is slightly faking credit for affiliate fees, including Honey after PayPal bought it. It's standard course. It's not that I don't want to be critical.

It’s just—it doesn’t rise to the level of outrage in me. The whole area of affiliate marketing is scammy on top of scammy on top of scammy, with dark patterns, impossible-to-cancel subscriptions, and things placed in your purchase box or checkout that you never intended to buy. How did I buy 7 subscriptions to Outlook 365? Thank you very much, GoDaddy.

I don’t think GoDaddy could exist if it hadn’t perfected getting you to add things when you buy a domain name that you never knew about. I’ve got so many Microsoft 365 subscriptions on GoDaddy domains. I don’t even know how to shut them off, guys. I have to contact someone, but I need my PIN code, and everyone has a different PIN code at GoDaddy.

I don’t remember my PIN code from a domain I bought 11 years ago, but I’m paying $9.95 a month forever for Microsoft 365 for that domain, right? So when GoDaddy does that, I just can’t get that level of outrage. It’s not like Phoebe’s the director of marketing. I mean, she’s running the company, right? The team did it.

I just feel like we’ve got bigger things to be outraged about. They walked it back, and we move on.

I can go with that. Naughty, naughty. Shouldn’t have done it. Stop doing it. Yet another startup slightly misbehaves. I’m kind of with Jason. It’s not ideal. You shouldn’t do it, but move on.

Jason Calacanis

Well, I’ll raise the related question, just for fun. Harry, you can cut me off if you want. For investing and for startups, if everybody does it, is it okay?

10. Why AI Startups Are Worth Billions So Quickly

I’ll give you an example. Almost every GTM AI agentic-GTM provider is buying data sources that they probably shouldn’t be buying. Almost every startup that’s doing GTM is buying from—whether it’s a Chinese data center—they’re taking your data. You can’t be that good with data if you’re getting it the legitimate way.

And every startup is doing it. No, it wasn’t me; I used it through an API, right? But if you’re doing something where LinkedIn’s going to sue you when you get bigger, that’s not cool. Every startup’s doing it. Is it okay if we’re cutting corners, but when we’re bigger, we’ll cut back?

What if everybody’s doing it? What if everybody’s doing affiliate marketing? What if everybody’s using sketchy data sources? In B2B, I’ve always said no to those, but sometimes they get big. I do think, in the era of greed, when the outcomes are so big, it’s hard to tell a founder to say no when it’s standard in the industry.

There are so many examples. I know it’s extreme, but even when you look at Polymarket, that was some sketchy stuff, and pretty good returns on paper, right? To Harry’s point, I probably could sell my Polymarket stock if I had any, right? Or SK Hynix stock. I probably could have sold it. I mean, if everybody’s doing it.

You know, I will confess, when you started this conversation, I was like, “No, you shouldn’t do it,” because I can make the distinction, I think, between—again, for context, folks, it is illegal to scrape LinkedIn, and most people don’t do it, but there are companies that aggregate information from a number of sources, including LinkedIn, and many reputable companies buy from those companies.

So it’s like, I don’t do bad things, but I know a person who’s done bad things and I give him money. That’s the example you’re using, Jason. Part of me wanted to say, “No, you’re wrong,” because I think the cookie stuffing that Honey is alleged to have done is more intentional and more within your own volition than buying from someone who themselves have done something wrong. So they have a little bit of distance.

I could argue the toss with you, but when you started talking, I realized so many breakthroughs rely on a little bit of pushing the boundaries. Uber did it with taxi regulation. Airbnb did it with regulation. You’re right, Polymarket did it in terms of how they’re regulated.

The interesting thing is—and, actually, as another reminder, Anthropic and OpenAI did it in terms of how they trained on information—and generally, my observation is that all those chickens come home to roost. You end up paying the tax. You end up getting caught with it. But if you succeed, it’s all okay.

I’m going to do my trip down history. There’s the Elizabethan quote: “Treason does not succeed, but what’s the reason? If it does succeed, no one calls it treason.” When you succeed as Uber, you just get what you did legalized. When you succeed as Airbnb, you get what you did legalized. So, yes, there are examples of that.

I still think cookie stuffing is a little bit unlikely to get over that line, Jason. But fair point: there are a fair amount of ethical gray areas. Going back to where we started with Apple, there are gray areas, and then there are areas that you step over at your peril. That maybe would be my nuanced comment.

