OpenAI & SpaceX S1 Drops | Layoffs at Cloudflare & ClickUp | OpenRouter & Polsia Raise Mega Rounds

SPEAKER_00

I can't remember where corporate America was so convinced of the ROI of something as it is right now of AI. So, Anthropic has done as much in Q1 as all of last year, and OpenAI has done 30% of what they've done last year in Q1. Within a couple of quarters, Anthropic will be visibly and obviously ahead. Profitable, growing more quickly and bigger. That's the dominant out of all three vectors. This week, starting off, OpenAI confidentially files their S1. What does this mean for Anthropic's plans to go public? SpaceX then drops theirs, the largest IPO in history. Anthropic then hits $44 billion in ARR and laps OpenAI in revenue. And then NVIDIA, print $81.6 billion in revenue. And the market? Yours. Then, layoffs, ClickUp, and more. It could be the Geocities deal of the AI era. We could look back on this. And listen, I'm on Team Elon, okay? But 100 times trailing sales, we may look back and say these were some good companies. But 100 times, my god. I love the S1. I think it's all madness. I wouldn't buy a share. I just love the optimism. It's SolarCity on steroids, and we're all here for it because we love AI, but it makes no sense. The cliche, the picks and shovels of the agentic revolution is just a good place to be investing. Ready to go?

SPEAKER_00

Boys, I am so excited for this. I have to admit, I was wondering where we start, given the sheer amount of news that came out in the last few days. Jason, you were telling me on email, oh, not that much this week. I'm like, if this isn't much, I'm glad we do the show today and not three, four years ago. Because Jesus, NVIDIA, we got to start with NVIDIA, $81.6 billion revenue quarter, $91 billion Q2 guidance, $80 billion buyback. And after all of that, the stock barely moved. How did you guys analyze NVIDIA's quarter, the stock not moving? Help me understand.

SPEAKER_01

[SPEAKER_02] I mean, let's do the trivial bit. First of all, I think you forgot the most compelling number, right? It's not the $81 billion in revenue, which is pretty damn compelling. It's the 50-something billion in profits, which makes it the most profitable company on the planet because typically, Google clocks in around $100 to $120 billion now a year, right? These guys in a quarter made $50 billion plus, $56 billion of profits. So if that were continued all year and there was some non-recurring stuff, that's $200 billion of profit a year. That actually, to me, was the big compelling number, right? So first, it's not just a great revenue business growing at 80%. It's a wildly profitable operating margin business, right? And then the second thing to your point about what you make, you were noting about the stock on the day. And, you know, honestly, don't look at the stock on the day. The big picture is over the last 12 months, I mean, that's six, nine months, the stock's up 20%. I mean, it's basically, it had a one-off jump whenever, you know, from 23 on when everyone just totally internalized in the space of a year the amount of spend that would take place in AI CapEx. And it, you know, gapped up and it got to $140, $150 about a year, a year and a half ago. And then it's just growing more steadily because even though the news is amazing, what you got to ask yourself, the stock moves not on the total news, but on the delta news. So they beat slightly. It's gone up slightly over six months. But I wouldn't look at the day's movement. I look overall, the big picture comment is a year, a year and a half ago. It was a credible question to ask, is this CapEx thing sustainable? And I remember asking it myself. I even, shock horror, priced Nvidia puts to say at $140, $150, is this thing done? And I decided that would be a done bet. And thank God I did. Because here we are a year later, the company continues to grow at 80%. The stock doesn't grow at 80% because stock markets anticipate. So about a year and a half ago, the stock grew 3, 4x. And now it's growing at 20%. What it's saying here is, this is great. This is amazing. We trade at a mid-20s PE. We think this is really cool, but we're not sure it's going to have 5x in terms of growth from here. So it all makes sense. The market's pretty happy at Nvidia right now.

SPEAKER_01

[SPEAKER_00] Well, maybe just two thoughts for sure. That's a good summary. One is, I think in today's market, not falling after your quarter is a strong sign because the markets are expecting insane growth, right? They're expecting. And so this growth was just what we expected. It's good enough. And I'm not being facetious, especially when Nvidia is 77% of all of our 401ks. It's not just a niche stock. It's not even just us. It's 7% of all Americans' life savings are essentially in Nvidia this morning with concentration. You know, we're all in, you know, S&P 500 and VTI inversion. So everyone's on board with, whether they realize it or not, everyone's on board with Nvidia, right? When Nvidia falls, we all fall with 7% of our life savings. We're all living, we're all AI investors, whether we really realize it or not, but it's so hard. You got to beat, raise and accelerate. That's the mantra today, beat, raise and accelerate. And so even just trading flat, I'm all for it. That's a good quarter.

SPEAKER_01

AI CapEx is going to be about a hundred billion this year. You look at it and go, most of the time, GPUs are 50% of total CapEx, 400 billion. Nvidia has pretty commanding market share, not alone, but that probably says, you know, that Nvidia business, no surprise, is a $300 billion a year run rate. And there it is at 320 billion. When you look at the announcements of the hyperscalers last week, you kind of know what they're going to do this week in Nvidia. It's plus or minus 2%. So we're really dealing with small numbers. And at that point, I think market reaction is in the noise. Jensen Huang said this week, sorry, on AI CapEx infrastructure spend, that we'd reach three to four trillion by 2030. Do you think that is an extrapolation or exaggeration? Or can you feasibly see that happening? 2030.

SPEAKER_01

Let's just do the quick math here. And what does it mean? If we just use the same guideline of 50% of 20, when you say two to three trillion, that would be one and a half trillion of total semiconductor CapEx. That would be, plus or minus 2%. So we're really dealing with small numbers. And at that point, I think market reaction is in the noise. Jensen Wang said this week, sorry, on AI CapEx infrastructure spend, that we'd reach three to four trillion by 2030. Do you think that is an extrapolation over exaggeration? Or can you feasibly see that happening? 2030.

SPEAKER_01

Let's just do the quick math here. And what does it mean? As Harry always says, what does it mean? If we just use the same guideline of 50% of 20, when you say two to three trillion, that would be one and a half trillion of total semiconductor CapEx. That would be, and that's to say at that point, there were only 70% market share because training and all the other stuff. That's roughly a trillion dollars in revenue from 300 now. I don't think the market, I mean, brutal comment. I don't think the market's quite anticipating that to say the least.

SPEAKER_01

Extrapolating the past gets you there. Extrapolating the growth rate of the last three or four years gets you there. But I think the $64 trillion question is, does that growth continue? Or do you start to hit, not technical constraints, obviously, but economic constraints? Is the ROI there economically on the next $2 trillion? To go from roughly a trillion dollars of CapEx right now to $3 trillion, which is what he's saying it'll happen in, it's now shockingly only four more years, there has to be an ROI on that next, on that next $2 trillion. And I think that's the question.

SPEAKER_00

Well, the Uber COO says no, right? The Uber COO says there's, we've reached, we've already, and if they're not talking their own game, even Microsoft allegedly moving off Anthropic to saying it's too expensive to use Opus, which I think is talking your own game. But at least to the markets, they're saying the ROI, the incremental ROI isn't there, to Rory's point. It's not there.

SPEAKER_00

How did you read that? Uber COO saying that they spent four, well, the whole year's Anthropic credits in four months, but they weren't seeing the gains or the productivity or efficiency gains that they... Did he say the second part of that? Just want to be clear. He said it wasn't measurable. He said he thought it was probably there, but it wasn't measurable. The COO. Yeah. Right. And, oh, got it.

SPEAKER_00

What do I think? Well, first of all, a COO is often a bean counter with a different name, okay? So I'm very, I think, listen, I think we all need COOs to scale, right? But sometimes they're not the most creative person in the organization. Sometimes they think they're product centric, but they really just sit in meetings. And so I always take a grain of salt when a COO says, we don't need more engineers, we don't need more product, we don't need any of this, because they're usually not, and I should have looked up his background, they're usually not deep engineering product background. So they often live in a pre-AI world where products are static.

SPEAKER_00

And I would argue Uber, in many ways, is a static product. Uber, Uber Eats and everything, these are not radically different products than four or five years ago. These are clever marketplaces that scale. So the second point I'll say is, I think what actually, what I do think is more interesting from this is, I hate to say that everything's bifurcated, right? But I do think we're going to enter a world where things are more bifurcated. And we will see folks who get more and more gains from AI spend, they will token max forever, right? Often folks well north of a million in revenue per employee, 2 million in revenue per employee will token max until there's no tomorrow. If you're already hyper efficient, you will find more ways to use AI. The folks that are less efficient, that are larger organizations and more traditional,

SPEAKER_00

[SPEAKER_02] I think will, as the year goes on, become more skeptical, especially if prices go up. And maybe prices are, if you look at Anthropic raising prices, right? If you look at Nebius raising prices, as prices go up, the area of experimentation is ending and people will see very different results from that. I think it is the core question of the year. I think we said that last week, right? And as I say, one of the reasons I, and it almost sounds like heresy to raise it, maybe you won't get a return on your marginal dollar of AI. As I say, it feels sometimes when I say it in San Francisco, it's literally wandering around the Vatican saying, is this Pope guy the right guy, right? It's criticism, the core thing. But I think you've got to, when it's the only question that determines the answer to whether you get to 3 trillion in revenue in CapEx or not, you got to ask it. And I was reading, it's funny, over the weekend, actually, our data scientists sent me an archive paper, the price of progress, price performance and the future of AI. It's a recent paper, basically benchmarking exactly this, right? And the summary, no surprise, is, and basically the comment they're making is, don't just look at benchmarks in the abstract, look at benchmarks and pricing at the same time. Because implicitly, it's the word, you have to start thinking about ROI. And obviously, benchmarks get faster. Price per token, price for all those metrics going down amazingly every year. But token consumption going up every year, because as these models do more agentic reasoning, the total compute cost goes up. So the net cost of serving the customer goes up on aggregate, right? And it's a clear concept, and they don't actually come to a definitive conclusion. But it at least starts to say, the trend here is exponentially increasing costs as you move from simple chat interaction all the way to full-on agents, right? So even though the cost per token stories are amazing going down, the actual cost to do something goes up a lot. So as you look at those SWE benchmarks, right? The thing that's here is not just one tent, it might be 100 of the thing that's next level out to the right. So you're taking on, so you can consume vast amounts of token dollars. So you're right, Jason, at some point, and maybe it is what you said, maybe there are businesses that have a lot of white space ahead of them, where they can write a lot of software, and they can get a lot of value. And maybe there are other businesses where you hit a marginal

SPEAKER_00

[SPEAKER_02] amazing going down, the actual cost to do something goes up a lot. So, as you look at those SWE benchmarks, right? The thing that's here is not just one tent, it might be 100 of the thing that's next level out to the right. So you're taking on, so you can consume vast amounts of token dollars. So you're right, Jason, at some point, I mean, and maybe it is what you said, maybe there are businesses that have a lot of white space ahead of them, where they can write a lot of software, and they can get a lot of value. And maybe there are other businesses where you hit a marginal return much quicker. I don't know. I don't have an answer here. But I do know that it's probably the most important question. I mean, if you, because when you're spending 3 million, you can be a bit laissez-faire and say, it's all fine, we're probably getting a return. When you're spending 300 million, someone probably needs to know. The other thing is, I think that Uber's gross profit margins are 39.75%, okay? This isn't NVIDIA, but it's an awfully good business. Yeah. Right? It's an awful. And I think folks like that, they're going to have different perspectives on the ROI of AI. When you're at those 40% margins, it sounds great, but you're very focused on protecting them in most cases, right? We're not all Zuck, right? And so if job number one is protecting the profit margin, you're going to be really skeptical of spending too much on AI, right? And it's a trade-off. And in some ways, folks with margins like this, we just have to take what they say with a grain of salt, not because it's not important. It is extremely important to understand that we can't all spend like a drunken sailor with 2 million of free credits out of YC, right? We can't spend that way. But it also means there's a backwards-looking bias in here. There's a skepticism in the org where, because, I think what also the CEO said, it's there, but we can't measure it yet, right? So if you're Zuck, not literally Zuck, but if you're Zuck-ass, you're going to lean the hell in, right? When the team is telling you it works, you're going to say, let's give it the rest of the year. If you're a skeptic maintaining those profit margins, you're going to be like, okay, this is performative token maxing. Get me out here, right? And maybe there's going to be a case study of DoorDash versus Uber. DoorDash is still founder-led, one of the most aggressive CEOs out there, right? I mean, JFC. I haven't seen DoorDash say, let's use less token, guys. I haven't seen, if I see it from both. Now, interestingly, if I see it from both, from one of the most aggressive founder-led companies out there, Uber's, one of Uber's biggest competitors, then I'll consider myself chastised, right? If Tony says the same, then I'll be like, okay, we have, we got ahead of ourselves for real companies. It's, we're not all, but I haven't heard it from DoorDash.

