OpenAI 砍掉 Sora，广告业务 ARR 达1亿美元 | Oura 准备上市，Whoop 以100亿美元估值融资

So we start with Anthropic’s monster week.

Jason Lemkin

We may be at the stage where we throw the humans under the bus, not the AI anymore, which I think at some level is pretty terrifying.

We move to OpenAI killing Sora.

Jason Lemkin

I think shooting it in the head is even more significant.

Rory O’Driscoll

A big part of the whole strategic direction of the company was flawed.

Agreed. You’re saying the economists and the accountants have wandered into the room and said, “We have a scarce resource here. Let’s optimize it. Let’s devote this compute to the people who can pay the most for it.”

And then we finish on the man with the biggest balls in tech, Masa.

Jason Lemkin

You haven’t lived until you’ve seen an 85% decline in an index.

Rory O’Driscoll

This is one where it’s just back-ass backwards.

Jason Lemkin

I don’t believe there’s right or wrong in money. There’s just money.

Rory O’Driscoll

I just don’t think raising at $5 billion or $8 billion when you’re at $80 million or $100 million of suspect ARR is the most exciting accomplishment in the world.

Jason Lemkin

Let me be direct: get the fuck over it. You should conform your company around your customers and your model, not your VCs.

Rory O’Driscoll

Being mean to a billionaire is actually a feature.

Ready again? Boys, welcome back. This is “This Week in Anthropic,” otherwise known as the SaaS OGs, which has been renamed.

1. The "Claude Mythos" Leak: 10 Trillion Parameters

I want to start with, you guessed it, Anthropic’s unbelievable 28-day month of February, where they did $6 billion in revenue, which was more than Databricks has done in its entire lifetime. Do you know what I think was the most interesting news out of Anthropic this week? It was actually the accidental leak of Claude Mythos.

Essentially, 3,000 unpublished assets leaked. It’s apparently a 10-trillion-parameter model that is this next-level step change in capabilities, which they’re not releasing because of how powerful it is. This is by far the most interesting to me. Jason, how did you think about this news?

Jason Lemkin

Well, look, obviously it’s embarrassing for Anthropic to leak it. I actually just think we’re going to see more and more of this accelerate. The faster we vibe-code, the faster we ship, the more corners we cut in general on application-level security. It happens.

I mean, so many folks are accidentally uploading code to insecure GitHubs, to databases, to Supabases that are, by default, open. So this is accelerating our data, which is just open on the internet. You could say, “By God, it shouldn’t happen at Anthropic,” and I’m sure someone will get scolded. I’m sure they will. But overall, this is accelerating, and it’s going to accelerate even more as we let our AI agents make decisions.

Our agents are going to decide where to put code; they’re going to decide what level of security to use; and this is going to become happenstance. You know, it’s funny. People are like, “Oh, how could Anthropic have a new security agent and have this happen at the same time?” I think it makes perfect sense.

2. OpenAI Kills "Sora": A Massive Strategic Own Goal?

The Anthropic AI security agents, which I’ve basically used in Replit, are very, very good. And it also makes sense that, as we rush, we’re going to leak source code, data, PII, right? I don’t know whether it’s happened, but it was reported today that all of Mercor’s data leaked. It’s being held hostage—all of it: every single interview, every single piece of PII, every single piece of human data.

We used to mock these things. I think it’s going to start happening daily and weekly in the agentic era. It doesn’t excuse it, but it’s a reality. Agents are goal-seeking, and agents are going to make—not only are they going to make the same mistakes as humans, they’re going to work 1,000 times faster. So even if they make mistakes 10% as often, Rory, help me with the math: if they’re 1,000 times more productive, they’re still going to make 100 times more mistakes. So we’re going to see it everywhere.

Again, just for perspective, because there are 2 things going on here: you have Anthropic, with some data leaked from Anthropic about its new model, Mythos, which is itself meant to be amazingly powerful in dealing with cybersecurity. There was a whole consequence that we’ll talk about in a second in terms of how that impacted cybersecurity stocks.

But as Jason pointed out, the level of irony here is acute because it was an inadvertent leak. So you had a situation where a model that’s meant to be amazing for cybersecurity actually leaked via a cybersecurity leak. We’re toggling between the 2.

On the cybersecurity leak, it was noteworthy that Anthropic blamed “human error.” We may be at the stage where we throw the humans under the bus, not the AI anymore, which I think at some level is pretty terrifying. But you know exactly what happened. You often see this down in the weeds, where you’re about to do a big announcement and you have your content management system. You stage all the assets—be it the Fed press release, or, in the UK, the budget, if you remember, Harry.

You have the press release ready to hit play the minute the budget has ended, and someone inadvertently forgets and puts it on the public side in advance. It’s the same thing here. It probably was human error: there was a whole bunch of content ready for, I don’t know, pick a date, March 15, for the announcement of Mythos; they forgot to secure it correctly, and out it goes.

So that’s the first thing. That’s the embarrassing part of it. Then the interesting part of it—and you really do have to do this without sniggering, despite the fact that it all leaked—is that you also have to separately talk about the fact that there are some big claims on Mythos, right, that Anthropic was making via this leaked memo.

Reminder

no one else has seen it. The actual model is not publicly available. Obviously, some people have seen it, but it’s not publicly available. Even I was trying to get copies of the leaked memo. There are just a few screenshots at this stage, and it’s hard to track it down, at least quickly.

But the statement is that it’s way more powerful. The second thing is that it’s going to be way more expensive for them to serve, and therefore way more expensive for customers to buy. The third thing is a particular focus on cybersecurity. It’s meant to be “extremely good at detecting cyber issues,” and the result of that was a 4% or 5% decline in the average cybersecurity stock last Friday when this leak happened.

Jason Lemkin

Yeah, just 2 other things on the leak, just for this trade-off.

You know, I’m dating myself, but when I was at EchoSign and we were acquired by Adobe, we were an early customer of GitHub. We were putting source code in the cloud, and that was banned at Adobe at the time. It was banned because the source code was their crown jewel.

It was pretty easy to make a crappy PDF reader or a crappy image generator, but to do what Photoshop or Adobe Acrobat did—all the exceptions, all the corners, the thousands, tens of thousands of corner cases—was the crown jewel of the company. And so we got the first exemption to be able to use source code in the cloud.

Pros and cons, but when they used this on-premises source-code-management tool, it took a month to do a release. A month. Okay, now we’re doing 60 releases a day, right? Even Anthropic, the fastest-growing enterprise company of all time, is still doing massive releases every month or 2 and dropping features every day, right?

So we went from something that took 30 days at a tech leader to something that takes hours. There are trade-offs there, and I’ll take them, but we’re going to see it explode.

In terms of the stuff that was published today, going back a few threads on show number 50 to Rory, 1 of the things that I thought was pretty cool in Kairos was 2 things: an always-on background assistant that works constantly. Our AI is working with us 24/7. And agents that can sleep, wake, and self-resume without any prompt.

The autonomous agents—I’ve been talking about how this is going to consume orders of magnitude more tokens and change our lives—I’m excited to see more of them coming. OpenClaw was just this brief thing that woke us up to what Anthropic appears to be all in on: truly autonomous agents running 24/7, hopefully safely, hopefully not leaking all of our source code. But it’s coming soon, right?

Not this whole idea that we’ve been doing since we started this podcast, where you went on to ChatGPT or Claude. No one had heard of Claude when we started this. I was a quirky guy using Claude, and you talked to it and went back the next day.

The next release is going to be on all the time—all the time—debating Harry’s latest investment: Was it big enough? Is he too concentrated in the fund? Where should he go? What was Rory thinking on that deal? Why was Rory abusing Harry by email again for the second time in a day?

I’m excited to see it coming sooner. When our agents are 24/7, they’re literally around us, and we give up all of our personal freedoms and autonomy as part of it.

Are you focused on the embarrassment of it being leaked and the human-error element? While Anthropic has Mythos, which is supposedly as powerful as it is, you’re juxtaposing that with OpenAI screwing around with killing Sora, ads not really working, and people being unhappy with it.