Do you just have a bunch of these quotes percolating in your head?

I’ve got so much noise in my head, Harry. It’s terrifying. But that’s between me and—

Like, you know, a quote on treason being normalized. We call this reading, but your generation’s given up. I don’t read anything longer than a tweet.

Jason Calacanis

I know. Signs on it.

I think we should do the stuff at the end: the Carta stuff, the Constellation Software and ZoomInfo stuff. I think there’s some interesting stuff there.

Jason Calacanis

Okay, let’s totally do it.

Let’s start with the Carta stuff. The top 5% of seed rounds hit a $200 million valuation. Bro, why don’t you take us away there?

Jason Calacanis

Look, I thought Carta was super interesting. They made a comment—you look at seed, and all seed pricing has gone up, but there’s this small percentage that are at a $200 million pre-money valuation. What’s going on?

I think, really, when you break it down—and I think Ben Bratman had a good tweet on this—is that there are a couple of consensus-bet areas where people are willing, and to some extent have to, bet aggressively in terms of dollar size out of the gate. That’s why the prices are high.

I mean, all these neoclouds—if it’s going to take you $300 million to get something done, there’s no point raising $20 million at a $20 million pre-money valuation. You’re just not going to get there. So there are a bunch of those deals where it’s $200 million or $200 million. The dilution may be even more than that. The dilution might be 20% or more, but the check size is $20 million plus, right? And the pre-money is $100 million.

That’s one category: the neoclouds. The other category is the kind of agentic-inference, known-big-market, maybe-a-little-bit-capital-intensive businesses—not quite like building a neocloud, but building out the AI infrastructure—where, again, there’s a belief that there’s a wall of money around the spend there that you can access. People are just willing to write bigger checks.

So, yeah, it totally makes sense. The point Carta was making in the tweet was that it’s not the norm. Ordinary pricing has gone up 10% or 20%, but top-decile pricing has gone up 6x or something like that.

That’s really just a function of there being a class of bets that frankly didn’t exist 5 years ago. There weren’t—I mean, Dylan Patel has that great list of 60-plus neoclouds. Six years ago, there were none except OpenAI, right? That bet didn’t exist. The inference bets are some of the other heavy-dollar bets. Five years ago, people weren’t doing those.

I have, for what it’s worth, a slightly different and maybe less interesting take on it. There’s one thing that’s new here, but the one thing that’s old in this Carta thing is that the large funds, the largest funds—this was like the one and only board I was ever on with John Doerr—when they want to really get ownership, if they have the money, they’ll do it.

In the old days, it was, “We’ll do the seed and the A at the same time. That’s how we’ll get our 20%,” right? It’s the same today. If you can raise a $100 million seed, if you have nothing, you may not need revenue, but if you see an S-tier team and it’s in the right space and you want to hit your number—your 20% ownership—and your fund is in the billions, the math just makes sense.

It’s a bet. It’s just a bet. This has been true since I started investing. That even happens with all the top YC companies: “Okay, sell only 6%, only sell 8%.” But if someone wants to come in and do 2.5 rounds, then you can sell it to Marc and friends, right? That’s how you get the big ownership out of YC: you do 2 rounds.

It’s not new. It’s just that the outcomes have changed it. I don’t know that capital needs are the main driver. I think it’s part of it, but we don’t give startups massive valuations just because they need the capital. Though sometimes you have to, to solve it, right?

I did this in my first startup. I needed $10 million to start, and I got to do it at a $3 million pre-money valuation. It was great. It was great. It was great. Thank you very much. I hated VCs for so many years after, when I gave them a 5.5x return in 12 months and no one showed up to the closing dinner, and they bought 80% of my seed. Thank you very much.

I think it’s just to hit my target when the outcomes can be huge. I don’t think it’s more complicated. If you have a $3 billion or $4 billion fund, the math pencils out.

I think it’s smart to say it’s kind of combining multiple rounds into one. The other way to look at that is the rise of tranched rounds, where you see companies often cited as doing one at $50 million and then the majority come in at $500 million, and it’s—

Jason Calacanis

But even that isn’t new.

11. Can Big Tech Keep Spending at This Pace?

It's just exacerbated by AI. It's not brand new. These things are not brand new, right? What's changed is that they've just become normal. Instead of being reserved for potential outliers, this behavior has become normal for potential outliers, and there are simply more potential outliers.