SPEAKER_00

[SPEAKER_01] Is it not a very simple realization of what is your core business? Uber's core business is moving people, hard logistics, real world, and Facebook's precision of advertising between, consumer and provider or advertiser. And so of course you're willing to spend on one. It is your core business and then physical world. No, that sounds credible and some level is, but I mean, if you think of it really in terms of both of them have lots of software developers and the big picture question is how much leverage does a software developer get, using AI. And I would argue that's much, that should be more similar across different entities. It's not obvious to me that a software developer working at Uber still has a software problem to work on. If AI gives lift across all industries in terms of software development, then he or she should experience the same lift. So no, I mean, what might be true, Harry is to Jason's point, if you start with 90% gross margins and you're ideologically committed to the program anyway, then you might close your eyes. And I think it's what exactly what Jason said, the burden of proof, you don't impose the burden of proof. There's no one at Facebook yet. So that might change saying, are we getting value from our billions of dollars in AI? We just assume we are and keep going, right? I do think in companies where they have the same idea, same software engineers, but where just the way they run the business because of the overall margin structure is more careful, then you might have people asking the same question. I mean, because to be clear, and just to nail your point for good, if what you're saying is you only get AI software productivity lift in industries that themselves are digital, that would be really bad news because that's a significant sub-segment of the tab.

SPEAKER_00

[SPEAKER_02] I'm saying your willingness to spend is much greater when your core business is digital versus a normal business. Then we're saying the same thing. You're like, the more you're ideologically committed to belief, right? The more you spend, but just to put it out there now, I can't remember a year in the internet in the mid-nineties where corporate America was so convinced of the ROI of something as it is right now of AI, right? The sluice gates are open. The money, the money is being authorized. I mean, the proof that the sluice gates are open is you just look at Anthropic going from five billion in gap revenue to 10 billion in gap revenue. No one is stopping. There's not a ton of querying going on right now, right? So there is a willingness to spend.

SPEAKER_00

So therefore, I do think the next shoe to drop will be, okay, that was fun. We spent 10 billion. What did we get? And it's just the way the narrative has to play out.

SPEAKER_02

Well, I think that was a very prescient recognition of Anthropic's revenue growth, but you missed the core number there. I think you missed it. Okay. [SPEAKER_01] Personally, gross margins, gross margins expanded from 38% to 70%. And they're going to have a 559 million operating profit in Q2 projected. [SPEAKER_01] Yeah. Wow. Wow. Predictable. This business is getting better.

SPEAKER_02

But yes. And if you remember, we talked about last year's margins, which I think were 34%, don't quote, right? The prior year, they were negative 60. So the trajectory was really strong, right? In other words, the margin profile went from negative at the gross margin level in 24 to positive 34, right? That's two points. That's a trend. So no surprise when on top of that, [SPEAKER_01] And they're going to have a 559 million operating profit in Q2 projected. [SPEAKER_01] Yeah. Wow. [SPEAKER_01] Wow. Predictable. This business is getting better.

SPEAKER_01

[SPEAKER_02] But yes. And if you remember, we talked about last year's margins, which I think were 34%, don't quote, right? The prior year, they were negative 60. So the trajectory was really strong, right? In other words, the margin profile went from negative at the gross margin level in 24 to positive 34, right? That's two points. That's a trend. So no surprise when on top of that, when your revenue keeps going, so you have more fixed cost recovery, and you got a little bit of a $50, right? And if it was, it's not seven, let's say it was even 50%, right? You have $2.5 billion of cash coming to cover OPEX, right? It's just hard to spend that much money. High growth and improving margins meant profits were inevitable. So I, yeah. Now, will they keep them? Are they about to spend a whole ton with SpaceX? Yes. But I wasn't surprised. We run an internal model just to understand the ecosystem. And it's funny this happened because Anthropic had only talked about getting profits in 27. And I kept running this internal model, which is just growth, gross margins, training costs. And I'm like, it's hard to lose money with this level of growth and any kind of decent margin structure. So I wasn't surprised by it, but obviously impressed.

SPEAKER_02

What one from the profit, some folks say some of it may be partially due to a discount with SpaceX the first couple months, right? So I just put an asterisk and a dagger. It has nothing to do with Rory's trend. It just may mean that the headline metric is not as impressive as it looks. It is great. There is, but the flip side of it is there's a bit of a bear case on Anthropic out of this, which is that Anthropic has benefited from two things versus OpenAI at a minimum, right? It has become a premium product since the end of last year. It is twice the cost of its competitor, right? Especially at the after Four-Six, it started to charge enterprises clearly per token, per use case. They charge them directly prices doubled. It is a premium product. It is twice the price of its competitor. And it, in a sense it lucked out by not going into video, which ended up being the biggest cash sink that OpenAI ever did. Probably Grok abandons it too. But there is an existential—everything is go, go, go at Anthropic, right? Every single light across the room is green. But if the market, if even if DoorDash is, Tony at DoorDash is spend everything right toward Ash compared to Uber said, we make so much more money on AI. We just, we, with engineering savings, we just justify our token savings with our engineering savings. He views the ROI is very clear. Uber says it's not there. Okay. If even half the world says the ROI isn't there for a premium product, Claude's growth is impacted. Either it has to cut its prices, right? To match the competition or it will maintain a premium product like Apple. But won't be able to maintain its market share, right? Just at the end of the day, it's twice as expensive as its competitor and given the pace of this world. How long can it maintain it? I'm not sure it's clear it can maintain it infinitely. And that's just, it doesn't feel like it today, but that might be a hint of the bear cases that the premium cannot be maintained. The demand is there, but the premium may not be able to be maintained.

SPEAKER_02

[SPEAKER_01] But I think right now, even though they are the premium product, the proof is the revenue traction. The revenue traction Q1 to Q2 is so strong. What it means is, maybe this way it's, it would require a change in behavior from where the default switches are now in corporate America for that number to change. Now you're exactly right, Jason, it could happen and it will all be about probably two things. One is the ROI there. We just competing products in the marketplace, competing strongly. And there were obviously we're talking about OpenAI and obviously to some extent, we're talking about Gemini. And the truth is right now, Anthropic stole the show in enterprise. And one of the other two was distracted. And I think that's OpenAI with consumer. And then Google, despite having amazing raw tech, just didn't do a great job productizing. So right now the inertia bet is more of the same and something would actually have to change in terms of competitive dynamics for that to be true. And as I say, I think there can only be one of two things, either the green eye shades come out and start talking about ROI or the competitive dynamics change. Absent one of those two things, the trend is your friend, as they say in trading. Listen, the Uber CEO complaining and the Microsoft stuff, we can take some potshots at it, but also I just think it's showing you the future. More folks are going to come out and say the same by definition, right? Going to happen. So just the bear case is just as that happens, it might seem like the reaction is no more AI guys go back to pulling up your old ID and coding by hand, but more likely is we will just get better at token budgets. And as folks struggle to manage it themselves, that will, that may degrade pricing power for Anthropic as folks just go, I'm just going to use a combination of GPT-4, it's older and some Sonnet and some DeepSeek. Cause that's the way I'm going to spend my budget, right? It just benefit. It's just a bear case. That's all.

SPEAKER_02

[SPEAKER_01] I think you're exactly right. And the reason I think this is going to happen, that's why I was reading that paper is this. At some point, when you go from experimental to, we spent $300 million on tokens. We don't want to change our guidance. We got some lift in efficiency, so we can touch a hot, we don't $100 million of better off operating income. So yay us. But that means we have $200 million of OPEX we didn't expect to have, and we have to terminate a whole bunch of people. At that point, HR comes in, right? And if you think you can do that just on vibes or no ROI, you're fooling yourself, right? So I actually agree with you. I think once AI spend starts to eat big chunks of wages in large companies, you're just going to have to have that discussion. And it's going to become more, more quantitative. It's going to have to be more provable.

SPEAKER_01

hot, we don't $100 million of better off operating income. So yay us. But that means we have $200 million of OPEX we didn't expect to have, and we have to terminate a whole bunch of people. At that point, HR comes in, right? And if you think you can do that just on vibes or no ROI, you're fooling yourself, right? So I actually agree with you. I think once AI spend starts to eat big chunks of wages in large companies, you're just going to have to have that discussion. And it's going to become more quantitative. It's going to have to be more provable, right? Yep. Big, and not just for humane and political reasons, but just because economically it makes sense. I mean, someone should stand there and say, okay, we've never done a layoff in five years. We're about to terminate 10% of our people. Are we sure that we're getting the ROI from this investment? And if the person leading the investment says, I think so, but we haven't really checked. Hopefully a CEO will say, well, before we terminate a thousand people, shouldn't we check, perhaps take another week here, guys?

SPEAKER_01

That's the ultimate symbol though, that the layoffs that we're seeing is ultimately a case of overhiring during COVID, not AI efficiency.

SPEAKER_00

That's the dumbest thing I've ever seen on Twitter is that this is overhiring during COVID. Dumbest take from the smartest people, including Mark Andreessen, who's got 40 IQ points on me. The dumbest take I've seen since our last show, the dumbest take. Well, why? Because every single...

SPEAKER_00

What was COVID? I'm starting to forget COVID. Even 2020. I mean, what is now, Rory, what is natural attrition across your portfolio? 15% a year, 20% a year? 25% a year, right? This is not overhiring. It might be that your worst employees didn't leave. That's a related, but different issue, right? And actually, I think that's part of it is the unreskillable need to go. But this is not overhiring with 20% attrition. I can't, what's 20% times six? How does that compound? Yeah, it's... It doesn't matter. Yeah, exactly. You get to the same place.

SPEAKER_00

You're saying the 16,000 cut from Intuit, the thousands cut from Coinbase, the 800 cut from LinkedIn. You're saying that is not overhiring. That is AI efficiency. I'm saying there is already more natural attrition since 2020 than these numbers. So blaming it directly on overhiring is clickbait. It's clickbait because it's not mathematically true. [SPEAKER_02] I saw a really cool tweet that basically said, it's five years since COVID people. Haven't you heard of performance reviews? That was me. That was you. That was my tweet. I love that.

SPEAKER_01

That was my point. I mean, you have no excuse. You had five years to manage your low performers out and now you're blaming it on overhiring? Hang on. The only thing that could be true, just articulate the case. The only thing that could be true is you hired to this level and you have this human anchoring whereby you perceive you need all these people. So every time they churn, you just rehired because you had myopia and you didn't realize you could get by on much less. And now suddenly AI has opened your eyes. It is possible in mature organizations with monopoly type profits. You see that behavior, but I'm with you, Jason. It's weird to see it across the board. Right?