It seems like this massive chasm between the progression of force that is Dario and Anthropic, continuing faster and harder than ever, and a faltering, confused, and dazed OpenAI wandering around the product desert trying to find some water.

Jason Lemkin

You’re just being mean. Again, as I said last week—I’m sorry to repeat myself—the narrative is overdone on both sides. I think some parts of it are true. Obviously, you’re throwing a bunch of different things together.

The decision to shoot Sora in the head is almost certainly a good decision. Look, it’s obviously embarrassing to say something is going to be amazing less than 4 or 5 months ago and then shoot it in the head, but if it’s a mistake, give them credit for at least saying it’s a mistake and moving on, right?

And, yeah, that relationship with Disney—again, I think I’m going to give you—I wasn’t me.

I was sneering at it in real time when it happened. I think someone else on this podcast said it was really significant. Just saying, right?

Rory O’Driscoll

I think it’s massively significant.

I think shooting it in the head is even more significant.

Rory O’Driscoll

I think it’s saying that a big part of the whole strategic direction of the company was flawed.

Agreed. The whole idea that we are going all in on consumer—from what I read, Sora made single-digit millions of revenue, right?—and was consuming $1 million a week, which actually sounds way too low, right? It must have consumed billions and made single-digit millions. It makes no sense as a product, either in the short term or the long term.

If you want to own the whole consumer experience with AI, they decided, “We have to own image and video,” and Anthropic never even attempted to do it, right? So, it’s a massive retreat. It’s probably the right decision, to your point. In fact, it almost certainly is, but, man, our strategy was wrong. This is a huge own goal. Our strategy was wrong.

Rory O’Driscoll

Agreed, but I still think that, as I say, Harry’s overegging it a little, because you made a comment about ads that I think is, quote-unquote, hyperbole, effectively implying that the ad strategy hasn’t worked. That’s a bit of a bigger leap. I mean, Sora hasn’t worked; they’ve killed it. I think I’m with Jason. I think that’s smart, because one of the things you’re seeing right now is that, in a world of scarce compute—and astonishingly, despite all the investments that we’ve seen in terms of actual available compute for people to sell AI on, we’re in a scarcity mode—you don’t devote compute to things that are highly compute-intensive and low-revenue-intensive.

Video generation is extraordinarily compute-intensive, relatively speaking, and the revenue is almost minuscule. Conversely, code gen, while it is compute-intensive, is orders of magnitude less compute-intensive, and there are real dollars attached to it. What’s happening right now? I actually think, at a higher level, it’s very healthy. You’re seeing the economists and the accountants wander into the room, and they’ve said, “We have a scarce resource here. Let’s optimize it. Let’s devote this compute to the people who can pay the most for it.”

3. OpenAI Hits $100M in Ads: Why OpenAI Must Make Ads Work

So, that’s the Sora comment. On the ads comment, Harry, it’s early days for ChatGPT ads, but again, I cite that quote from Brian Kim that I thought was really good. Of course, they’re going to run ads, because there’s no other way to build a mass consumer business, and they’ve got no choice, right? Their consumer conversion rate runs roughly 5%. It gets them to, I think, a roughly $10–15 billion consumer business, right? And that’s out of their 500 million uniques or whatever it is.

One of 2 things has to happen in the consumer business. Again, I’m going to leave the enterprise business out. On the consumer business, either A, they take that conversion rate to a number we’ve never seen before from a typical consumer business. I think that’s unlikely. I don’t think most consumers are going to pay $20 a month for this.

Or option B is you make an ad business work. They’ve got no choice but to make it work. By working, I don’t mean $100 million. People are kind of ragging on the $100 million. It’s in the noise. It’s scale. Big picture here: Facebook and Google each do $200 billion, plus or minus, a year in digital ads. If these guys aren’t doing $20 billion within a couple of years, they’re not even in the game.

And to get to the market cap—I mean, remember, Facebook has a $1.7-whatever-it-is-trillion market cap doing $200 billion. Alphabet/Google has a $2 trillion market cap doing $260 billion-plus. If they’re going to grow into the market cap on the consumer side, $20 billion isn’t enough. They have to do $50 billion, $70 billion of ads.

4. The Manus Scandal: Founders Trapped in China After Meta Deal

So, unlike Sora, this is not going to be, “Try the ads and then fold.” There are only 2 existential bets for this company. One of them is ads to make the consumer business work. The other is, “Oh my God, we should have done coding all along. Let’s get a competitive coding and enterprise model out there and compete with Anthropic on that side.”

Those are the only 2 things they’re doing, and they’re the only 2 things they should be doing. Straightforward. I actually see this as good news. At least we’ve gone from the “let’s wander around the world feeling cool” to “there are only 2 things to do. Let’s get them done.” It’s net-net a positive. Better late than never.

Man, did you see? They had a story in The Wall Street Journal this week about why Dario left OpenAI. Did you read that story?

Rory O’Driscoll

Yeah, I did.

I mean, there was the amount of tension at OpenAI, the fact that Greg Brockman attempted to recruit them, and no one would work for him. He and his sister would not work for Greg Brockman or talk to him. They would not allow him to be part of the LLM or GPT groups.

Then Sam had to constantly tell each of them that they were in charge, right? He told Dario he was the boss, then told Ilya and Greg they could fire him at any time if they wanted to fire Sam, right? Then they begged Dario to come back. Then Dario said he would stay only if he directly reported to the board and nobody else.

I mean, the level of dysfunction—and then firing Sam, bringing him back, then Sora and DALL·E, and we’re not doing coding—it’s just, I mean, I’m exhausted. Maybe I’m wrong. I have to think at least someone like me would feel much more comfortable at Anthropic, where it appears there’s a much more consistent process and leadership, right? Same founders, same thing, same goals. I have to think a company organized like that is just going to out-execute someone with that level of drama. I almost can’t take it.

You’re going to kill me for this, Rory. Is the best thing for OpenAI not to buy Sierra, incorporate that as its customer support product, and have Bret Taylor come in as the day-to-day CEO, with Sam as the fundraiser?

Rory O’Driscoll

I’m not in the boardroom, so I hear you. Look, at the end of the day, Harry, I think you’re right, and I would favor that as a board member, but I’m not going to say that publicly because I don’t want Sam to break my balls. I’m too unimportant for Sam to even give a shit about, right? So, I don’t worry about that at all.

So, let me say this delicately: the amount of board-level and senior-team-level turnover over an extended period of time is probably the highest warning signal that you can have as a board member about how your CEO is doing, right? If it was anything other than a founder-led company and this level of drama was going on, you’d probably sit down with the CEO and ask, “How’s it going, at least, and what are you thinking of doing about this?”

I don’t think you turn on people just when things go to shit, but you probably want to cut down the drama from here, build a team, and try and call a shot and play it for more than 6 months at a time.

When you’ve worked at or observed startups where the CEO is spending so much of their time load-balancing talent that can’t work together versus when you’ve worked at one, or with one, where the talent’s rowing in the same direction, to say that it’s night and day would be an understatement, right? It’s like the backside of Pluto and the front side of Mercury.

And I think Sam—we can criticize him. Actually, when I read everything I’ve seen, and then when I read the Wall Street Journal article, it’s like, “My God, this guy has spent so much time load-balancing the drama of these extremely brilliant personalities.” Oh my God, that can consume most of your time as CEO. Most of your time load-balancing.

Rory O’Driscoll

You’re exactly right. It is the drama of—yeah, we’re not dealing with a bunch of people just trying to crank out some B2B software and make a paycheck. We’re dealing with people who are angsting about whether this is going to change the world, who have fears about the technology, and who have desires to be seen as credited for the technology despite the fears about it.

This is, as is often the case, extraordinarily talented people coming with an extraordinarily high bandwidth and a demand on attention and care and feeding. It’s been a real slog, I say.