Jason Calacanis

Yeah. I'm just keying off a comment you made: you don't have to give them a high price just because they need a lot of capital, and that was actually very wise. I was reflecting on that because we used to do, 2 decades ago, a fair amount of biotech, where you knew going in it would take $100 million. But you're exactly right: what would happen is the price would still be low, which was tough on the entrepreneur, because the investors would frequently commit up front and then tranche it.

So they're signing up for the $100 million, the price is low, maybe stepped up for the later one, because they were saying, “I'm taking the risk, so I have to get the ownership.” What's happening in some of these new labs' cases is that the capital need is there, but because the demand from the big funds is greater than the supply of entrepreneurs, and because the perception, rightly or not, is that the outcomes are so huge, the supply-and-demand equation allows the entrepreneur to get a valuation that, 10 years ago, if you were funding a biotech or a semiconductor company where the capital need was $100 million, you would be getting a 20-on-20 in a tranched agreement for $80 million more. You'd be down to 7–10% ownership by the time you did your big round.

I think you're right. At least a good portion of this technology is now doing capital-intensive deals again, and they're doing it on terms that are way more favorable to the entrepreneur than you would have seen in the past. They're doing that for 1 reason: they have the money, and there's a perception that the existing stuff has worked well enough to be worth the risk. We'll see.

I'm not sure if you saw, but a very hot round got announced today: Chai Discovery, $400 million at a $3.8 billion valuation, led by Index, Kleiner Perkins, Sequoia, and Dimension. In terms of dilution, good round for the founders there.

[Speaker?]

Great.

Yeah. Not too late for 10% from 20VC. You can always squeeze that one in on top.

[Speaker?]

No comment.

Not too late. Not too late.

[Speaker?]

No comment.

Not too late. I mean, genuine comment here, and I'll avoid the specifics. One, it's super interesting, but it will be interesting to see how much those companies end up being like an LLM and a language model, and how much they end up being like a biotech company, which is a very different trajectory. We can revisit that another day.

Jason Calacanis

Do you know what's really interesting? It's the blast radius from rounds. What I mean by that is, Isomorphic was incredibly oversubscribed, to the tune of 8–10x. The majority of people did not get anything at all when they wanted huge amounts. That then causes a blast radius. I'm not saying Chai is second-best by any means, but there's a lot of interest in Chai, and then it means there's a lot of interest in another company, Latent Labs in London. It's a blast radius from a very hot deal, which was originally Isomorphic.

It's 1 reason, honestly, sometimes not to announce a deal at all.

Jason Calacanis

I've seen this a lot. I just don't do it. There are more reasons in my career today not to announce, if you can get away with it, not to announce a hot round, right? If you've already got enough going on to attract talent for recruiting, especially if you're not selling directly to tech buyers, to the exec folks, I say don't announce it, man.

Yeah, it will be—

Jason Calacanis

It brings out the daggers, funds your competitors, and creates issues. You just become a target, right? That press release better be worth it because it's like a Dorito or a Pringle: it's gone the next day, right? It better be worth it. If no one's heard of you, do it.

It's also interesting: Thrive in Chai. I'm like, wait a minute. I thought you just led Isomorphic.

Jason Calacanis

Different partners, Harry. Different partners, and we've got paper walls right between them. They're thin. We can hear the other side, but there it is. It is a wall. I think the overlap in terms of competitor risk is fairly low, especially if you're not on the board. So I actually think that's 1 argument for why it's totally fine to invest in both.

Let me make another point, which goes back to Harry's late-stage thing. One of the biggest debates I've had is this: in the early stage, when you're on the board, you can't invest in 2 competitors. In the late stage, structurally, you have to—and we'll talk about that. Not have to, but you should be able to, because if you believe that the late stage is the replacement for what used to be the public markets for high-growth companies, Fidelity Growth would have invested in OpenAI and Anthropic. They're not on the board of either; they want to make the secular bet.

I think the companies that are replacing them in the late-stage private markets are going to do roughly the same thing. Thrive is a master of late-stage investing now. Ironically, you're right: they said that they wouldn't invest in Anthropic when they invested in OpenAI, so it is odd. But generally speaking, I think it's not a problem if you have limited information rights to be invested as a pure late-stage investor or non-board member in multiple companies in the same broad thematic area. It's not our business, but especially—

I will say 1 thing that's changed over the years too, more recently, is—

Jason Calacanis

Monogamous.