SPEAKER_01

Yeah. It just can't be true. It's an excuse. And I don't think any, I don't totally get this mea culpa, this Cloudflare mea culpa on Twitter. I don't totally get what the vibe is here. And I do think they're 80% honest and direct, but this bloat thing seems like the excuse. It can't. Jason, can you just explain the Clickup and the Cloudflare just so people have some context?

SPEAKER_01

Just publicly explaining why we're doing 20% layoffs in thousand line thousand, 2000 word tweets. And I just actually don't, I'm not, I'm literally saying I don't understand the PR benefit. And the meta problem I find is even if it's true, they're sort of blaming the employees, or even if they're not, it's perceived to be blaming the employees for not understanding AI, not being able to change, not being able to evolve. So we've got to let you go. Before the Cloudflare and Clickups, they were always like our very best people have to go, please hire them. Here's a Google sheet. There's no one better than Harry, Rory and Jason we've ever hired. It has nothing to do with organizational change or the world, or that there are low performers. It's just a flip of the coin. We had to let somebody go. We flipped a four headed coin and Harry, Rory and Jason came up, but they're just as good as the folks we kept. That was the vibe, even through last year. Now it's like it's ruthless. Cloudflare, Matthew's like Cloudflare is doing great. And we're still walking 20% of the people out the door. And look, it's some messaging to the markets, existing employees and others that I don't totally, I think CEOs feel that they should be direct, right? In a way that everyone was uncomfortable to do through 2024. Everyone was uncomfortable being direct. Everyone had to say everyone's great. We're all great. And I don't know if it's gone too far, but it's very public. I don't need to know. If Cloudflare is laying off 20, 21%. I mean, I know they're a public company, but that's about all I need to know.

SPEAKER_01

It's funny. You're damned if you do damned if you don't, because the Intuit CEO did it. It's not about AI and also got slammed for that. People saying, well, of course it's about AI because look at these deals you did on Tropic you're being competed with. And yeah, it's always about AI one way or the other, isn't it? Well, directly or our growth is slowed because we don't have an agentic product or we're not being renewed by CEOs because they need budget for Anthropic tokens. It's the world isn't static, right? That's the honest message. That would be my tweet, guys. Sorry. The world isn't static. I partially screwed up. This is what I'm doing to try to fix it.

SPEAKER_01

[SPEAKER_02] Well, of course it's about AI because look at these deals you did on Tropic you're being competed with. And yeah, it's always about AI one way or the other, isn't it? Well, directly or our growth is slowed because we don't have an agentic product or we're not being renewed by CEOs because they need budget for anthropic tokens. It's the world isn't static, right? That's the honest message. That would be my tweet guys. Sorry. The world isn't static. I partially screwed up. This is what I'm doing to try to fix it. Blame me.

SPEAKER_01

[SPEAKER_00] There's no easy way to give bad news, right? There's no easy way to lay off a thousand people. So no matter what you say, you're going to get dinged on it, which is what you know, and I think it's been evolving over the last six months, how people have approached it. I liked what Zeb said though, from ClickUp, even though he got hazed more than math. But one of the things I liked that he said, and it was very transparent. It's like, listen, I'm laying off 22% of my company so I can pay a million dollars to my high performers in the age of AI because one it's, well, he didn't really say it's market, but that's one reason you have to do it.

SPEAKER_01

The other point is if a 10X employee, if a 10X engineer is a hundred X engineer or $1 million bookings a year, sales rep is now two and a half million. I got to pay them. I got to pay them. I got to pay what was a three or 400 or $500,000 a year and a million dollar employee because they're delivering that much value. And I've lived it. I mean, we've gone not to be a broke record. We're doing with two and a half people. We're doing what we did with 20 and everyone's get paid more on my little team, right? Everyone should make a million bucks because we don't have the 18 other people's salaries to pay and we're more productive. Right. And I don't know if it's going to happen at the Uber scale, but it's going to happen all across startups. Right. And with $2 million or more per employee in startups, you got to pay, man, you got to pay your high performers. And to be fair, you're consistent. Cause when I did my three months ago, I was doing my four different ways, four different reasons for AI, the layoffs, all of only some of which are AI you added that fifth, which is I'm terminating people to get different people, which effectively is what this guy is doing and saying, I'm terminating people to make room for people that might cost more per head that are just of different skills.

SPEAKER_01

Well, that might be what he's doing, but what Zeb really said is I just want to pay my current high performers more.

SPEAKER_01

Yeah. It's to say I'm making space for people who you can keep more salary dollars to. And the funny thing is guys, this is the counter argument to the last discussion, right? We're actually repeating ourselves in a weird way. We just had a discussion on this. Does the AI, does 20% of spend, does 20% of R and D spend on tokens make sense? And if it does, that's, you know, first, because the first section of the podcast was is, you know, does token spend make sense and has it a positive ROI and is it good for, and thus it's good for anthropic. And then the second section is I'm shocked to discover this layoff. So if you zoom out a million miles, that's the consequence of the first, right? These are actually the same discussions. If there is efficiency gains from AI across R and D and sales and marketing particular, and then on top of that, if there's also product impacts, which I always like to separate internal impact versus product impact. If those things are true, then everyone's going to be doing 20% layoff just because it's 20% more efficient.

SPEAKER_01

My aha is this discussion and the last discussion are in fact, the same discussion. In one case, we're focusing on the winner, which is anthropic. And in one case, we're focusing on the loser, which is employees. Well, I think, yeah, but the top employees, I think what ClickUp share, I mean, we read that when you see them on social media, Anthropic's hiring a social media manager in King's cross for $450,000 a year. And you're like, how could that be? Right. But their revenue per employee is so astronomically high that it all ties. Right. And I think the new paradigm when the three of us met 2 million per employee was what you reach someday, right? That was Apple and others that was the, you didn't even have to think startups never got there. I think 2 million per employee is going to become the new normal as you scale. And that will mean that your high performers can make two to three times what they used to make, just like an anthropic. Because you just don't need as many people and you're going to funnel your compensation to your high performers instead of the high performers making 40 or 50% more than your mid pack. It should be five X.

SPEAKER_01

What you're basically saying is a new set of tools have come to the floor in the last two or three years. And if you know how to use those tools, your productivity relative to someone who doesn't have those tools is so great that we can afford to pay you half a million bucks because you're obviating the need for three people.

SPEAKER_01

[SPEAKER_00] Being an agentic, an agentic expert should be like Excel in 12 or 24 months. Everybody should be able to do it because prompting has gotten so much, you don't even need to know how to prompt anymore, right? It should be true. But I have some lingering concerns that instead what happens is the agentic experts get another 10X better and everyone else falls further and further and further behind and are unemployable. And we go to 4 million in revenue per employee and 5 million in revenue per employee. And it's exhausting. I'm the most tired I've ever been in a sense, but my productivity rate is off the charts.

SPEAKER_01

[SPEAKER_00] You're right, Jason. What you're saying is if you're on this exponential curve that, you know, Dario and Sam keep talking about, then the knowledge frontier keeps growing every year and it's just super hard to keep up, right? In which case it does get pretty brutal. I mean, one of my favorite Max Planck quotes is, you know, science advances funeral by funeral. In other words, people don't learn and change their views. Just the people with the wrong views [SPEAKER_00] 4 million in revenue per employee and 5 million in revenue per employee. And it's exhausting. I'm the most tired I've ever been in a sense, but my productivity rate is off the charts.

SPEAKER_01

[SPEAKER_00] You're right, Jason. What you're saying is if you're on this exponential curve that Dario and Sam keep talking about, then the knowledge frontier keeps growing every year and it's just super hard to keep up, right? In which case it does get pretty brutal. One of my favorite Max Planck quotes is that science advances funeral by funeral. In other words, people don't learn and change their views. Just the people with the wrong views die. Right. And it's pretty brutal, but it's true. Gradually all the people who believe the earth is flat died and retraining is hard. It is, which explains the unpopularity. It's uncomfortable to have to learn new stuff beyond a certain age. And it must be pretty dystopian to have to do that or die.

SPEAKER_01

Don't worry, Roy, as a podcaster now, you're safe. The venture investors are screwed, but now you're a podcaster. Good to know, Harry. I got you. Don't worry. I look forward to seeing the passive income from your advertising dollars coming into my account. How far I've looked and they haven't hit, Harry. Don't worry. They're coming soon. They're coming soon. Just like productivity gains. Okay. Yeah.

SPEAKER_01

Listen, S1. Let's start with OpenAI. Roy, you teased it out at the end of last episode, brilliantly as it came breaking through. Confidential files S1. $852 billion to a trillion dollar valuation for a Q4 listing. The main question that I had, I think probably is on the minds of everyone, is OpenAI forcing its way out before Anthropic? Let's start with that.

SPEAKER_01

Yes. They have to, and they should, and they must, and they got to go now. Because they've been lapped. They've been lapped. It's pretty grim. Last year, Anthropic Gap revenue, $4.5 billion. OpenAI Gap revenue, $13 billion. Q1, Anthropic Gap revenue, $5 billion. More than the entire last year. Q1, OpenAI revenue, $5.5 billion, $5.4 billion. So still bigger than Anthropic, which is the point they were making by leaking that number. But if you look at it, it's only 35% of last year's revenue. So Anthropic has done as much in Q1 as all of last year, and OpenAI has done 30% of what they've done last year in Q1, and they're only slightly bigger than Anthropic. You play that out for a couple more quarters. We know Anthropic, but if they do anything close to 10, and then they flat for the rest of the year, let's say 10 in Q2, Q3, Q4, that's $35 billion in Gap for the year. One company is growing 10x year on year, and the other company is growing 2x, 3x year on year, and they're pretty close now. Within a couple of quarters, Anthropic, on the current trajectory, and it could change, will be visibly and obviously ahead. It could be something like high 20s Gap revenue for one, high 30s Gap revenue for Anthropic. That's not a good look, right? Especially when the one with the high 30s is profitable. Profitable, growing more quickly, and bigger. Dominant, that's Pareto dominant on all three vectors. And OpenAI could be smaller, growing less quickly, still unprofitable. That's a horrible strategic place to be, right? But that's a tough message to come out into market with. But to tell you this, would you prefer, yes, it is, but you've got two choices. You go first, and at least you can say, here's a chance to buy the first foundation model, go now, right? Which all your chat GPT, everyone knows who you are.

SPEAKER_01

[SPEAKER_02] Are you waiting, and are you waiting till the other guys go, and then you go right after them as, hey, we're like Anthropic, but smaller, and not as good? Or are you saying you wait two years and grow your way out of it? Because nothing's forever. You re-accelerate growth, and you go public in 2029. No, that's craziness, right? Unfortunately, the rule of life in all these markets is the number one gets to choose what happens, and the number two has to respond. OpenAI was number one, had that degrees of freedom. Now they could be number two. When you're number two, what you do not do is wait till the number one prices goes public, and then you come out as a pale version of that. So I think they're smart to go and they got to go. And this might be one, actually, where the risk is you exhaust some of the capital pool available, because these IPOs are—this is always something people would say, but it was never really true, right? But traditionally, folks that bought into the IPOs is a fairly slim segment of the market, because you can always buy the next day, right? Now we have Cerebus really was maximally oversubscribed, right? As Roy made to the point, it priced to the exact highest thing you could do without refiling, right? Without taking the risk of refiling with the SEC. But it's still a relatively thin investor pool, right? Even with Schwab and everyone coming in for SpaceX. So this might be the one time where it really does help to go just to not exhaust the capital pool. And tell the story. Yeah. And, you know, look, and given the trajectory of the last five years, they've turned title to tell the story. They did build a category, right? Yeah. And I would much prefer to go public as the first foundation model in the category, then go public, if in fact, the numbers that we've had leaked. And I always caveat this, because no one sees the numbers, right? If the numbers that have been leaked are correct, such that by the middle to latter part of this year, you're visibly number two, losing money and growing more slowly, then you do not want to wait until that becomes painfully obvious.