5. Masa Son's $40BN Bridge Loan: Investing More Into OpenAI

Okay, the man with the most balls in investing, Masayoshi Son. SoftBank gets a $40 bridge loan to buy OpenAI stock. How deep can Masayoshi go?

Rory O’Driscoll

He’ll go as deep as they let him. I mean, that’s the one thing we know. If they give him another $20 billion, he’ll borrow that, too, right?

I mean, look, this is hot. I checked SoftBank. You’ve got SoftBank Holdings—I have to be careful here. There’s the telco group, which is reasonably levered at the Japan level, and then SoftBank Group is around 2× levered, right? 1.5–2× levered in terms of equity, right? What that means is a 30–40% decline wipes them out. It’s a very aggressive stance.

It would be like me taking our $800 venture—no, no, no, $1 venture fund, borrowing $1.8 billion, and investing it all. If it works, I really juice my return, but if it goes wrong by 30%, I’m done, right? And it’s super aggressive.

I suppose his lesson is that Masayoshi survived 2002, when, I remind everyone, the Nasdaq went down 85%. You haven’t lived until you’ve seen an 85% decline in an index, right? Obviously, if that happened or anything like it, you’d just be way underwater.

Right? So, it's a fairly high amount of leverage for an investment fund, to say the least.

Rory O'Driscoll

Yeah, it's dramatic. Having said that, real estate investment funds get the maximum leverage they can by design, right? That is how they work. I would imagine that if venture had access to more debt, we'd all load up on it.

If we could do the growth rounds in your hottest company, maybe we would, and we could get all the carry from it. The worst thing is we leave the keys to Fund 7 on the table. We might load up, too. I'm not sure, but certainly real estate funds load up as much as they can.

But just pushing back again, real estate funds load up because the cash flows are predictable.

Rory O'Driscoll

Well, they can load up because the cash flows are predictable.

Agreed.

Rory O'Driscoll

Right? We don't. It's just harder for my little fund to go to Silicon Valley Bank and borrow $20 million. In terms of risk, I would argue the SoftBank portfolio—not the telecom company at the subsidiary level, but the SoftBank portfolio—is more like Jason's fund than it is a real estate fund. So, I think it's a high level of risk.

Well, plus, what did he lose on WeWork? $12 billion? He knows what it's like.

6. Cybersecurity Stocks Tank: Is the Anthropic Panic Justified?

Rory O'Driscoll

The 2 big assets, from memory, are obviously the OpenAI position and, I think, the Arm position, which I still think is in the holding company. They're amazing, world-class companies, but it's easily imaginable that both of them decline 30%. So, yeah, it's a hell of a way to live.

Speaking of declining 30% and being in the hole, we might have touched on it earlier, but obviously the Mythos leak hammered cyber stocks. CrowdStrike, Palo Alto Networks, and Zscaler were all down 6%; Okta and Netskope were down 7%; Tanium was down 9%. Was this a justified dip, or is this an unjust reaction to Anthropic's news?

I was listening carefully to the names, and there are different aspects of security. For some of them, I can say, “Yeah, maybe that overlaps,” and then for others I go, “That's just a different thing.” When you listen to all the names being thrown out, you say that's just throwing the baby out with the bathwater.

Stepping back, how does Anthropic make security better? At the code-development stage, they can look at code and find security flaws. There are companies that already do something like that—application-security companies—and you could argue that this is a different way of doing it. Maybe some of those companies will be impacted.

What they're not doing, for example, is real-time perimeter defense. They're not running in real time, blocking people like a firewall. Nor are they doing what, for example, Okta does, which is single sign-on and authentication. That's simply not what they do; it's a different thing. The fact that both of those kinds of stocks sold off says it's just a knee-jerk reaction rather than anything thought through.

It will have an impact. If you were doing application security or secure-code review, you're probably going to have to either incorporate how this works into your analysis, or you'll be redundant. Just as GitHub had to roll in competing models and figure out how to adapt, for some of these companies this is really going to matter, and for others it's just a different thing.

7. The Golden Age of Cyber: Why AI Agents are a "Golden Goose" for Security

Stepping back, I think we're in the panicky stage. I think we're at the stage where, because these companies are doing so well, because they're private and no one sees the numbers, and because AI is so sexy and so potentially amazing, anything can cause a panic.

Robinhood was down about 10% because Elon didn't potentially give them the tender and was going straight through E-Trade, and that alone was a massive win for them.

Jason Lemkin

Well, obviously there's a panic in the market, and the question is: is the panic justified? The panic is that this revenue is not durable. That's the panic.

The cybersecurity one is really interesting. In my experience and opinion, this is one where it's just ass-backwards. If you're in the agentic world, this is the golden age of security. The number of security threats and issues is going up orders of magnitude.

Cloud leaking its source code—it doesn't matter. The number of apps is exploding. There are so many mobile apps that the App Store is taking a month to get your app reviewed versus a week. Everything is exploding. These apps are being built by agents, they're being built in unpredictable ways, folks aren't looking at the code, and the pace of features and products being shipped is accelerating while corners are being cut.

This is a golden age for taking any mature category and acknowledging, “Good news for us: there are more threats.” Whether it's application-level or perimeter security, the good news is that threats are exploding.

I'm not a real cybersecurity expert, although I'm doing another investment right now for just this reason. The whole shtick in my lifetime has been, “Look, you've got to constantly buy new products because new threats keep emerging.” There's been a golden goose in cybersecurity that has allowed new entrants to come into a conservative category.

Someone like Wiz will show up and say, “We know how to do this on the web.” People are so terrified of new threats that they'll take the meeting. This should be the golden age for new and existing investors because the threats are terrifying, and you can't stop the rogue engineers who vibe-coded something that accessed your data.

This should benefit everybody. Everyone should be a rocket ship. Everybody building, everybody monetizing GPUs, is a rocket ship. The fact that the market doesn't see it shows we're in a true panic, which makes it hard to predict a bottom. I don't get it. Everyone should be benefiting when you see an explosion in application production and a change in the paradigm.

The change in the paradigm is good for everybody except Windows Defender from 1996. It probably doesn't help that product, or whatever the hell they have. But everyone with engineers should benefit.

Rory O'Driscoll

I broadly agree with Jason. There might be more than Windows Defender 2006 that are impacted. As I say, some of the application-security and code-review products could be impacted. But big picture, Jason's right.

Instead of having people trying to get into your firewall, everyone is now downloading an agent, giving it full root access to their computer, and telling it, “Have a go.” As Jason just pointed out, it's going to work overnight. No one yet knows the exact approach we're going to have to take to defend against agents running within the organization, but everyone 100% understands that this is an emerging mega-threat because of the velocity of adoption times the power of the solution.

So, I agree with Jason. It might not be the old guard that takes advantage of it, but they tend to be put out of the game. One of the things I admire about the security companies—the CrowdStrikes and the Palo Alto Networks of this world—is that they know damn fine that when a new threat emerges and a new solution emerges for that threat, when an earlier winner comes out, you better spend your $300 million or $500 million and swoop up the winner and add it to your product.

I think there'll be a ton of fast acquisitions as agent-security solutions emerge. If people are smart—and I think those 2 companies are extraordinarily smart—they'll be doing acquisitions long before it's “certain,” because you're going to have CIOs coming to you.

One thing worth mentioning is that somebody leaked information from Anthropic. They are masters at selling fear. One of the things they're doing is releasing the Mythos model first to CISOs within companies. It's kind of like, “Oh, it's so scary. We're going to give you this model and give you time to figure out how to use it.”

Of course, part of that time will involve giving $1 million to Anthropic, so it's just great marketing. They're actually leaning into that and saying to the CISOs, “You're going to have to figure this out. This is the new terrifying weapon we've invented. Please give us $1 million and we'll let you defend yourself with it, too.” Great marketing.

But it speaks to how afraid every security CISO should be, given the pace of agentic AI adoption in the enterprise.

Jason Lemkin

The golden age of cyber—I mean, how hard is it to get a meeting? Whoever you are, if you have any established brand and say, “We've got a new agentic product. We're going to help protect you from this,” you're going to get a meeting that afternoon. Right?