Harry probably has even more data. He can challenge me on it, but in my experience, these smaller late-stage investors get nothing from the company. 100%, 100%.

Nothing.

Jason Calacanis

So let's say I'm a founder. You could ask me, “Whoever invested in my competitor is only an 80% competitor to Harry's partner, or 20% or 40%.” This might really bother me at the seed stage. But if I like their brand and I'm going to give them nothing, they're going to learn literally less than a Google search, let alone an Anthropic deep dive. Maybe I just literally don't care. Maybe I even like to usurp that investor from my competitor, right?

In the old days, you asked—and I still think you ask—the founder. I hope you ask the founder. But you can position it in a way that they're getting nothing, right? The founder might like it. You never know. Maybe I got the latter part a little bit backward, but my point is there's just no disclosure now. It's like doing an AngelList investment from the old days. You might not hear from anybody for 5 years. Congratulations, Harry. Here's your 2.5x check. The round on TechCrunch looked like a massive exit. It's 2.5x, but here's your money. Thanks for being a part of it. Thanks for the mention on 20VC, by the way. Very helpful a year ago.

Dude, normally 2x, but yeah, thanks.

Jason Calacanis

I think the point is that there are really 3 categories of investor. There are passive early angels—I don't have an opinion on that; it's not my business. Then there's kind of A and B, where we play, where you're taking a board seat. You absolutely can't do competitive deals. And then there is this new market that, in the last 15 years, has become the late-stage replacement for public companies.

I think you can, because you are replicating the economic terms of what used to be the NASDAQ small-cap marketplace, and there was no prohibition on cross-investing. So it's inefficient to create 1.

We're going to discuss a final topic, which could be in the same realm in some respects. I actually really wanted to do this, which we mentioned last week: Bending Spoons, their M&A, very successful, and the need for the B2B version of that.

Constellation buys TouchBistro—respectfully, I'm sure no one's heard of it—a $70 million ARR Toast competitor that's not growing. They bought a $70 million ARR company for $70 million. How do we think about this? Is this the future of pre-AI SaaS? I just had the Wix founder on the show.

Jason Calacanis

Yeah. $2.1 billion in revenue, $2.1 billion market cap. The TouchBistro thing, first of all, is your destiny if, on top of a slow-growth company, you also put a whole bunch of leverage on it. They did a round, I think, in 2019 at a $600 million pre-money valuation, but then they took money from Francisco Partners in the form of debt. When they missed that, the debt was converted to senior equity.

My guess is the company was running out of money. The senior equity, which had seniority, could put in $100 million, and they could get all the money back in a sale. No one else had any incentive to put money in. So the lesson here is that slow-growth businesses with clean cap tables probably could have done better. They probably could have kept going. Maybe they could have made that $70 million into $140 million. It's never going to be transformative.

But if you have a slow-growth company and, on top of that, you have a big-ass wall of debt, then yes, this could be your future. I read that case a couple of days ago, and I'm like, it's just the real danger of adding a big slug of venture debt on top of a slow-growth business. You just end up trapped. You put your equity at risk, and you end up with a cap table that's very misaligned.

You end up with a bunch of people at the top of the cap table looking out for themselves, doing their job of minding their money. They're like, “I've lent you $100 million. I want my $100 million back.” And 1 level down, they're like, “I've got $200 million in equity in here. Unless I get to $300 million, the first $100 million is senior to me, so I have nothing to play for.” So, yeah, that's the lesson here.

For what it's worth, I thought it was modestly interesting because, structurally, they took on way too much debt. There are going to be a lot more of these stories coming in the coming years. I really don't want to talk to them, but I just don't. I hate debt now.

Jason Calacanis

I hate it. I hate it. I hate it. I hate it. I used to be a fan, but people take on too much. And certainly, don't do debt instead of an equity round.

That's the sucker bet, right? You better be the hottest thing on planet Earth, or it's going to kill you.

It took me a while to see it, but more interestingly—and it's not a huge point, because all the equity holders were basically wiped out right in the conversion, right?—I think it made $2 million on its $100 million. It became a clean look at a stalled unicorn at scale. There weren't egos in it. There weren't 2021 markups to maintain.