SPEAKER_01

[SPEAKER_02] If you're Anthropic, do you change anything? [SPEAKER_02] That's a good question. [SPEAKER_02] Seeing OpenAI now force this out. [SPEAKER_02] And again, I'm going to repeat myself, because you're number one, you don't have to change on what the other guy does, right? As a total historical analogy, I remember when Britain ruled the waves. I like to do my little Britain digs here, right? When Britain ruled the waves and the British Navy ruled the world, right? Jackie Fisher, the Lord of the Admiralty said, why do you have the British Navy? He said, to go anywhere you want, whatever force you want, right? It was—

SPEAKER_01

[SPEAKER_02] If you're Anthropic, do you change anything? [SPEAKER_02] That's a good question. [SPEAKER_02] Seeing OpenAI now force this out.

SPEAKER_01

[SPEAKER_02] And again, I'm going to repeat myself, because you're number one, you don't have to change on what the other guy does, right? As a total historical analogy, I remember when Britain ruled the waves. I like to do my little Britain digs here, right? When Britain ruled the waves and the British Navy ruled the world, right? Jackie Fisher, the Lord of the Admiralty said, why do you have the British Navy? He said, to go anywhere you damn well please, whatever force you want, right? It was, you can do whatever you want. That's the point of being number one, right? It was, this is what you want. You don't want to be sitting there agonizing about what the other guy does. It doesn't matter. Anthropic could decide, oh my God, OpenAI is going to go public in September, maybe we go in November, and maybe they do bring it up a little.

SPEAKER_01

[SPEAKER_02] But to your point on exhausting capital supplies, we've got space.

SPEAKER_01

[SPEAKER_02] And let's worry about that, because to some extent it's a thing, but remember, if you're profitable and growing nicely, let's just say that, if you are the better company, right, and you have the more attractive metrics, the way this market is, and the other guy's going out three months ahead of you, everyone who's investing will think, I'm going to put some money in this, but I'm going to keep some money for the good one, especially if they have telegraphed when they're going out. But the truth is this, operational excellence gives you a fair amount of strategic independence. And it's, yeah, you've got way more degrees of freedom.

SPEAKER_01

[SPEAKER_02] Or maybe, if Anthropic truly is profitable and truly has infinite demand for its shares, it doesn't need to IPO at all. Why would I, if I'm Dario, I've already pledged 90% of my shares to charity, right? I still have an altruistic mission at my heart. If I'm profitable and I can do infinite secondaries, maybe ironically, they stay like Stripe. They're like, listen, we crushed it. Why would I want to float a small amount of my shares and deal with the headaches? I am going to stay true to my mission.

SPEAKER_01

I'll tell you why. And I actually think that the dynamics of staying private longer are super different for people like Stripe than Anthropic, because at Stripe, you're in a fair degree of predictability. The business model is understood. The capital needs are light. You're kicking off cash. You're buying back shares. Those guys can actually stay private as long as they want. It's a payment processing company at scale with profits. It's a glorious thing. Who knows how it'll be valued? And that's still TBD in the public markets, but you can chunk on forever. I just think the capital needs for Anthropic are such that you would be wise to access the public markets when you can, right? Because unlike Stripe, where your capital needs are light, in the case of Anthropic, even though you're not bearing those capital needs today because you are lucky enough to have hyperscalers foolish enough to build on your behalf, for every dollar of revenue that you add, someone has to invest four or five dollars in CapEx up front. So if one day you want to do 100 billion in revenue, someone's got to invest 500 billion, and you probably want the operational freedom to get more of that money for yourself.

SPEAKER_01

And there's a dollar sum above which only the public markets can do it. Because even though, again, people talk about, oh my God, the private markets. SpaceX, as its prospectus points out so wisely, was pretty capital efficient. $23 billion raised, right? You know, it's not 100 billion, right? And it's entirely credible. It's entirely credible that if Jensen's correct, going back to the thing, and if CapEx is shock-houred, three trillion a year, and you're the leader in that industry, you might need to raise 300 billion because maybe two years from now, Microsoft doesn't want to, or Amazon doesn't want to be your hyperscaler provider. So I think they go public because the capital needs dwarf anything we've seen. Or at least it may provide some independence, right? NVIDIA's got 200 billion of cash flow a year it needs to do something with, right? And the other thing, going back to the beginning of the conversation, that's why I don't think this circular revenue is as negative as you think. NVIDIA has so much cash, it has to recirculate this capital. It has no other use for it, right?

SPEAKER_01

Just going back to this, before we do SpaceX, if OpenAI were to go out, what do you think the reception would be? Do you think it will be overwhelmingly positive?

SPEAKER_01

[SPEAKER_02] In today's market, yes, because we're risk-on and people want to make an AI bet, and they're the leader. They are the leader, and they built the category. You go outside the valley and you utter the word Claude and people look at you. ChatGPT is the thing, right? So yeah, and they're going to have stellar revenue, albeit lost. And clearly, losses aren't a problem at the moment, when you look at where SpaceX is going to come out. So I think that it would be really good. You know, really good might mean not quite as good as Anthropic, which is all that, plus profits too. But I think right now, the public doesn't have any pure-play AIs other than something like CoreWeave, now something like Cerebus. It's obvious that the top of the heap are the foundational model companies, and I think it'll be a great reception. That's why they'd be crazy not to do it. You've got to remember all this leaked. We're not going to be ready yet from the CFO, 2027, 2028. I would imagine we had a good internal session where we're doing this, guys. Thank you for your input. But the day after the litigation settled, we are going for this thing.

SPEAKER_01

[SPEAKER_02] And I think they're right. I just think two things. One, we're going to get a little preview, which is if this retail demand for SpaceX is super high, then OpenAI can copy it, right? Elon's putting 30% of this massive IPO, the largest in history so far, into retail. It sounds like he's being democratic and meritocratic in Robinhood, and maybe he is, but I think it will benefit the stock price because they want to own Elon, not the jury in Oakland, but a lot of the country wants to own it. And if OpenAI clones that for its fan base, and it drives the stock up, they'll get a little preview of how well that strategy works.

SPEAKER_01

And the other thing is, it's fun to talk about this two-horse race on the 20 VC. But again, I don't think there's anything wrong with owning both as a public market equities investor. I think [SPEAKER_02] Elon's putting 30% of this massive IPO, the largest in history so far, into retail. It sounds like he's being democratic and meritocratic in Robinhood, and maybe he is, but I think it will benefit the stock price because they want to own Elon, not the jury in Oakland, but a lot of the country wants to own it. And if OpenAI clones that for its fan base, and it drives the stock up, they'll get a little preview of how well that strategy works.

SPEAKER_01

And the other thing is, it's fun to talk about this two-horse race on the 20 VC. But again, I don't think there's anything wrong with owning both as a public market equities investor. I think owning both is a good strategy. Why? There's so much change. Even if Anthropic is better, why not buy into the IPOs of both? Put in 50 million into each, 100 million, 250 million to each, and see how it plays out over the next years. The overall trends are strong. Not everyone needs to be a savant.

SPEAKER_01

One of the ways to think about this is, these are three CEOs trying to do something incredibly hard, which is stay on top of the most dynamic, evolving industry and tech trend we've ever seen. And no one, none of the three of them, and by that I mean SpaceX, Anthropic, and OpenAI is pitching a perfect game. Right? You can look at aspects of what any of the three of them have done and go, oh, that's not as good. Right? But the truth is, all three of them from a standing start have built companies worth plus or minus a trillion dollars, and they are riding this trend, and no one else is even close. I mean, so if you look at the report card, yeah, you're right. Anthropic did an amazing job of building the model, which ironically meant that their CapEx forecasting was wrong because they underestimated the success of their own model. To some extent, OpenAI has been the other way around. They really nailed it on CapEx. They nailed it on compute, but maybe should have spent a little more time on the enterprise features of their model. So they're behind there. And obviously, if you look at the X.AI part of SpaceX, leaving aside the thing, just focusing on their AI business, they're the world's best builders of fast data centers. But so they get the points for that. But on the other hand, they lose points because they weren't able to fill it with their own stuff. And you're right, it's not quite perhaps as modern or as SOC 2 compliant as some of the stuff OpenAI got. So yeah, you can make knocks on all three, but zooming out from a standing start, they've each created a trillion dollars in AI in the last five years. What did you do on your summer vacation? Well, and Elon founded both. Elon founded both of them.

SPEAKER_01

Oh my God. You have to remember that. The entire pivot. Plus Neuralink, plus he's controlling people's bodies with his brains, plus the Hyperloop, plus Tesla. [SPEAKER_00] Yeah. [SPEAKER_00] Well, we can talk about that the entire. Because SpaceX, when we come to talk about SpaceX, the proportion of their future value they attribute to AI. It seems to me, this is Elon, effectively losing OpenAI and says, God damn, I'm going to have one of my own and made it happen, realized it from whole cloth.

SPEAKER_01

[SPEAKER_02] Well, let's discuss SpaceX directly then with the S1 dropping. Rory, given the fact that I didn't correctly identify the most important numbers, what did you think were the most important numbers when you looked at SpaceX's S1 dropping?

SPEAKER_01

[SPEAKER_02] Cynical answer, I don't think the numbers matter a damn, right? Because the point is this, everyone said the same thing. You run the sum of parts on those three businesses. There's a lowest growth, but tech enabling, amazing space business. There's a good growth, profitable, Starlink business that has much a bigger TAM than the simple rocket business and the launch business, right? But it's growing 30%, 40%. And then there's X.AI that at the time of filing was a great big gaping hole of 15 billion in CapEx and no obvious revenue. Since then, obviously, he's done two amazing deals to fund the revenue, and now he's got a $15 billion run rate business. But you add it all up, right? And the truth is, you get to a launch business, you have a Starlink business, which you can value, and you have a core weave type business, because that's what the AI business has now become. And the sum of those parts is so far lower than the proposed valuation. And the only difference between the two is the Elon premium, and that's not a number that comes out of the prospectus. So I looked at the numbers, I read them, I understood them, but they offer me no value in trying to value this company, right? I mean, other than saying utterly stupid things like the fundamental value of Tesla is 200 billion, and the trading price is six times that. So maybe the Elon premium is 6x value. I don't know. It could be the GeoCities deal of the AI era. We could look back on this. And listen, I'm all on team Elon, okay? But 100 times trailing sales, we may look back and say these were some good companies. But 100 times, my God, my God.

SPEAKER_01

No, I'm going to defend. No matter what happens, these are three amazing companies. Twitter's just, growth has fallen off, it's shrunk 50% since he bought it. Elon's the magician of revenue compression in Twitter. Revenue has fallen 50%. [SPEAKER_00] Yes, Twitter, agree. Twitter is, I mean, again, if you disaggregate the three businesses, the launch business is a good, stable, boring business with amazing technology.

SPEAKER_02

Barely growing.

SPEAKER_02

That probably would get tapped out at 10 to 20% growth rate. The Starlink business is a great business with a much bigger TAM, but 30, 40% growth rate, 10, well, think about 14 billion in revenue and EBITDA positive, right? So actually, I will give you the most interesting chart, Harry, and the number in the thing with the TAM analysis. And I want to say for memory, it was something like 28 trillion, largest TAM in history. That's actually not as interesting. The interesting part is, even though, let's say 10 years ago, this was a launch company, and eight and two years ago, this was a launch company with a communications business. 90% of the TAM that they identified is not in those two sectors. It's all about AI, right? So the interesting thing is the story they're trying to tell is a story that's 90% not focused on the two things where they're unique and differentiated, right? I think they're not unique.