Wish I bought them over Figma. That's a depressing chart that I'm looking at.

Rory O'Driscoll

You need to let go, Harry.

It's down 30% in a month, Rory. It's hard to let go after 30% in a month.

Rory O'Driscoll

Okay. No crying in the casino. Move on.

8. Gross vs. Net: The Truth Behind AI Revenue Accounting

I do want to discuss the questionability of the revenue. We've got Anthropic recognizing revenue in a very different way from OpenAI.

And then you also have questionability around Emergent Labs. Is it okay if ARR is kind of questionable in how it’s accounted for? How do we think about that? You can choose which one you want to take.

Jason Lemkin

Let me just—can I just—maybe Rory can dig into it, but I’ll tell you this: one startup I invested in is over $100 million ARR, and I get them—I own just enough to get the investor updates. I’m not on the board. And I get 3 numbers every month, 3 revenue numbers. I don’t know what the hell they are. It’s over $100 million, but the smallest one is ARR.

Now, I’ve been in since seed. I don’t really care. I’m in the money. I don’t have a choice. But I can’t understand this company’s doing great, but I can’t understand—for the life of me, I cannot understand these 3 numbers. There are asterisks and daggers, and there are charts that go everywhere, but they keep going up and to the right. Which I think was on this Emergent thing we could talk about next. I think that’s what some investors said: who cares? But I can’t tell what the hell the difference is, or what an ARR is in 2026.

Rory O'Driscoll

What I always get is pipeline, which is complete bullshit. There’s contracted, and then there’s live. So, first of all, stepping back: to be fair to both Anthropic and OpenAI, they have a very clear and sensible way they define ARR.

What they take is the average of the last 4 weeks, to smooth it out, times 13, because there are 13 4-week periods in a year, which is more sensible than monthly because you have these varying months. So basically, what they’re saying is realized revenue for the last 4 weeks, averaged—the average of the last 4 weeks times 13. Obviously, if it’s the average, it’s times 52. But basically, it’s actual GAAP revenue. What did we bill for the last 4 weeks? The average of the calculated revenue of the last 4 weeks, to take into account all this. That’s the run rate, right?

So it’s actually pretty—it's not committed, to be fair to them. It’s not committed to any of the bullshit kind of higher-level stuff. It’s actual money flowing through the system. Anthropic is roughly at $19 billion, based on that kind of trailing 4-week metric, and OpenAI is around $25 billion.

Now let’s talk about your thing. There was this whole meme that OpenAI reports net on its partner revenue and Anthropic reports gross. What they’re saying there is, if OpenAI sells through Microsoft and Microsoft takes the money off the top, OpenAI only reports the net amount. If Anthropic sells through AWS and they sell $100 worth of revenue, they report the gross amount, and then they give $20 back to Amazon as a cost of sale. So there are 2 different methods for what looks like the same revenue mix, the same revenue approach.

9. The "Vibe Coding" Era: Reselling Tokens and Triple-Counting ARR

Jason Lemkin

I thought you were going to extend that. I thought part of where you were going was to Michael Cannon-Brookes’ point on the show: that a lot of this revenue is getting double- or triple-counted because of how it’s being recognized. And not only does this happen, but then Cursor’s selling it again and recognizing the revenue, right? The same tokens—people keep reselling these tokens again and again and recognizing them as their own ARR. How many times do we get to resell these poor little tokens?

Rory O'Driscoll

I think that’s actually a great point, Jason. I hadn’t got to it. You’re exactly right. No, it’s like—everyone’s got amazing revenue growth because it’s the same little token going. I just pick the same token.

I mean, if we all agree to have essentially 0% gross margins, an infinite number of us can keep reselling tokens to each other, can’t we? This is our new 20VC ScaleSaster Demo Day. We all resell a million tokens to each other in the first week, so everyone in Batch 0001 has $1 million ARR in its first week because we just resold our tokens to each other. So then it’s completely fair. The VCs don’t mind.

Rory O'Driscoll

And you’re exactly right. The sentence that you added in passing is the key one. Until we all have to get profitable, all this can continue. And then, at some point—that’s why I said I think you’re starting to see it—someone’s going to have to say, assuming we want to have a net present value and a cash flow, what’s going on here? And then all this becomes more clear. I didn’t comment on Emergent Labs being the fastest to $100 million—

Sure, Jason—you actually tried it, didn’t you?

Jason Lemkin

I did. Listen, it’s hard for me to know the criticism, right? Some folks in the press in the Indian B2B environment tried to make this some sort of scandal, right? And, in a sense, fair enough. If you go to Emergent Labs—and Emergent Labs is sort of an Indian competitor to Replit and Lovable, which I’ll show you what I learned in a minute—and you go right now to the homepage, they say zero to $100 million, I think, in 8 months. It’s right there. It’s the biggest banner.

In all fairness, if you’re going to put yourself out there, not just as a tweet, but if it’s going to be right there on your website, one would expect 70% to 80% accuracy in that number, ideally higher, right? So if it’s lower than that, I think it’s fair that some daggers came out. But I was curious. I don’t actually know what happened. Is it triple-counting too?

I can tell you one thing that I learned, which I don’t love, and a lot of AI startups do this, so this is not unique to Emergent. Instead of getting you to use the free version, they try to get you to immediately do a free trial that says it’s $0 and $20 a month thereafter.

Now, so many folks do this. It’s not unique to them. It’s probably best practice in most accelerators. But I’m pretty sure that means they recognized all $240 in ARR that first month when you’re paying $0. And they trick you because you don’t even—yeah, you do have to click on the Stripe link, but you almost think you’re just using the free product.

So if I do a $0-a-month product that’s discounted as a marketing cost and I churn after 30 days, does that count as $240 of ARR? I think for a lot of startups it does. That’s a fair criticism. I’m not saying this is what Emergent did, but a lot of startups will instantly recognize that as $240 in ARR. That’s how they rocket if you’re self-serve; otherwise, you can’t get there that quickly, right? So they clearly did that.

I will say what was interesting is I overall think the criticism is probably unfounded because I thought the product was pretty good, much better than Make—like an order of magnitude better than the disaster of Make—because I do have a 6-part test. The first part is the awareness test. So I ask it to redo the saster.ai homepage.

Actually, of all the platforms, it did the best job. It beat all of them—all of the leaders—because I redid this recently. I redid it. And they’re all good at it: Replit, Lovable, v0; they’re all good at it. They all passed the test, but it actually was probably the best, and it passed a bunch of the other tests.

So I’m not going to switch to Emergent Labs, but I would say it’s in the top 10% of vibe-coding apps. That’s pretty good. So that tells me it’s a legit business. They did the work.

And a lot of these are just—the truth is, if you play with a lot of these, even from leaders, Make’s not the only one that’s crappy, okay? Because they’re basically relying on the fact that Claude Code does 90% of the work for you, right? They’re just putting the simplest wrapper around this. And so they did a good job.

But I really didn’t like the way they do the billing. We’d probably have to shoot half our portfolio companies that do PLG AI, because I think it’s a sus practice. I just don’t like tricking you with this $0 for the first month when you think you’re using a free trial, right? That’s the sus part. I don’t love that kind of gray area. But the product’s pretty good.

You know what I didn’t like when it comes to confusing? I was wondering whether to go off on one on this show, and then I thought, “Fuck it. Let’s go off on one. It’s been a long day.” I’m pissed off by these tranche rounds. I see them all the freaking time. The amount of Sequoia rounds where it’s like, “Oh, you know, X raises money from Sequoia at $5 billion.” Trust me, Sequoia got in at $1 billion, but they just club it together and then announce the sum and then the latest valuation.

And it’s just very misleading. The tier 1s get in early. A tier 2 or tier 3 instantly marks it up the same day.

Jason Lemkin

Isn’t it? For years. What’s the difference?