There wasn't even founder drama to deal with. What is it worth to just get out of this thing? It was worth 1x. I think it's almost a clean look, because too many of these situations just aren't clean or are obscured, right? It's a clean look at what 8%–6% growth with deteriorating market share at $70 million ARR, where you're not raising cash, is worth. It's worth 1x.

Jason Calacanis

And that's a brutal lesson to founders. If you don't want to be worth 1x, do something before it's too late, man.

The other thing I just look for is: was it still founder-led? Was the CEO still a founder? No—the founder stepped away in March 2021.

Jason Calacanis

I mean, that's part of why it got in that position, right? But it also makes the deal possible, right? If the founders have a veto, 1x may not be exciting when there's nothing—when it doesn't clear the preference stack.

They're made out of being trusted. And I agree, I don't want to pick on them. Stuff happens. I've had deals fail, too. But this is the laundry list of everything that can go wrong: the slow-growth, non-wildly-exciting business, the misaligned cap table. You've already made the CEO change, and you've got no move left.

This is the low end of the outcomes, which segues to the buyer, Constellation. We touched on it, and you mentioned it in the context of Bending Spoons, and we had a good discussion about that last time. You have Bending Spoons buying consumer businesses trading at 12x, and you have Constellation buying B2B businesses trading at 3x.

But as Jason pointed out, we said it last time: there's more disruption in the B2B space, and these pre-AI companies might not have the durability of revenue that Bending Spoons does. I still think the difference between 12x and 3x for 2 rollups is just a little much.

One thought I had since then is that one of the biggest advantages Bending Spoons might have is that there's only 1 buyer of consumer assets, while there are a ton of buyers of B2B assets. So they probably get the pick of the litter, right? There's a way in which there are 100 PE firms that want to buy your shitty B2B company, but no one was touching consumer assets. I think, a little bit, they're getting the premium for being the only player in town.

On the B2B side, the interesting question is this: let me put it back at you, Jason. At 3x run-rate revenues for Constellation, which is a buyer of cash-flow businesses buying pre-AI businesses, do you think that's a good bet, or do you think that's fairly valued, dirt cheap, or going to zero from here? How do you think about that?

Jason Calacanis

Again, I don't want to pick on TouchBistro, but I think in the agentic world they can't be turned around unless you have an incredible new Bending Spoons leader who wants to do it. They have sticky revenue, right? Anything with POS has sticky revenue because it's work, right? But it is terminal in the AI era, right? It's nothing new. This is just a crisp case study.

But do you think—I agree TouchBistro was a given. The interesting thing is, can Constellation, at 3x revenues—it's probably, I don't know, 20. It's probably trading at 20.3, or 12x–13x cash flow. Salesforce is below 10x forecast cash flow. These guys are effectively forecasting that these kinds of companies don't last 10 years.

Jason Calacanis

I don't think they do.

You don't? Interesting. That's a big—

Jason Calacanis

I think when the renewals come, they all just die, like Marketo. They're all going to die.

If you don't have the new revenue, you really are toast.

Jason Calacanis

I think, listen, it's just 1 because I'm close to it for fun. I was also one of the first 10 customers. I'm just watching Marketo, which Adobe bought. I've used it as a case study, but now I'm connected to so many folks there. It is going to zero. It is, but it is one of the slowest decays you can get, right? Moving off a marketing-automation system in the enterprise is very slow, but it is without question going to zero.

There are no net-new logos, everyone hates it, and they just raised prices on us again—20%—while deprecating the API, so we're finally leaving. Salesforce did an LLM lift. It took 1 day to leave. It used to be a year; now it's 1 day. I just think all of these ones are—and I don't know what Constellation does with them if they're in terminal decay. I guess if you have 40% free cash flow, it's just a spreadsheet.

But the terminal decay is accelerating faster than I would have thought when we started this podcast, for sure.

That's the sentence. That's helpful. You're right that it may well be priced correctly, because the fact is all these rollups have terminal decay in the future. But your comment is correct: if the speed of terminal decay accelerates, then you're right—the spreadsheet blows up. Interesting. I'm going to look at those numbers again.

The final one, I think—though it may not be a discussion or a continuation of what we discussed before about Menlo and their fund size—is that Greylock announced that they have raised their newest fund, Greylock 18. I love Greylock 18: a phenomenal firm throughout its vintages, with a really unparalleled track record across so many different years. But a $1.5 billion fund—I thought, similar to the Menlo style, they could have raised much more. It's a very disciplined choice at $1.5 billion, given the size of the partnership.