SPEAKER_02

Harry, and the number in the thing with the TAM analysis. And I want to say for memory, [SPEAKER_01] it was something like 28 trillion, largest TAM in history. That's actually not as interesting. [SPEAKER_01] The interesting part is, even though, let's say 10 years ago, this was a launch company, [SPEAKER_01] and eight and two years ago, this was a launch company with a communications business. [SPEAKER_01] 90% of the TAM that they identified is not in those two sectors. It's all about AI, right? [SPEAKER_01] So the interesting thing is the story they're trying to tell is a story that's 90% not focused

SPEAKER_02

[SPEAKER_01] on the two things where they're unique and differentiated, right? I think they're not unique [SPEAKER_01] and differentiated in the AI story because Grok as a foundation model has gone nowhere. They did make an enormously clever deal in the sense that they had Colossus. They built it quick. And again, the S1 is very clear on that. We think we build faster than anyone else. And they've sold that to Anthropic. They've rented that to Anthropic for 1.25 billion a month, [SPEAKER_01] right? With, I think, a 90-day cancellation clause on either side. So basically,

SPEAKER_02

[SPEAKER_01] it's $15 billion a year, right? In one sense, that's amazing because it's amazing to think that [SPEAKER_01] you had an $18 billion business built over 20 years, and then you just added $15 billion to that [SPEAKER_01] in one transaction on a product you only built a year and a half ago. So in one sense, it validates [SPEAKER_01] the comment that the AI business is so much more dynamic even than the launch or the comms business. [SPEAKER_01] If you're willing to spend the capital up front and invest $10 billion in a hole in the ground, [SPEAKER_01] and the compute demand is there, you can get to $15 billion in revenues pretty quickly.

SPEAKER_02

[SPEAKER_01] But what you are is a more efficient core wave. Now, they will say that- [SPEAKER_01] That's worth $2 trillion? [SPEAKER_01] That's my point. [SPEAKER_01] Renting out chips because Jensen's your buddy and you're a good customer, [SPEAKER_01] that's worth $2 trillion, $3 trillion? Or are the rockets growing 10% or the Twitter [SPEAKER_01] declining 50%? Which of them exactly is worth $2 to $3 trillion? And that's why I jumped on the Geocity thing. I love the S1, right? I think it's all madness. I wouldn't buy a share. I just love the optimism. I just love the fact that they've done this thing

SPEAKER_01

[SPEAKER_02] for 20 years. And some of the assertions of why they're great were totally true. We've been fairly [SPEAKER_02] capital efficient for people who put rockets in space. We build faster than anyone else. [SPEAKER_02] We have organizing principles around engineering that the algorithm to make things [SPEAKER_02] efficient. There's a lot of bold claims here that are justified on the basis of 20 years [SPEAKER_02] achievement from probably one of the most talented entrepreneurs everywhere. It was a great read. [SPEAKER_02] Go America that we can have this, right? So after I say all that, you're right, Jason,

SPEAKER_01

[SPEAKER_02] I still don't get it. I'm so far from $2 trillion in DCF that my head hurts. [SPEAKER_00] Well, it's almost worse. The 100x trailing revenues is, it's hard to tie, [SPEAKER_02] right? But the real, and listen, I'm on Team Elon. I've bought five Teslas. I've got three Starlink [SPEAKER_02] subscriptions across. I am from the very early days. But in some ways, this is just very cynical. [SPEAKER_02] This is financial engineering, which I have mad respect for. But my God, you're taking a bunch [SPEAKER_02] of disparate assets that most of them have no connection. You're throwing an S1 talking about

SPEAKER_01

[SPEAKER_02] the history of AI, which wasn't even your business a year ago. A little bit of financial engineering, [SPEAKER_02] I'm all for. But man, this is so much fun. This is bailing out my failed Twitter acquisition,

SPEAKER_00

[SPEAKER_02] my idea that I'll compete with OpenAI just by buying chips. That didn't work. I'm bailing both [SPEAKER_02] of them out, right? I had a quiet business, which I should have kept private with SpaceX, right? It was [SPEAKER_02] a great private business that blended, was doing great things. Now I'm bailing everybody out in this [SPEAKER_02] massive confab. It's SolarCity on steroids. And we're all here for it because we love AI,

SPEAKER_01

[SPEAKER_02] but it makes no sense. This conglomeration of friends of Elon makes no sense to anyone but [SPEAKER_02] the folks getting bailed out on Twitter. It just doesn't make any sense. [SPEAKER_02] The answer is the launch business enables Starlink, right? Yep. [SPEAKER_02] And the launch business will enable data centers in space, which is how, and we've proven that we can [SPEAKER_02] build capacity on the ground. So therefore, we'll be able to build capacity in space and we'll be able [SPEAKER_02] to get 100 gigawatts of capacity a year, five years from now. Data centers in space is the thing that

SPEAKER_01

[SPEAKER_02] joins all the dots together, which is different than me saying, even for a second, I ascribe a high [SPEAKER_02] probability to that. There is a coherence to the narrative, but only if the next thing happens. [SPEAKER_02] And that's always the way. If you buy data centers in space will work, then it makes sense to have a [SPEAKER_02] launch business in the same company that has a data center business. [SPEAKER_02] And the tweets, where do the tweets fit into all this? The tweets don't fit in at all. Let's get real here. [SPEAKER_02] The tweets don't fit in at all. And the truth is that the 44 billion on Twitter was a miserable

SPEAKER_01

[SPEAKER_02] acquisition and the value of that company is less today than we bought it. And that was pretty obvious [SPEAKER_02] two years ago, right? He's chosen to bail everyone out by rolling it into X.ai. He's chosen to bail X.ai out [SPEAKER_02] by rolling it into SpaceX. And to be fair to X.ai, even though we just use the words bail it out, [SPEAKER_02] if they can keep the $15 billion a year, they're going to cut. That's a very high. I mean, the total [SPEAKER_02] CapEx in X.ai over the last two years, I think it was $12 billion and $7 billion, $19 billion,

SPEAKER_01

[SPEAKER_02] right? One year and a bit of Anthropic revenue at $1.25 billion a month, and you covered your nut, [SPEAKER_02] which is a pretty high cash on cash return for a data center project. So look, I sat there thinking [SPEAKER_02] it's all crazy too. And then I'm like, he built something right at the time when capacity is at a [SPEAKER_02] premium and he's found a way to monetize it. I mean, Jason, I repeat, [SPEAKER_02] I don't get to, I mean, even, so you apply the core weave multiple and you get to under $100 billion. [SPEAKER_02] I'm not sure what to make of that information, but it was definitely a get out of jail free card

SPEAKER_01

[SPEAKER_02] when he got the Anthropic business. By 2030, what will that core business be? Will it be Starlink? [SPEAKER_02] Will it be data centers in space? Will it be NeoCloud? Good question. I think Starlink will be the vast [SPEAKER_02] portion of the value. The value, I don't think data centers in space will be a meaningful percentage

SPEAKER_02

of revenue. And I think their existing data center business will be a high revenue, relatively low return on equity business. How that compares to Starlink, TBD. But I think we'll

SPEAKER_02

I'm not sure what to make of that information, but it was definitely a get out of jail free card when he got the Anthropic business. By 2030, what will that core business be? Will it be Starlink? Will it be data centers in space? Will it be NeoCloud? Good question. I think Starlink will be the vast portion of the value. The value, I don't think data centers in space will be a meaningful percentage of revenue. And I think their existing data center business will be a high revenue, relatively low return on equity business. How that compares to Starlink, TBD. But I think we'll discover that, as I say, it's a very core weave with the world best engineering and construction team. So implicitly, I'm saying not a whole ton of data center in space revenue on top.

SPEAKER_02

[SPEAKER_01] And that will only be worth anything like the current value if the narrative is tracking towards data centers, which again means it goes back to the first sentence of the show, which means if Jensen is right that you're spending $3 trillion a year and the ROE is there for the full $3 trillion, as Jason has pointed out, and if we can't build shit anywhere else, then at some point you'll want to stuff them into space regardless of all the pain and suffering involved in doing that. And then Elon will be right. If any of those if clauses turns out not to be true, then it'll be a bridge too far.

SPEAKER_02

Or more correctly, a rocket too far. But who knows? If the one human being on the face of the planet who's earned the right to say, give me a trillion dollars, I'm going to make the bet is Elon. And if you want that bet, I'm glad it's going to be available for people. If you want to scratch that itch, go scratch the itch.

SPEAKER_02

Look, I think he made a compromise to get it all done. I think he's going to have to build a $50 billion core weave business to make the math tie. And I think he had to give up on going to Mars and go to the moon. He reached a certain point in life where to achieve our goals, we have to be practical again. He's got into a practical phase. He's got to take this thing public. He's got to mash it all together. And the anthropic thing either has to die or keep going to make this work. So he's got to commit to building a 50 plus billion dollar core weave and going to the moon, which is pointless, but it doesn't mean it won't get him there in the end.

SPEAKER_02

[SPEAKER_01] He could also, if the world changes, I mean, we're on such an intense trajectory, right? But if the world changes, he could also dump this whole core weave business and just move on, right? If it is possible, you can write off the chips, you can write it all off. If the world goes a different way than we expect.

SPEAKER_00

[SPEAKER_02] And proving how little we all know at this point, how little you can know, another option is, if everyone else proves incredibly mediocre at building new data center capacity, and Anthropic keeps continuing to grow, then, even though they, I was thinking they did this like $1.25 billion a month, it's like when you buy storage capacity. When you send it away to storage, extra furniture, and you pay $100 a month, and you think I'll be out of this thing in six months. And then five years later, you're still paying and they're now charging you $400 a month, right? This is public storage, but for compute, and he could well be there for the next five years when Anthropic is just needing the capacity and making a ton of money.

SPEAKER_00

Consumer storage solutions, one of the best businesses. Yeah, exactly. This is like that. And you go in, I'm sure Anthropic went in thinking, we're only going to pay him $1.25 billion for four months. But they're relying on other people building data centers, and life is tricky. And not everyone will bulldoze like Elon did. And a year or two from now, because remember, this is very profitable revenue. You cannot take it away in his CapEx, in a shitty CapEx business where Core Weave is a good company, but makes only okay profits, at $1.25 billion, it's a great deal, right? Relative to his CapEx. So his ROE on that is pretty high in a low ROI business, return on equity business.

SPEAKER_00

And don't forget, it's a poor analogy, but it's the same guy. There's no one in the US or Europe that can produce EVs remotely as efficiently as Elon, like the investments he's made. And I'm not saying it's the same. He can't control the whole supply chain, but the man has a history of doing things that are both at the cutting edge and radically more efficient by investing hard, like Colossus and all of this, by doing things that are radically, maybe he does build his own massive fabs in the US and they work and we can take shots at it like we did when the Tesla Roadster came out, but no one can, outside of China, no one can approach what he's done in EVs. Maybe it happens in data centers. I mean, Core Weave has nothing. What do Nebius and Core Weave have? Nothing. No special technology, right? He could accelerate past everybody. He's got the capabilities, right? And we could turn around in five years and say, my God, he owns every data center in the galaxy.

SPEAKER_00

[SPEAKER_01] Because, I mean, if nothing else. Who else? You can't compete with a Model 3. A Model 3 you can lease in the US for $299 a month. It's self-driving, it needs no fuel, and it's better than all its competition. And the only reason we don't all drive them in the US is 60% of the country hates them. Otherwise, there would be no other car in this country, right? He could, what if he does the same for data centers? It might be with Colossus and friends and everything. He's actually doing this for real. Because remember, the one thing you need to be successful in a CapEx intensive data center business is a low cost of capital. And I think we can agree that a $2 trillion pre-money, that's a pretty damn low cost of capital. So yeah, he can sell. I mean, if he gets $75 billion, he can put $25 billion into another Colossus. Sell that for another $15 billion. Away you go.