Well, I think crypto—

Jason Lemkin

Andreessen’s crypto fund—you know, essentially 80% off the token. What’s the difference? It’s the same thing, isn’t it? You’re paying for the signal.

Rory O'Driscoll

I think if you break it down, first of all, just so everyone’s on the same page—because, interestingly, neither Claude nor GPT was on the same page and didn’t know what a tranche round was. They gave the old conventional venture tranche round based on performance milestones, from back in the day when we actually ran businesses, right? So they didn’t have a clear understanding of this.

So let’s be clear on the practice here. The practice here is when a company, a hot company, raises a round where there are effectively 2 different prices per share. First, let’s call it a first closing and a second closing, even if they’re at or near contemporaneous, where the first one might be at $250 million pre, and the second one is at $1 billion pre. The highlight and the headline is always at $1 billion pre.

There are 2 impacts of this. First, let’s do the simple one where there’s just a single participant in the round, right? That’s where, if I’m the new investor, I want to pay $600.

Jason Lemkin

The company wants a headline of $1 billion. To win the deal, someone says, “Okay, let me put some money in at $250 million, some money in at $1 billion.” I can do math because I’m paid to do math because I’m an investor. So, I know my overall basis is $600 million. So, I’m getting what I want, and the company’s getting what it wants, which is a headline number of $1 billion, right?

It’s silly, but that’s all that’s happening in that case. That’s the single-participant tranche deal, right? If a company wants a headline, that’s what they get, right? Generally, those things come back to bite you because, by definition, if you’re the company, just as the investor can do math, presumably you can do math.

If you accept that combined deal, you’re implicitly saying, “I know I’m only worth $600 million, but I’d like the optics of $1 billion.” You better be damn sure that your next round you’re at $1.5 billion; otherwise, you’ll have the optics of a down round. And if you’re an optics believer, that’s probably worse than the optic, right?

So, that’s the single-participant version. The much more annoying version that Harry clearly was getting on his high horse about is when you have the same structure, but access to those rounds where the lead investor maybe does all of the $250 million pre-round and only half of the $1 billion round, and then some new investors just get to do the $1 billion round.

So, literally at the same time, the lead investor was investing at $600 million, and the follower investor, the less-marquee investor, is investing in the same asset at $1 billion. I don’t believe there’s right or wrong in money; there’s just money, right? At a minimum, you have to look yourself in the mirror as the other investor and say, “Wow, that’s the price of being cool, right? That’s the price of access. I’m paying 50% more because I just can’t access that deal, right?”

And that feels like a pretty invidious thing to do. Again, try to avoid morality and saying, “Oh, because it would feel shitty.” You really would feel like a loser if you did that. But let’s play it out.

There’s a situation where the lead investor, let’s say it’s Sequoia because everything good and strong should be Sequoia, is admitting it’s only worth $600 million on average, and they’re just doing this fakey transaction. The company’s admitting it’s only worth $600 million on average because they’re taking the money at a blended cost of $600 million.

So, what you’re saying by doing it at $1 billion is either, “I have a lower cost of capital and I’m willing to take a lower return than everyone else,” or the only positive spin you can come up with is, “The company thinks it’s worth $600 million, Sequoia thinks it’s worth $600 million, but I am smart enough, even though I don’t have access, to know that it’s really worth $1 billion, and I should do it at $1 billion, even though I can’t get the $600 million.”

I’m willing to put up with the upfront tax and foolishness because I think 6–12 months from now it’ll be obvious that I bought it at a great price and maybe I’ll look like a genius.

Rory O’Driscoll

Yeah, but we’ve entered an era, though. The meta thing—maybe this wasn’t exactly what you were queuing up, Harry—but it is tough. We’ve entered an era where so many founders are obsessed with headline prices. Obsessed. They’re obsessed coming out of Demo Day. They’re obsessed once they cross $1 billion, which I think should be a moment to take a pause because of the M&A options.

They’re obsessed about driving to $11 billion and $9 billion and one-upping their competition, and the numbers have become a joke to many founders, right? “Joke” is the wrong term. They just don’t think through any of the ramifications of the valuation they’re at, and they don’t care.

I’m not even saying that’s bad. I think burning the bridges is a good way to have a big outcome, but it’s become utterly gamified on many levels, right? It’s just become gamified. These multi-team tranches in a round are just part of gamifying it, right?

It’s been true of YC since I started investing. There was always a cheaper price before Demo Day, a higher price at Demo Day if you’re reasonably hot, and then a 20% or 30% higher price after Demo Day. So, that version has just become institutionalized, and so be it if it’s what the founders want. If they want to gamify it, so be it, right?

I just don’t think raising at $5 billion or $8 billion when you’re at $80 million or $100 million of suspect ARR is the most exciting accomplishment in the world. I’m going to send a few thumb emojis in the email, but that’s about it. That’s about it. They’re all fake anyway. They’re all just bets, right? These are not public companies.

Yeah, it goes back to your point about Lovable and the growth during the 8 months to $100 million. The gamification of the race to $100 million. I’m not choosing Lovable—

Rory O’Driscoll

Listen, I think they built a good product. I’m sure they’ve been overly lambasted because whether it’s $100 million, $80 million, or $60 million, I don’t care. It’s pretty damn good, right? Whatever it is. But if you’re going to do that, you deserve the daggers to come out when it’s not $100 million, right?

10. Oura Going Public & Whoop Raises $500M at $10BN Valuation

One that I thought was fantastic and exciting—I would like to see a potential IPO or two shortly. I thought this was fascinating. It’s been an incredible journey, actually, from Scandinavia, these founders building this business. It’s had a couple of CEO changes. The business is actually in incredible shape.

Whoop announced today that they raised, I think it was, $500 million at $10 billion. Fitness and health data. Do you know what, actually, Rory? Jason’s annoying me again. I don’t know if you remember his predictions, but he predicted, if I’m not wrong, that 2027 would be the year for human healthcare data and longevity.

Rory O’Driscoll

Yes, and it looks like it might even be 2026. The great thing about both stories is that they’re very defensible. I mean, the hardware side is very defensible. This is not an AI-heavy story. They use AI in what they do, but these are fundamentally standalone products with a clear consumer value proposition.

They’re not going to be cloud-coded on Friday. I totally see it, and they clearly have had critical mass in terms of revenues. I think it’s awesome.

Jason Lemkin

I think the interesting thing for these products, obviously, is that they have exploded. They are recurring-revenue products, right? Going back to the topic of ARR, these are recurring-revenue products, right? For the most part, right? Fairly expensive subscriptions. And they’re exciting until, like Peloton, when they aren’t, right?

There’s not a $2,000 cost here, but I’m not being critical. I think that they’re exciting, but there’s also a faddishness. People can switch. So, the RR, the ARR, the pirate R—what multiples do these companies deserve? What it is, I’m not smart enough to know, but the acceleration is a force of nature, right? I’d love to be a seed investor, don’t get me wrong.

I think you can switch, Harry. You’re into fitness. I’m not so much, but I run 360 days a year, 5 miles a day, for 10 years. So, if there’s a better treadmill, a better device, a better thing, I would switch.

And, you know, whatever, you’re fairly fit, Harry. If you wear Oura and you love it, but Whoop is better and you care, you’re going to switch. So, it’s not ServiceNow ARR, right? You’re loyal, but there’s just some disruption.

Look at Peloton. When Peloton blew up, actually, as the world changed, even though people loved Peloton—super-high NPS—the Peloton addicts of 2020 on Zoom loved it. But when the world changed, they just switched. The simple answer to Peloton is they just switched. They just switched.

And Whoop is different from Oura. It could be a Whoop or an Oura, and maybe one is your ankle and it has your AI rock from Jony Ive, and then we’ll switch.

Rory O'Driscoll

Two comments on this. One disclosure: we are lucky enough to have a small investment in Oura through the acquisition of one of our companies. So, I don’t have a ton of information, and I’m not going to breach any confidentialities, but just in an abundance of caution, I’m not going to comment on numbers at all, right? Great products, right?