Jason Calacanis

When was the last fundraise?

Sorry?

Jason Calacanis

When was the last fundraise?

2 and a half years ago.

Jason Calacanis

Oh, I guess that's disciplined, then.

Yeah. No, this is a really disciplined firm, again, with a broad bunch of partners on a capital-per-partner basis.

Jason Calacanis

Well, look, it still would be nice to get into carry mode faster. As great as a mega-fund is, all things being equal, it's a little longer to get those carry checks. To state the obvious, those are different things, right? Are you optimizing for the office, or are you optimizing for the carry check?

Let's be clear: the billion-dollar-plus funds have been able to make it work in this market because there have been places to put the money. Going right back to it, I'm not going to be a skeptic. You can't deny a fact. As Senator Moynihan said, you can have different opinions, but you have to have the same facts. So far, those big-ass funds have worked, given the market we've been in, right?

But Jason's right. All other things being equal, if you know you can raise pretty much indefinitely because you have a track record of 50 years of success, and you don't want to play in the super-late-stage business where you have to write $200 million checks, then slightly smaller funds raised regularly increase the probability of a carry distribution before you have grandchildren.

Yeah. Yeah, I don't know if it's disciplined. I'm sure they have plenty of annex funds and SPVs and SMAs and triple-layered, quadruple-layered sources of capital to tap into, but they say disciplined. Listen, if you can raise infinite capital and you don't have a massive partnership, what's the amount of money I can deploy with my proper reserve models in 18 months and then just raise another fund? Is that disciplined, or is that just how you should run your business?

I mean, 2 and a half to 3 years would be more the standard, but yes.

Jason Calacanis

No, man. I want to know: what can I deploy? I do think you don't make the most money if you wait 3 years to deploy your fund. You make the most money if you can thoughtfully deploy it in 18 months, right? Then get going on the next fund, right?

You're better at the math than me. But yeah, as they say, that word “thoughtfully” is very—as the LLM would say—it's very load-bearing in that sense, Jason.

Jason Calacanis

Well, I guess, just to Harry's point, I don't think that this is—what was the term you used?

Discipline.

Jason Calacanis

Yeah. This is one of these things we manufacture, like how wonderful it is to benchmark as a flat partnership. I mean, who cares, right? It's not disciplined; it's optimized to make the most money.

It's not disciplined. The guys at Greylock didn't sit back and say, “You know, we could make 2 times as much money, 3 times as much money with a fund, but we'd like to make half as much money as our goal for the next fund,” right? It would just be nicer for everybody if we made less money, right?

Jason Calacanis

You can be Andreessen, Lightspeed, General Catalyst, or Thrive, where you are building a multibillion-dollar platform, or you can be a Matrix, a Greylock, or a Menlo, where you are maybe not building that platform to IPO.

The truth is, it goes back to that—remember we talked about the 3 things: the secular, the cyclical, and the crowded. I think you should raise the money that allows you to pursue your business, provided your business is still strategically relevant. You shouldn't raise an amount of money that's dictated by what other people are doing, except to the extent that perceptions change.

And I think that's what they're doing. They're saying, to do a different business, if you decided, “I want to be in that late-stage, multi-platform thing,” then you raise $5 billion. If you want to be predominantly early-stage, but with the ability to follow on rare occasions, then you raise the $1 billion to $2 billion that the Menlos of this world and the Greylocks of this world have done.

More than that, I think the real point to this is that something Jason said, I think, is correct: You use positive words like “discipline,” but the real test should be that, when you're alone as the GP, you should say, “What are we good at? What's going to maximize the probability of us succeeding at what we're good at and minimize the probability of blowing up the franchise by getting it wrong?”

And if that's the meaning of the word discipline, then yes, it's disciplined. It's also, as Jason points out, to some extent, that what is disciplined for the LP should also be long-term maximizing for the GP. It's the Goldman Sachs rule: I want to be long-term greedy, right? And part of being long-term greedy is preserving the franchise.

It's worth noting, just to make a point in passing, that 2 of the names you cited, Greylock and Menlo, both celebrated their 50th anniversary plus, right? They've survived 3 very tough cycles, and some of the other firms haven't really been through that same level of downturn, so there's no doubt that there's a survivor bias in that discipline. And who's to argue with that, boys? As a wrap—