SPEAKER_01

There are two private market stories that I want to touch on that we haven't touched on. One, AI Slop, spelled backwards is Pulsier. You're like we're being pumped. Listen, it is a company that enables a single person to build a business, and that business is then run by AI. It is bluntly not very popular. It's sent out a huge

SPEAKER_02

[SPEAKER_01] that's a pretty damn low cost of capital. So yeah, he can sell. If he gets $75 billion, he can put $25 billion into another Colossus, sell that for another $15 billion. Away you go.

SPEAKER_01

There are two private market stories that I want to touch on that we haven't touched on. One, AI Slop, spelled backwards is Pulsier. You're thinking we're being pumped. Listen, it is a company that enables a single person to build a business, and that business is then run by AI. It is bluntly not very popular. It's sent out a huge amount of unwanted emails, which is why a lot of people don't like it. The founder of OpenClaw, Peter Steinberger, even responded to the funding announcement saying how much he disliked these unsolicited emails. But they raised 30 to 40 million bucks. I can't remember the exact amount. It had a $250 million valuation, some pretty well-known firms in there. Jason, I'm intrigued. I think you'll have a thought on this one. I feel I could do it. I feel I could just call it AI Slop instead of Pulsier. I feel they can raise the 250. I feel the only question, if I did this, could I take the 40 million out of secondary? Is that okay? Do they care? I would do this if I could do it all. Listen, it would be nice. Unfortunately, I wouldn't get QSBS on it, right? Because it would be very. That's the negative, but I wouldn't mind taking 40 million. I think I could do 10 million like this if I knew I could. I don't want to raise from VCs and be stuck with the moral obligation of turning it into 2.5 billion. But I don't know. Listen, the fact that you have a startup named AI Slop raising $250 million, if this is the peak, we're going to look back and make fun of this one. But I don't know enough about it in all fairness to be sure I'm wrong on this. I feel so punked by a startup called AI Slop raising $250 million. I feel this is a bigger FU than Maltbook. Maltbook pretending these agents are talking to each other in a social network when it's all humans telling the agents to go onto Maltbook and talk to each other. I don't know, is this the next Maltbook? I literally don't know, but I would at least have asked them to please change the name.

SPEAKER_01

[SPEAKER_02] I did admire the tweak though. It is just such a two fingers salute to the great AI marketplace. I'll tell you what I don't like about it. There is, and I'm, listen, we can talk about what it's actually having just begun to build an app while we're on this Zoom or whatever. Yeah, it's pretty good. From a marketing perspective, it's pretty good. So I asked it to build an inbound AI SDR for me. I've already done this and bought and used a bunch of these tools. The way it qualified and spec'd out this product, pretty good. Certainly better than Replit or Lovable, which are horizontal tools. The journey it guided me through, it did what make couldn't do while we're here. It researched sastardoc.ai. It figured out what I did. It analyzed what tool I would likely want to build. It spec'd out the tool, did a great job. I give it a 10 out of 10 in terms of the journey so far. And then it immediately asked me to give a credit card and pay $49 a month before I got any value. My general experience with agentic tools like this is that when they ask for money before they deliver any value, it's not that good. The marketing exceeds the value of the tool. I can't think of an exception. I'm not saying they're out there, but every time I go through one of these, I immediately try to figure out where the cancel link is and it doesn't exist. But I give them a 10 out of 10 for the journey before I put in my credit card. It's pretty good. They definitely make you feel like they can build something pretty badass for you. I have to say one thing that is impressive. I met them as part of the fundraise to be very candid.

SPEAKER_01

It is one founder. There's no team.

SPEAKER_01

You're right. It is one guy. It's one guy. So it's pretty, I'm impressed with the aesthetics of it. And I know a lot of founders disagree with me, and a lot of accelerators tell you to charge instantly for these agentic products, but the ones that I love, the ones I use every day, they give you value before you have to put in a credit card. And this goes back to the Y Combinator 2 million of tokens, invest $5 in me or a dollar of tokens in me or $2 to earn 49 bucks a month, which is the entry points for the product. You can't invest two bucks in me. Disappointed. But I get why a lot of young founders or others think it's not worth two bucks, but to me, I'm out.

SPEAKER_01

[SPEAKER_02] Boys, what other stories have I missed that you think we should cover?

SPEAKER_02

Yeah, Extra raised $250 at $2.2 billion to build the search engine for AI agents today. Open Router announced their round, which was $150 million at $1.3 billion, led by Capital G. Manus founders are trying to now buy back the company and it's quite a contorted deal.

SPEAKER_02

Yeah. I mean, look, all these have in common—and the first two have in common, and the third one's obviously very different—is look, super interesting infrastructure companies all about the buildout in AI. Right. And the big picture here is enterprises are adopting AI. Yes, they're using LLMs, but there's a coterie of other tools that developers and people need to build great agents. And X is doing—I mean, what X does is around search. Because if you think about it, when ChatGPT first came out, it was dumb about anything other than what the model had been trained on. And then people like Perplexity added internet search. So now when you do a ChatGPT query or a Perplexity query or an Anthropic query, it queries the web at the same time. So it knows what's going on. Well, in the same way, if you're building as an enterprise, building an agent of your own, and you have access to the OpenAI API or the Anthropic API, what you don't have, what you also will want is some version of infrastructure and web search, either across the whole web or across a set of defined...

SPEAKER_02

[SPEAKER_01] If you remember back when ChatGPT first came out, it was dumb about anything other than what the model had been trained on. And then people like Perplexity add internet search. So now when you do a ChatGPT query or a Perplexity query or an Anthropic query, it queries the web at the same time. So it knows what's going on. Well, in the same way, if you're building as an enterprise, building an agent of your own, and you have access to the OpenAI API or the Anthropic API, what you don't have, what you also will want is some version of infrastructure and web search, either across the whole web or across a set of defined websites, so that when you build your own internal agent, it's a smart agent that knows what's going on and executes to that need. It's done really nicely themselves in parallel are the two companies doing that that have recently raised. And my point is merely, there's just a whole bunch of infrastructure going on one level below the foundation models that are just super interesting.

SPEAKER_02

We talked about OpenRouter before, I think, which is a company that allows their customers to switch between various different models, both the foundation models and the open source models. They have, I think, 50 plus models hosted on there to allow, obviously, enterprises to access the lowest cost model. And these are all just interesting trends because, I think the OpenRouter one is interesting because, going back to what Jason said, price cut. If people are spending $300 million in a premium product and having to lay off people as a result of that, there's going to be some interest in exploring cheaper costs. And I think companies like OpenRouter, the value proposition is, hey, not every query you need to make needs to be done on the most expensive foundation model. And there are other places to do that, right? So I think, again, it's just that the cliché, the picks and shovels of the agentic revolution is just a good place to be investing. And I think it will be going forward, strong, interesting teams doing good stuff. You know, you can squint one way and say foundation models hoover it all up, but I don't think they do. I think there's a lot to be done around building your great next generation agents. I mean, Jason, do you guys use Parallels or Exa or what for web search when you're building your agents?

SPEAKER_02

[SPEAKER_00] Yeah, and I'm a small investor in Exa. It's just a great product, right? It's another one where I was a user before an investor, an early user because, and the reason I think it's super interesting, even more interesting than OpenRouter, right? Which is a great one is OpenRouter is really interesting because it's a, well, first of all, they nailed the ability to dynamically and easily pick your LLM, right? They nailed something. And so it's a proxy for all these discussions, right? It's a great proxy insight. The world needs that product. Exa is more interesting and some of its peers because these are products that have no use without agents. And when we throw these term agents around, agents have a lot of meaning, right? But literally, if you're not, at least as far as I'm aware of, if you're not building an agentic product, you have no use for a tool like Exa, right? Agents don't hop on Google and do search. They don't. And they don't create files on Dropbox and they don't hop on Zoom. Agents have a whole, true agents have a different set of workflows and tools they use. And Exa nailed one of the core, it wasn't even obvious to me until I started. There's an obvious issue. Your agents do need to find current information. It's just that simple. Just like Google was the killer use app for humans, right? Your agents need it, but they don't need Google. And putting it aside whether Google should have built it rather than promote Exa, which was a ball bouncing the right way, to see the type of growth, even at the early stage for something that agents need and humans directly don't, it's a look into the future. And then it also says that a lot of stuff isn't BS, that a lot of this stuff is real. And we, the idea that we may all manage more agents than humans, isn't just BS. It's real. It's real. And they needed a whole different set of tools.

SPEAKER_02

[SPEAKER_01] They don't need Zoom. They don't need a traditional CRM. They need tools like Exa. Maybe we should only be investing in those things and leave those human tools behind guys. Enough investing in human software. A very dangerous thing to do. Can you paint to me the upside case here at 2.2 billion? What is that upside case from here? Is this a hundred billion dollar company, which powers the future of agents? And what is the market composition of this market? We've got Parag at Parallel doing it as well. Is this an Uber Lift where one of the providers takes 90%? Is this an AWS Google Cloud? Just help me understand that.

SPEAKER_02

[SPEAKER_01] I think it's more the latter than the former. It's not a, because it's a developer tool. There's not huge network effects. So you can easily contemplate two companies being pretty successful here, right? I mean, we did some work on this. We looked at Exa back in the last run, I think super team and good product. Someone bit higher and they won the deal and great name to benchmark. So what can you do? Right. I actually think it's a super good company. I like the team. I think they're smart. I think the market's there. And I don't think it's a winner take all. I think it'll be a couple of players because it just doesn't pay, you're not going to have 10 because it's not economically efficient to have that. But at the same time, it's not like Google where the winner take all emerged because it was a consumer product. And then the ad network and the structure of the market for ad sales kind of pointed to a winner take most. In this case, individual developers will be making decisions. So I can imagine comfortably two players here dividing reasonable market share, just like most developer tools. I think next level, how big can it be is the question. I think Jason's right. It's all about the more you believe in agents, the more you believe they're going to want to access structured information. And this thing's going to coattach to all that. So that's the bet here, right? You are betting on lots of agents needing lots of information, doing lots of reasoning, doing lots of search. So it's a derivative bet on agents. And right now, that's been a good bet. I think how you think about getting to a billion dollars in revenue, which is probably what you need to be in a normal market to be worth 10 billion in,

SPEAKER_02

[SPEAKER_01] Jason's right. It's all about the more you believe in agents, the more you believe they're going to want to access structured information. And this thing's going to co-attach to all that. So that's the bet here, right? You are betting on lots of agents needing lots of information, doing lots of reasoning, just doing lots of search. So it's a derivative bet on agents. And right now, that's been a good bet. I think how you think about getting to a billion dollars in revenue, which is probably what you need to be in a normal market to be worth 10 billion in market cap. I'm just thinking about it on the fly here, because I remember we looked at the numbers a year and a half ago. It's not crazy in a world. Look, if we're spending, going back to first principles, if we're spending 3 trillion on CapEx, right? That's the Jensen comment. Then the second comment is, if you look at the SpaceX S1, remember I said that 90% of the market is AI? One of the next interesting things is within that AI, they said 90% of that is enterprise. So they're basically saying the vast bulk of AI spend is going to be enterprise building shit. Right. And if that's the case, and they're building enterprises to frankly replace human work, then I think there's a pretty compelling market for the product that allows structured information to be accessible to that agent. I think you'll expand beyond just web search. I think you'll do curated lists. I think you'll do internal information, but there's totally a need for that product.

SPEAKER_02

[SPEAKER_00] That's wild. Another benchmark company. This fund is going to be one of the best performing funds in history. I mean, seriously. [SPEAKER_01] Yes. Good for them. [SPEAKER_01] That's the job. [SPEAKER_01] That's the job. [SPEAKER_00] Every company, Jason. The number of hits is extraordinary.