But to your point, Jason, it’s not ARR like ServiceNow. Let me be direct: get the fuck over it, right? Not every business on the planet has 5-year design-in. If you’re running a bar down the street, every night I can go drink at a different bar. If you’re selling Coca-Cola, every day I can switch to Pepsi.

If you’re running Amazon, as a consumer, every day I can go and search and go on Walmart. Not every business is going to have enduring, long-term lock-in. Obviously, you prefer to have lock-in, right? But there are lots of businesses that have been around for 50 years where every day they have to earn the right for the consumer to go to them.

There’s no doubt in my mind that any kind of consumer hardware-software combination product has some residual asset from the subscription. But then, yeah, every new device has to be awesome. You’re in competition with other awesome products. It turns out capitalism is hard. If you want to make $10 billion in value, you’ve got to deliver value to your consumers.

Jason Lemkin

And I think, for what it's worth, on Peloton, I actually think what really happened to them is a little like the Zoom story: it was demand that would have been wonderful. It would have been the greatest stock ever had that demand been spread out over 5 or 6 years, increasing at 20% a year. We'd be talking about the Peloton compounding machine. Instead, everyone bought the damn thing at the same time. They staffed up to meet that demand, the market was widely saturated, and then the stock went down and it broke the narrative.

So I do agree: there's nothing you can do to make a market bigger than what it is, but I think they got whiplash by virtue of the COVID demand spike followed by the demand falloff.

Rory O'Driscoll

No, I think the meta question—listen, Oura, as Harry said, I guess I called it. These are great markets. They're large markets. They're markets where people will pay relatively high subscription fees for data. I write a lot of checks.

The meta question for venture is, you know, the classic Peter Thiel Zero to One: “Competition is for losers,” is what Dr. Thiel said. “Competition is for losers.” Competition destroys profits; monopolies drive innovation. You want to invest in monopolies. And so that's just my meta anxiety: if these are unmonopolizable markets, are they good ones for venture or not?

Obviously, there are 2 sides to it, but I would feel more comfortable investing in things that become monopolies. I mean, it's a better landing place than investing in brands.

And you can't ascribe the same durability of revenue to this as you can what? Like, as much as I love—

11. Epic Games Layoffs: The Reality of the Attention Economy

Rory O'Driscoll

But on the other hand, you can ascribe super-high growth and you can ascribe big TAM. Look, if there were enough monopolies to do even 1 good monopoly a year, I'd be in, right? And, speaking of the founders, they're about to get the all-time prize because they invested in a space monopoly and 20 years later they're going to cash in their chips, right?

Monopolies are better businesses than competitive markets. But I do think you can still build billions of dollars of value from a high-quality consumer product, right? There were lots of prior examples of that. We all understand the dynamics. I think it's much less competitive—I mean, actually, for what it's worth, I think if you look at consumer products that flame out, like the GoPro, it's much less a competition issue.

It's not like GoPro died because a competitor to GoPro emerged, right? It's that saturation is as big a problem as anything else.

Jason Lemkin

Well, Rory, I might disagree with you. I mean, there was a whole step function in the industry that they got left behind, right?

Rory O'Driscoll

Yeah.

Would you prefer $2 billion in consumer hardware revenue or $2 billion worth of 5-year contracts like Palantir?

Jason Lemkin

Yeah, I'll take the contracts with the 90% gross margin and the 5-year lock-in, please. You're starting with 10. But you've got to give a lot of respect to these—

Maybe the more interesting question, Rory, that you brought up, because so much has changed. This is our 50th show. So much has changed, right? When we started this show, durable public-company revenue, despite a slowdown in the top line, was the gold standard, right? It was the best revenue out there.

Fast-forward to today: Oura going public—do we give a crap what type of revenue it is? Because the durable software stuff is trading lower than the S&P 500. Maybe I'd rather have recurring revenue with a somewhat suspect customer lifetime value because the software value is so low. Maybe I don't care where my revenue comes from, but it used to matter. It used to matter, right?

We'd be in board meetings where you would torture companies so that they would have more ARR and less variable revenue. That seems archaic today.

Rory O'Driscoll

Yeah, and I remember doing that. I remember telling people not to do that because I'm a big believer that you should sell your product the way the customer wants to buy it.

One of the things I hated about venture was when people would say, “Oh, make it all recurring revenue.” And then the fun one that's actually really relevant right now is—remember everyone would say, “Oh, it's a hardware product, but all the value's in the software, so we're really a software company.” And now, hilariously, everyone's going, “Oh, thank God I've got hardware, because hardware's defensible, not software,” right?

A big-picture comment is: you should conform your company around your customers and your model, not your VCs. I agree with you, this kind of pretending it's ARR, but then next year we hate ARR—it's just a total waste of time for entrepreneurs. Things are what they are, and you do best in business if you actually say what they are and just live and die by that.

Most consumer products have high volatility associated with them. You'd better have a damn good R&D function and continue to build great products.

Look, today—this week—they also talked about how, I think, Allbirds was acquired for less than $30 million.

Jason Calacanis

It was acquired—yeah, I was literally about to bring this up, Harry. It was acquired by AMAX for $39 million.

So my question is, if a company like Oura goes public and you see weakness in a quarter, should you dump this thing instantly like Allbirds versus forgive a little bit of weakness in a Salesforce or ServiceNow?

Rory O'Driscoll

Again, I'm going to avoid any specifics. Genuine comment here, right? Because it's not appropriate, but I would say something. Unlike the other 2 guys, I've run a textile manufacturing company 30 years ago.

The technology required to make an Allbirds shoe is not the same as the technology required to make a wearable electronic device that sits on the human finger and measures blood. Either of these kinds of consumer electronic products—they're not a monopoly in the same way NVIDIA is, but it's a pretty rare number of companies that can do that.

They're not GoPro. Put it this way, Jason: I'll name a wearable, you'll name a wearable, and then I'll name a sneaker, and you'll name a sneaker. We'll be done with wearables long before we're done with sneakers, because there are a lot of different sneaker companies.

And, yeah, it turns out sneakers are easier to make than wearables, which are easier to make than NVIDIA GPU chips.

Jason Lemkin

Speaking of whether we care what we actually care about, there were 2 things. I don't know if you guys know this, but I have wonderful partners, and 1 of my partners is much more intelligent than me, which, Rory, you're going to make some form of gag about, but he helps me put together some of the schedules, too.

And he was like, “Whoa, I had no idea about this.” He was like, “Whoa, Epic Games laid off 25%.” I didn't even hear about that.

Rory O'Driscoll

Yeah, and then I heard Marc Andreessen on your last pod sort of laughing about how we all overhired in 2021.

Well, Marc Andreessen was very clear. He thought that we were all using AI as an excuse.

Rory O'Driscoll

Yes.

And that we were all overstaffed by 50%, or at least 75%.

Rory O'Driscoll

Did any of his portfolio companies do that overhiring?

Overhiring. Just logically, it would be 75%, or at least 50%, overstaffed by 50%, or at least 75%. This doesn't make any logical sense, but keep going. Just picking up on the errors here.

Rory O'Driscoll

He's pissed now. He's pissed. I would love to see you do a day of my life.

Jason Calacanis

I would love to give you a [__] break. I feel actually bad. You've been cranking all day.

Rory O'Driscoll

I will give you 2 hours of sleep for 6 hours a day running 2 companies at once, and then you can feel guilty. Now I'm feeling on.

Jason Calacanis

But no.

Rory O'Driscoll

Don't worry. But point being—

Jason Calacanis

But completely under the radar.

Rory O'Driscoll

They didn't try to do an AI [__] story. They basically said, you know, daily active usage of their Fortnite game and their games is down. So your revenue is down, so you take your expenses down.

It struck me as a no-[__] layoff announcement. It's like, you know, we sell less stuff, we have fewer people. It sucks. And, again, I really do try never to be cavalier about people losing their jobs, because everyone that I know has to put food on the table. They're not earning the kind of money we're earning, and now they have to go out and find another job in a shitty job market. It sucks.

But the lesson is—and that's why I respect them—it's like, we're selling less, so we have to do what we have to do to keep the company profitable.

Guys, we keep talking about these layoffs and these big numbers. I mean, it was over 1,000 people laid off in this layoff. A thousand. The numbers are relatively meaningless, and we've had so many of these conversations. What happens to the labor markets?

Rory O'Driscoll

Well, one thing on the Epic thing: the Wall Street Journal did a good article on this one this week on the permanent decline of Hollywood employment. It's permanently in decline. It's in decline because fewer movies and TV shows are being made. TikTok and YouTube are doing it, and it's in permanent decline because every other country provides larger subsidies, right?

And so there's this permanent decline in Hollywood labor. I think entertainment sort of shows us the future. Epic Games is entertainment, too, right? They'll absorb as much AI and technology as they can to adapt, and it's just early. It's just early. They've had to adapt to YouTube. They've had to adapt to social gaming.

Jason Calacanis

And I think we talk about these 1,000 people that they laid off last year or whatever, but I think Epic Games is—I think it's a more interesting view of the future than Block. We talk about folks saying, “I vibe-code a B2B app,” but content’s already been massively disrupted.

And some part of that is, as you pointed out to me when I got it wrong a few episodes back, AI-related in terms of recommendation engines. But I think a lot of it is just a very competitive attention economy. Fortnite was the game everyone talked about; now it’s not. It’s the nature of the gaming industry. So, yes, what does that mean?

Jason Calacanis

It’s the Fortnite circle coming for everybody at the end of the game.

Yeah, yeah, the little circle coming for everybody.

Jason Calacanis

Even Fortnite. Yeah, yeah, the Fortnite circle has come to Fortnite itself; it has surrounded itself. Poor Epic Games is in the middle of its endgame of Fortnite.

The hidden content creator is shooting it out at the very end. It’s coming for all of us. The Fortnite circle’s coming for all of us.

The other one that was relatively, I think, maybe a little bit overlooked is the report of Manus’s founders—Manus, obviously, for context, being bought by Meta recently—being trapped or kept in China. So, just give people context.

Jason Lemkin

Manus was a company originally based in China and had some Chinese investors. Then it redomiciled to Singapore. Benchmark invested, and it was effectively refounded as a US-Singapore company. Meta acquired it. I want to say—and I use the past tense, “acquired,” because my understanding is the transaction has closed and the money has moved. Though, interestingly, neither ChatGPT nor Claude was clear on that, my understanding is that’s what happened.

But now the latest thing is that the Chinese government takes a dim view of this because they don’t want Chinese talent leaching overseas and going to the US and effectively not being Chinese anymore. They feel it as a brain drain. So, they did something that was pretty coercive, in the sense that 2 of the key founders of Manus, I think, were either in China or summoned to China, and they’re no longer able to leave. So, those are the facts.

And yeah, of course you care. I mean, starting from scratch, that sucks. I wish them the best because that’s not a pleasant place to be. I think you had the Jack Ma thing at Alibaba, of effectively going, as it were, under the radar for a few years when you incurred the displeasure of the administration. You also have people who’ve had significantly worse consequences than that.

So, let’s start with the basic: you wish them all the best, right? But to your point, would another deal like this happen to you? I think this whole Singapore-washing thing is over. It’s over. I’ll tell you who did notice. Maybe no one in America spent any time thinking about it, but every Chinese founder who was thinking about doing this is going, “Hmm. Hmm. I don’t know how I feel about this. I don’t know if I can do this deal. I do know that if I do this deal, I am never going home again.”

But I’m with you, Jason. I think all these other Singapore-washing deals are put on pause, or they’re put on reevaluation. The next thing is going to sound harsh: it’s a fairly coercive regime. If your family’s not out of the country, do you have exposure there? I think it just shows that authoritarian governments can take pretty dramatic, drastic steps to impact a citizenry if they want to.

And I agree, Jason, it makes it really hard to imagine doing another one of these deals without being worried about the consequences. Hopefully, they’ll kind of go, “Naughty you. Pay a 50% tax.” California makes it hard to leave, too, but if you pay them 13%, they’ll let you go to Nevada. Hopefully, it turns out to be something like that. And please God it’s not something more coercive. But I agree, Jason; I wouldn’t do another one.

Jason Calacanis

You know, in venture, you take risks. It’s part of the job. We’ve all had deals where there’s some rule, some corner that was cut, and we talked ourselves into, “It’s okay,” right? There’s something weird about this company, but we convince ourselves, by talking to some mediocre lawyer or asking an LLM today, that it’s okay. So, the Singapore-washing must work, right? They’ve moved to Singapore; it’s got to work. And you convince yourself if you talk to a few people, and you take the risk.

It appears to have bounced the right way for Benchmark and friends, right? It appears they’ve gotten their money. But you don’t do the next one, right? There are 242,000 millionaires in Singapore. The majority of the inflow is Chinese. You don’t do the next deal. Maybe other capital does the deal, and that’s fine, right? Capital’s fungible. But eventually, you just can’t do the next one like this. It’s too risky.

What do you do if you’re Meta? Part of the asset you’re acquiring is the team.

Jason Calacanis

$2 billion isn’t a lot for Meta, and they have the product. Yeah, what are you going to do, Harry? What would you recommend?

Well, listen, for Meta, I’ll just say one thing. I only have a tiny bit of information, but it appears to me Manus is running mostly smoothly as an application and a company. Now, I don’t know if the founders are working on it. I certainly feel strongly that when you lose your founders, you lose the heart and soul of your company. But in the short term, I don’t think it’s a big deal for Meta outside of the founders, because it’s running smoothly, right? That’s my view in the short term. It’s not down. The team’s functioning; they’re running. But it’s crazy.

And at the risk of being Pollyanna, but also wanting to assume the best of people, I would hope that the Meta management team and board, to the extent they do have any influence, can help these guys come to an amicable end. If it requires a tax settlement or whatever, you just don’t want to leave people you just acquired in limbo.

At some zoomed-out level, when you listen to the rhetoric on both capitals, you have to realize that trying to tread between these 2 countries is pretty hard right now. We have China hawks in the US government. They obviously have a whole ton of US hawks, or whatever the equivalent is. There’s a real perception of competition. We don’t let them buy the Nvidia chips, et cetera. You’re playing with fire in that thing, and sometimes it bites you.

Jason Calacanis

I just think overall it’s natural, given the outcomes in AI and given the growth, that I think it’s tied to taking the highest levels of risk we’ve also taken because the payoffs seem to be there. When the deal happened, folks kind of thought this was aggressive of Benchmark to ever have done a deal like this. Why are they doing a deal like this? It’s not even very cheap, right? It seems a little crazy. And they’re like, “Well, we’ve never seen anything grow like this, and the team’s incredibly talented,” right? So, they took a little bit of risk, and they made their profit.

We’re all taking more and more risk. Folks now get a week of revenue at a demo day: “I did $1 million my first week. It’s amazing. What about the second week?” “I don’t know.” But as long as it all works out in the aggregate—and I think this is why nobody cares, to Harry’s point—I cared about Manus. I added it to the list. I don’t think anybody cares. We’re all focused on getting $1 million our first week.

12. The Billionaire Tax: Why the Golden Geese are Leaving California

It’s just a good realization that the worst thing that can happen is not just, “Oh, you lose your money.” There are worse outcomes than that.

I mean, speaking about cashing that in and making billions of dollars, Steve Jurvetson has tied his career to Elon very smartly, so that’s not in any negative way in terms of the investments he has plowed in—they’ve doubled, quadrupled, everything in between. He leaves California and buys the most expensive home in Incline Village. And these were Jason’s notes: will anyone with liquidity be left in California? What if California is structurally bankrupt? Well, I mean—

Jason Calacanis

Yeah.

It’s not a great sign when they keep leaving, is it? It’s not a positive. But Rory’s staying, Jason.