SPEAKER_02

[SPEAKER_01] Yeah. And so they did exit 700, right? Or something like that. Right. I think you're right. This is 20 VC. So it'll be an Epic fund. I think also going in early and realizing that investing in a smidge of traction and breaking your rules a little bit was the right bet for today. Arguably, I think was maybe another insight too, right? Jack Altman just joined, right. And he just did Monaco, which I know—well, I'm a small shareholder. And they marked up the deal very quickly, almost two X from Founders Fund. But when Founders Fund invested, it had no revenue. So he got a much better deal for Benchmark by investing a couple of months later, but going from pre-revenue to blowing up, right? That's if you can get into the moment an AI leader blows up, that's when you want to invest. Right. And so then they've got they've tripled their evaluation. And I just think this going back to seed is for suckers. You want to invest the minute it blows up. The minute it blows up, you want to get the DM and just wire the money in AI. That's the play.

SPEAKER_02

Pushing on that a little. Because first of all, you got to give huge credit then to Lightspeed for doing the seed in XR. Yeah. For suckers. Well, I mean, if a sucker means you won a lot of percentage at low dollars. Yeah, no, but I think what you're saying is actually, so first of all, I just wanted to give that credit. But second, I think what you're saying is correct, which is there's a quantum reduction in risk when you go from no revenue to revenue. And then there's a linear reduction in risk thereafter. So the sweet spot, I always think on the risk return continuum is early product market fit, which is why we focus on it. Now you're right, Jason, the amazing thing about these AI companies is that moment used to be a year, year and a half, maybe two years of early product market. Now what you're seeing is you go from early product market fit to, "Oh my God, it's incredibly obvious." Just given the take rate, literally sometimes in weeks, if not months and definitely not years. I mean, the Lovable and Replit numbers. I think the same thing kind of happens here. You got to find that moment and yeah, you got to pay. I mean, in retrospect, I was wrong. You should have bid what I bid. You should have bid 750. Yeah. But also you have to abstract away. Not only does it happen so fast, but the counter side is you have to abstract away some of the competitive risks. You have to not worry that Google could—why didn't Google do it? Right. Because otherwise you'll do none of these deals. Even if Google wouldn't do it, the company didn't even exist a year ago in its current form. So what are the renewals going to look like? We don't know any of these data points, right? I mean, I agree. This is actually an interesting discussion about our job as this thing from being public market pontificators.

SPEAKER_02

You're right. You have to decide that you're willing to pay more on less information than the SaaS era, simply because the market and traction that you're going to see if you're right are so much quicker and so much bigger. That is the core of figuring out what's going on right now. And that's been the bet that's worked. Right. And I look back on the stuff we did that I feel right about. Those are the criteria. And then obviously you have to get the traction. If you don't, you feel like a moron. Right. And on the stuff you missed, you look back and you go, you should have leaned in even more. You should have been more willing to give conviction on the traction, even though the data points were sparse because the stuff that's working, that's in tune with what people are trying to build is the universality of the propensity to adopt right now. There's no 10 year cloud journey where some people decide now and more people decide two years from now. It goes back to the first thing we said, corporate America flipped the switch a year ago and said, "Thou shall do AI in 2026." And if you're selling anything on the picks and shovels business to help corporate America do AI, you get revenue this year. You don't get the luxury of waiting. "Oh, I'll wait to see one year's renewals before I bid." There's a thread, which is, can you make, if you have a generic products, can you make up in volume? So there is a software explosion. The number of apps we are building, it may not be benefiting Dropbox, right? Where Drew just stepped down, but overall the number of apps is exploding at a rate we've never seen before.

SPEAKER_02

[SPEAKER_01] It goes back to the first thing we said, corporate America flipped the switch a year ago and said, thou shall do AI in 2026. And if you're selling anything on the picks and shovels business to help corporate America do AI, you get revenue this year. You don't get the luxury of waiting.

SPEAKER_02

[SPEAKER_00] Oh, I'll wait to see one year's renewals before I bid. There's a thread, which is, can you make, if you have a genetic products, can you make up in volume? So there is a software explosion. The number of apps we are building, it may not be benefiting Dropbox, right? Where Drew just stepped down, but overall the number of apps is exploding at a rate we've never seen before.

SPEAKER_02

[SPEAKER_01] So apps are being built everywhere. Workflows are being built everywhere. Agents are being built. If you consider an agent a piece of software, it's exponential, right? At Databricks, Neon, which is their Supabase competitor, over 90% of the databases are built by agents, not by humans, okay? But they don't pay as much as they pay Mongo. So the question is, and I don't think Exa makes anywhere near what Google makes on a search, right? And so where I think we're going to get caught in a lot of these tools is where there aren't orders of magnitude more volume. And the reason Supabase and Exa and these groups are neat is because you don't need them for every app you build or every use case, but you sure need them for a lot. And we're going to make mistakes as investors. I know I've already made one or two and they're not fatal mistakes, but where it all is an agentic workflow. Agents need it, but they don't need it all the time. And so I can't make it up in volume. We need vast volume for these plays to work, vast.

SPEAKER_02

[SPEAKER_01] You're right, Jason, because you're right. The dollars per search of the whale. What you're saying is, what is the call it the harness, what are the core primitives that it takes to build an agent? And what are the other things that pretty much most enterprise developers will figure out, oh, I need a this, a that, and the other to make my agent. And if you pick the thing that only 10% of developers need, because it's corner case, you'll pay up and you'll have a tiny market. You'll be sad. But there probably is going to be five or six things like a database, like a search engine, like some kind of observability where over the next two or three years, revealed preference will say, the vast majority of developers realize they need this. It doesn't come in the box in the harness from Anthropic or OpenAI, and therefore standalone companies will be built doing it. You're right. And we just actually had that discussion last week. What are the five or six things that you need to own? Right. And I think it's credible argument that agent search, which is Exa. And I think I could be wrong on the Open Router thing is that a model switching layer could be two of those things that when you look back five years from now, every app company, either building a third party app or building in house said, oh, I need that product. So they end up with mass developer adoption. That's the best.

SPEAKER_02

Okay. We're going to do a final round, which as always is rage bait, but real. I just tweeted and me and Jason were talking about it. I just interviewed a CEO who said three things. Number one, we replaced our 600K Salesforce contract with a vibe coded CRM, which we built in three weeks. Number two, we will get rid of 80% of the SaaS we use internally. And number three, if Anthropic doubled pricing, we would not change our usage. Which do you think is the most rage bait statement there?

SPEAKER_02

I think the first one. The first one is rage bait. Yeah. And it's unhelpful. It's unhelpful. If you want to build your own CRM on top of Postgres, go for it. We moved to headless Salesforce. Okay. We don't log into Salesforce anymore. We could, we're a small enough organization. We could swap that out for our own database. It's not worth it. It's not worth maintaining that database. It's not worth fixing the connectors to our 10 external agents, which are all natively built into Salesforce. It's not worth giving up Agentforce, but if we needed no third party apps and we had time to burn, sure, because we already have our own layer, right? We already have our own autonomous agents running on top of Salesforce, but this is such a waste of time talking about vibe coding away Salesforce. And we have more important threats to our existing portfolio than this one, or more important things to invest in. So I just think it's a dated 2025 take. Even if there's truth in it, even if there's truth in it, his statement was actually that the functionality was so much better in what they were able to custom build with the integrations that they needed, given their healthcare focus company, that it was actually worth it to build their own and that they were able to do so much more given that. It might be if you have extreme vertical focus and willing to have an internal CRM that is not collaborative and you have a relatively small team using it and the benefits of knowing a traditional CRM interface and the thousands of apps that integrate with them are zero to you and you have time, because you don't save any money. This isn't worth it for the money. If all those things are true, vibe it away. But this is of all the threats there are to Salesforce and HubSpot and Monday and Atlassian. This ain't in the top 10, right? That's why I think it's rage bait because it matters, but it's not important in today's world. And it doesn't even explain the SaaSpocalypse. Other things explain the SaaSpocalypse. I like that answer on the first and agree. Yes. If you can get better functionality, knock yourself out, but vibe coding, that's silly. I actually want to take the third one and put it back on the CEO. What he basically said is if Anthropic doubled their prices, I wouldn't mind, right? Then you should be ashamed of yourself. You know why? Because what it basically says is the ROI on this AI is so good, I could pay twice as much. Well, in that case, let me tell you what you should do right now. You should go in and tell your people, use twice as much because logically there is some kind of ROI on every token you spend and the early ones are really great. And eventually you probably hit the marginal cost is exactly equal to marginal advantage. And then you're overspending. And if you're token maxing,

SPEAKER_02

[SPEAKER_01] actually put it back on the CEO. What he said is if Anthropic doubled their prices, I wouldn't mind, right? Then you should be ashamed of yourself. You know why? Because what it says is the ROI on this AI is so good, I could pay twice as much. Well, in that case, let me tell you what you should do right now. You should go in and tell your people, use twice as much because logically there is some kind of ROI on every token you spend and the early ones are really great. And eventually you probably hit the point where marginal cost is exactly equal to marginal advantage. And then you're overspending. And if you're token maxing, you're overspending, right? If you think you are willing to pay twice, if the projects that you're doing right now with AI really, not just rage bait, but really have a 2x return, have a return that's 2x better that could support twice the price, then what you should say to your engineers is you're not using enough, go do more, right? Because the two things have to be true at the same time, right? If Anthropic could double your prices and this project is still economically viable, then what's the next project on your to-do list? Go do it.

SPEAKER_02

[SPEAKER_01] I'm with you. But the company Harry's talking about is at 2 million in revenue per employee, so they can afford it. Let's be clear. They can afford to spend more on tokens. You're right. Of course you're right, Rory. If you're that efficient and you don't care what it costs, spend more, right? You haven't reached that. But I will tell you because we're there, I don't care what we spend on Anthropic either. Okay. Our AI VP of marketing customer success costs $257 a month. I don't care if it's 500, what's 257? We're getting toward the end of the pod. It doesn't matter. It doesn't matter. But here's the interesting thing.

SPEAKER_02

This is why I actually think it's interesting is we're all, if you're really good at this stuff, the issue isn't more money. The issue is idleness. Our agents are idle. We don't have enough brain sight. This is why these, what's the term for folks that can't sleep in San Francisco? They're coding all night. They have a, what's the term? I forget what it is. I could be cynical and say incels.

SPEAKER_02

No, there's this term where you're addicted to it. Your brain's been rewired. What the problem is, listen, there's the Uber issue where we spent too much on slop. Okay. Or we don't know what to measure. There's a different issue coming, which is coming for more of us in tech, which is that our humans just can't process all the output. It's not even about the money anymore. If you wake up and build me 50 features every single day, how many features can I even qualify overnight? Because we wake up every day, Rory, and our AI VP of marketing nags the hell out of us with three ideas we should be doing every day. It's exhausting. You can't implement 21 ideas a week, right? Every day, 1 PM, three ideas of the day, 10 K comes up with and pushes. You didn't get to my idea yesterday. Did you? Now you're behind, right?

SPEAKER_02

[SPEAKER_01] But stop. You're right. I agree with what you're saying. I just still go back to my comment though, is that it's, that paper I cited is that you can increase the complexity of the task and spend more money because he's implicitly saying he's getting way more value than he's paying for it. Right? So when you look and you're saying the same, 250 bucks, which is nothing. You're saying you get three great ideas a day and you can't process them. Correct? Yeah. We get more than that, but in addition, we get nagged. We got three things we need to do. We're going to end on a note of optimism. In that case, you need to hire another human.

SPEAKER_02

[SPEAKER_01] Oh, I agree. Because if the AI is giving more good ideas than the humans can process, and this is a whole nation of compliments and substitutes. If the AI is now giving more ideas than one human can process and they're good ideas, hire a second human who will be now more effective and the return on human capital will go up. But you're right, you know what the problem is to tie it all together. I desperately need a human for those ideas, but a 125 K laid off tech worker is worth nothing to me. I need a million dollar person to process these ideas and they need to turn that into 5 million and I will hire them today for a million bucks. No joke. But you got to be the million dollar person. Not think, not want to be that person. You got to be that person. The odd thing is every single discussion here has gone back to what's the optimal span between humans and tokens and what are the consequences of that configuration? That is the same question the whole time.