Jason Calacanis

I mean, look—first of all, you’re exactly right. All credit to Steve, and all power to him. I’ve known him intermittently for 30 years. He made a brilliant call to align with SpaceX. He’s been on the board of Tesla and SpaceX—Tesla for a while and then came off, obviously, for those who remember back in the day, but SpaceX too.

Yeah, he’s put his money in a compounding machine, and now he’s clearly hit the DPI moment, right? But going back to the thing, the truth is this: that’s why we said last week, high- and ultra-high-net-worth people have a high degree of mobility. Unfortunately, if you put the hammer up too high, they can leave and choose to go across the border to Incline Village and save 13% on any realized gains, plus, as we pointed out, 5% on all gains if this wealth tax passes.

Yeah, at the margin, why wouldn’t you? It’s not like you need to be in California to be a Tesla board member or a SpaceX board member, given they’re down in Texas. So, yeah, the actions have consequences.

What’s interesting also this week is that Washington State did pass its 9.9% state income tax on millionaires. And the governor said the reason—the governor said—is because there were a lot of folks who said, “Don’t do it,” right? Already, Howard Schultz left. He said today, “Well, they just deserve to pay more.”

Jason Calacanis

And that may well be true. It may well be true. I don't want to debate that. This is not political, right? I'm more concerned about the tipping point when we kill golden geese. Washington and California, and to a lesser extent New York, have been the golden geese. Washington said they're going to lose money; they're not going to make money on this.

It appears that most folks who are neutral or on the right have said California will lose money on the billionaire tax. Everyone's laughed, and the tax itself assumed massive amounts from Larry Ellison, who's been gone a half-decade, right? I do worry they're all leaving—everyone who doesn't work at OpenAI and Anthropic. On this show, we've done it 50 times, and I said at the beginning of this that you leave after the Series B. Now I see that used again and again by these folks who are on the right on it. They say all the founders will leave after the Series B, but it may happen by show 100.

Rory O'Driscoll

One of the arguments I make is this: the truth is, articulating the argument to the activist on the other side as, “You're being mean to the billionaires,” is of genuine interest, and being mean to a billionaire is actually a feature, right? But I think the real articulation is this: if you are actually losing revenue that won't be available to California, the marginal dollar in California probably goes into payment for homelessness, payment for young kids, payment for foster homes, and payment for marginal social welfare services that are easy to defund when times are tough, right?

By choosing to obtusely tax without any attention to the ability to collect that money, you've actually reduced the revenue that's available to you, right? That's the argument you have to make to someone on the other side of the table. You have literally chosen something instead of getting, you know, $250 million by picking up $50 million from the Larrys, Sergeys, and Jensens of this world. You went for $200 million, and now you're going to get zero.

What that means in real terms is that somewhere down the line, long after all these changes have been made, somewhere in Sacramento, someone will zero out a line item on the budget. Let me give you a clue: it won't be payments to the teachers, it won't be payments to the firemen; it'll be marginal services to marginal people that your crass stupidity and desire to make a political point has ended up costing them money. That's the only argument that moves the needle, because it's true. You were right, as you were saying, that it will have a net negative return. How do you now feel good?

Rory, if you were Steve, would you have left?

Rory O'Driscoll

From my perspective, I'm just so glad to be in California and so on. I moved around a lot early in my life; I have my friends here, and I've got my life here. At the margin, the whole point of having money is to be able to do what you want, and for 3, 4, 5, or even 13% of your income, do you really want to leave?

I will say that's why you can tax income relatively highly, because it comes all the time and you can't control the timing. Therefore, you have to uproot your whole life for the rest of your life to avoid it, and I don't think it's worth it. So I wouldn't move to avoid income tax.

13. Do VCs Actually Add Value? The Ron Conway vs. Matthew Prince Spat

Conversely, if you have this pending capital event where, literally, in 1 year you're going to sell, quote, all your SpaceX stock and realize a $2 billion gain, and you're going to pay an extra 13% of that in California, which is $260 million, maybe you turn to your wife and say, “Honey, for the next 2 years, why don't we live in Incline Village 165 days? I'll pay for the plane, we'll go back every week, you won't lose contact with anyone, and we will save $260 million.” And you go, “Hmm, that's real coin.”

That's not the life I live, and that's not the situation I'm in, but that's the argument you make. It's not crazy.

That's real coin, baby.

Jason Calacanis

That's real coin.

Is there any story that I haven't hit on, guys, that we should hit on? I just have to bring up the Ron Conway–Matthew Prince one because it was so—I highlighted that one on Twitter; it was just the funniest thing in the world. And, you know, I don't know Ron Conway. Do you want to provide some context?

Jason Lemkin

I know Ron Conway, but he's certainly viewed as one of the Silicon Valley gems, right? He's a seed investor in so many leaders, always out there as an advocate everywhere. He probably could have retired years ago, right? Very founder-centric.

He wrote that he had helped Cloudflare navigate some very significant issues earlier in the day. I think it was on Jack Altman's podcast—yeah, on Uncapped. They asked Matthew Prince, the CEO of Cloudflare, the question. He said, “Well, maybe. I don't remember any of it.”

It wasn't meant to be mean. Matthew can be fairly sharp, as Harry knows these days. The tweet was not mean. He literally just—maybe he couldn't remember getting any help from this beloved VC, and I think it just said so much to me about VCs adding value, but also VCs thinking they add value.

VCs possibly add a modest amount of value, but founders don't really think that modest value is consistent with the bravado of the VC. It just crystallized the whole value-add idea into a single tweet. It wasn't mean. It's just, “I don't remember Ron helping, but maybe he did.”

Rory O'Driscoll

Yeah, Jason, I did laugh at that, and I think it does. I think my big aha, to your point, is that both sides are, to some extent, right: as a VC, we all want to have agency. We all want to feel we help and want to be good people, and you look at it and go, “Hey, I spent some of my time helping the CEO. I feel I helped.”

But from the company's perspective, they're founding a company. They're doing a million things. On 1 or 2 things on a 10-year journey, you helped. You remember that vividly; they're like, “Dude, it just fades into the background of 100 things,” and you know better than me, Jason, what they have to do every day, right?

The truth is this: one of the proofs of this—an interesting way to check it—is that I often read business biographies and business stories of great companies, venture-backed companies, and how they're formed and what happened. And you know what I notice in them? Every single one of them has very little mention of VCs.

If you just read them, you eyeball them and say, “Oh, that's a biography there.” They crop in and come out a couple of times, right? I think that's right because, realistically, in the journey of what's going on, the only significant things we've done—as I said before on the podcast—are put in the money and put in more money when they need it.

We decide to hire or not hire, or fire, the CEO. We agree on the board's strategic direction, and anything after that is, at best, an assist, right? If you read the biographies of businesses, what you generally see is that the only time the VCs come in is on some version of those, right?

It's 5 pages of the journey early on, interspersed with about 200 pages in the first 5 chapters, and by the time they get to the IPO, it doesn't even rise to the level of a thing, right? I was reading the OpenAI biographies—a bunch of them—recently, and that's just the way it is. Microsoft, same thing, right?

The VC can feel those 5 minutes of impact were amazing, and they feel really good about them, and you feel warm and fuzzy. But the only thing founders really remember, for better or worse, is, “Oh my God, our backs were to the wall and no one would put in money, and they put in money.” They remember that.

Jason Lemkin

Sometimes. In my experience, it works.

Rory O'Driscoll

Sometimes they even forget that. But to your point, Jason—

Jason Calacanis

At least half the time they forget that.

Rory O'Driscoll

If they forget that, they're definitely going to forget the time you made that phone call to help them connect with X, Y, Z and that helped them do something, because that's something that happens 100 times a day. You know you won't.

Jason Calacanis

Yeah. Yeah.

Rory O'Driscoll

We're not the stars in the drama. We're bit players who get well paid for our part.

Boys, as always, the most humbling 90 minutes of my week.

Jason Calacanis

I actually think you'll get more. You'll be humbled tomorrow.

I'd be surprised.