SPEAKER_02

[SPEAKER_01] Jason also does not like other humans. And so that would be [SPEAKER_01] I do. I don't know why people say this. I'm a people person. [SPEAKER_01] That's what they all say. [SPEAKER_01] I'm a people person. As long as I don't have to interact with too many on any given day. [SPEAKER_01] Jason, I hate to be with you. If you have to say you're a people person, you're not a people person. [SPEAKER_01] I know. I know you're right. [SPEAKER_01] You're right.

SPEAKER_01

[SPEAKER_02] If you have to say you're right. Do you want me to be honest? [SPEAKER_02] I am not. [SPEAKER_02] Whatever the two by two is, Harry and I are at the opposite ends of whatever the two by two is. [SPEAKER_02] Okay. [SPEAKER_02] The extroversion, people person, everyone's Harry's best friend. I never get that. Why don't I have to say on social media? I was just in London meeting with my best friend, Harry Stebbins. I see like 11 of these on Twitter a day. [SPEAKER_02] Because you're not insincere like Harry. [SPEAKER_02] I see. That could be part of it. That might be part of it. Yeah. [SPEAKER_02] Yeah.

SPEAKER_01

[SPEAKER_02] Because I'm my big man. When I'm in, when people are in town, they come see me. [SPEAKER_02] But they're always my best friend. How many best friends can you have? 300? [SPEAKER_02] Oh, many. [SPEAKER_02] The Dunbar number is 160, just for the record. [SPEAKER_02] Best friends? [SPEAKER_02] Oh, dear. [SPEAKER_02] I see. That could be part of it. That might be part of it. Yeah. [SPEAKER_02] Yeah. [SPEAKER_02] Because I'm my big man. When I'm in, when people are in town, they come see me. [SPEAKER_02] But they're always my best friend. How many best friends can you have? 300? [SPEAKER_02] Oh, many. [SPEAKER_02] The Dunbar number is 160, just for the record.

SPEAKER_01

[SPEAKER_02] Best friends? [SPEAKER_02] Oh, dear. [SPEAKER_02] Yo, boy. [SPEAKER_02] Yo, boy. Yo, boy. on tokens. We don't want to change our guidance. We got some lift in efficiency, so we can touch a hot, we don't $100 million of better off operating income. So yay us. But that means we have $200 million of OPEX we didn't expect to have, and we have to terminate a whole bunch of people. At that point, HR comes in, right? And if you think you can do that just on vibes or no ROI, you're fooling yourself, right? So I actually agree with you. I think once, once AI spend starts to eat big chunks of wages in large companies, you're just going to have to have that discussion.

SPEAKER_01

And it's, it's going to become more, more quantitative. It's going to have to be more provable, right? Yep. Um, big, and not just for humane and political reasons, but just because economically, it makes sense. I mean, someone should stand there and say, okay, we've never done a layoff in five years. We're about to terminate 10% of our people. Are we sure that we're getting the ROI from this investment? And if the person leading the investment says, I don't think, I think so, but we haven't really checked. Hopefully a CEO will say, well, before we terminate a thousand people, shouldn't we check, perhaps take another week here, guys?

SPEAKER_01

That's the ultimate symbol though, that the layoffs that we're seeing is ultimately a case of overhiring during COVID, not AI efficiency. That's the dumbest thing I've ever, take I've ever seen on Twitter is that this is overhiring during COVID. Dumbest take from the smartest people, including Mark Andreessen, who's got 40 IQ points on me. The dumbest take I've seen since our last show, the dumbest take. Well, why? Because every single... What was COVID? I'm starting to forget COVID. Even 2020. I mean, what is now, Rory, what is natural attrition across your portfolio? 15% a year, 20% a year? 25% a year, right?

SPEAKER_01

This is not overhiring. It might be that your worst employees didn't leave. That's a related, but different issue, right? That's that. And actually, I think that's kind of part of it is the unreskillable need to go. But this is not overhiring with 20% attrition. I can't, what's 20% times six? How does that compound? Yeah, it's... It doesn't matter. Yeah, exactly. You get to the same place. You're saying the 16,000 cut from Intuit, the thousands cut from Coinbase, the 800 cut from LinkedIn. You're saying that is not overhiring. That is AI efficiency. I'm saying there is already more natural attrition since 2020 than these numbers. So blaming it directly

SPEAKER_02

on overhiring is clickbait. It's clickbait because it's not mathematically true. I saw a really cool tweet that basically said, it's five years since COVID people. Haven't you heard of performance reviews? That was me. That was you. That was my tweet. I love that. That was my point. I mean, you have no excuse. You had five years to manage your low performers out and now you're blaming it on overhiring? Hang on. The only thing that could be true, just articulate the case. The only thing that could be true is you're hired to this level and you have this kind of human anchoring whereby you perceive

SPEAKER_02

you need all these people. So every time they churn, you just rehired because you had myopia and you didn't realize you could get by on much less. And now suddenly terror has opened your eyes. It is possible in mature organizations with monopoly type profits. You see that behavior, but I'm with you, Jason. It's weird to see it across the board. Right? Yeah. It just can't be true. Like it's just, it's an excuse. And I don't think any, I, I don't totally get this mea culpa, this click up cloudflare mea culpa on Twitter. I don't totally get what the vibe is here. And I do think they're 80% honest and direct, but this bloat thing seems

SPEAKER_02

like the, the, the excuse, like it's can't, Jason, can you just explain the click up and the cloudflare just so people have some context? Just publicly explaining why we're doing 20% layoffs in thousand line thousand, 2000 word tweets. And I just actually don't, I'm not, I'm literally saying I don't understand the PR benefit. And the, the meta problem I find is even if it's true, they're sort of blaming the employees like, or, or even if they're not, it's perceived to be blaming the employees for not understanding AI, not being able to change, not being able to evolve. So we've got to let you go before the cloud

SPEAKER_02

for a click ups. They were always like our very best people have to go, please hire them. Here's a Google sheet. There's no one better than Harry Rory and Jason we've ever hired. It has nothing to do with organizational change or the world, or that there are low performers. It's just a flip of the coin. Like we had to let somebody go. We flipped a three headed coin, a four headed coin and Harry Rory and Jason came up, but they're just as good as the folks we kept. That was the vibe. Like even through last year. Now it's like, it's this, it's ruthless. Like to, like cloudflare, like Matthew's like cloudflare is doing great. And we're still walking 20% of the

SPEAKER_02

people out the door. And look, it's some sort of messaging to the markets, existing employees and others that I don't totally, the, the, the, I think CEOs feel that they should be direct, right. In a way that everyone was uncomfortable to do through 2024. Everyone was uncomfortable being direct. Everyone had to say, everyone's great. We're all great. And I, I, I don't know if it's gone too far, but, but it's, it's very public. Like, I don't need to know, like if cloudflare is laying off 20, 21%. I mean, I know they're a public company, but that's about all I need to know. It's funny. Cause I, you're damned if you do damned if you don't,

SPEAKER_02

cause the Intuit CEO did it. It's not about AI and also got slammed for that. People saying, well, of course it's about AI because look at these deals you did on Tropic you're being competed with. And yeah. Um, it's always about AI one way or the other, isn't it? Well, directly or our growth is slowed because we don't have an agentic product or, or we're, or we're not being renewed by CEOs because they need budget for anthropic tokens. It's it's it's the world isn't static, right? That's the honest message. That would be my tweet guys. Sorry. The world isn't static. I partially screwed up. This is what I'm doing to try to fix it. Blame me.

SPEAKER_00

There's no easy way to give bad news, right? There's no easy way to lay off a thousand people. So no matter what you say, you know, you're going to get dinged on it, which is what you know, and I think it's, it's been evolving over the last six months, how people have approached it. I liked what Zeb said though, from ClickUp, even though he got hazed more than math. But one of the things I liked that he said, and it was very transparent. It's like, listen, I'm laying off 22% of my company so I can pay a million dollars to my high performers in the age of AI because one it's, well, he didn't really say it's market, but that's one reason you have

SPEAKER_01

to do it. The other point is if a 10 X employee, if a 10 X engineer is a hundred X engineer or $1 million bookings a year, sales rep is now two and a half million. I, I got to pay them. I got to pay them. I got to pay what was a three or 400 or $500,000 a year and a million dollar employee because they're delivering that much value. And I I've lived it. I mean, we've gone not to be a broke record. We're doing with two and a half people. We're doing what we did with 20 and everyone's get paid more on my little team, right? Everyone should make a million bucks because we don't have the 18 other

SPEAKER_01

people's salaries to pay and we're more productive. Right. And I, I don't know if it's going to happen at the Uber scale, but it's going to happen all across startups. Right. And with $2 million or more per employee in startups, you got to pay, man, you got to pay your high performers. And to be fair, you're consistent. Cause when I did my three months ago, I was kind of doing my four different ways, four different reasons for AI, the layoffs, all of only some of which are AI you added that fifth, which is I'm terminating people to get different people, which effectively is what this guy is doing and saying, I'm terminating people to make room

SPEAKER_01

for people that might cost more per head that are just of different skills. Well, that, that, that might be what he's doing, but I, but what Zeb really said is I just want to pay my current high performers more. Yeah. It's, it's, it's, it's to say I'm making space for people who you can keep more salary dollars to. And the funny thing is guys, this is the counter argument to the last discussion, right? We're actually repeating ourselves in a weird way. We just had a discussion on this. Does the AI, you know, does, does 20% of spend, does 20% of R and D spend on tokens make sense?

SPEAKER_01

And if it does, that's, you know, first, because the first, the first section of the, of the podcast was is, you know, does token spend make sense and has it a positive ROI and is it good for, and thus it's good for entropic. And then the second section is I'm shocked to discover this layoff. So if you zoom out a million miles, that's the consequence of the first, right? These are actually the same discussions. If there is efficiency gains from AI across R and D and

SPEAKER_00

sales and marketing particular, and then on top of that, if there's also product impacts,

SPEAKER_01

which I always, I always like to separate internal impact versus product impact. If those things are true, then everyone's going to be doing 20% layoff just because it's 20% more efficient. My aha is this discussion and the last discussion are in fact, the same discussion. In one case, we're focusing on the winner, which is entropic. And in one case, we're focusing on the loser, which is employees. Well, I think, yeah, but the, but the top employees, I think what ClickUp share, I mean, we, we, we, we read that when you see them on social media, Anthropics hiring a social media manager in King's cross for $450,000 a year. And you're like, how, how could that be?

SPEAKER_01

Right. But their revenue per employee is so astronomically high that it all ties. Right. And I think the new paradigm when, when the three of us met 2 million per employee was what you reach someday, right? That was Apple and others that, that was the, you didn't even have to think startups never got there. I think 2 million per employee is going to become the new normal as you scale. And that will mean that your high performers can make two to three times what they used to make, just like an anthropic. Um, because you just don't need as many people and you're going to do what ClickUp did is you're going to funnel your compensation to your high performers instead of

SPEAKER_01

the high performers making 40 or 50% more than your mid pack. It should be five X. What you're basically saying is a new set of tools have come to the floor in the last two or three years. And if you know how to use those tools, your productivity relative to someone who doesn't have those tools is so great that we can afford to pay you half a million bucks because you're

SPEAKER_00

obviating the need for three people. Being an agentic, uh, an agentic expert should be like Excel in 12 or 24 months. Everybody should be able to do it because, because prompting has gotten so much, you don't even need to know how to prompt anymore, right? It should be true. But I, I have some lingering concerns that instead what happens is the agentic experts get another 10 X better and everyone else falls further and further and further behind and are unemployable. And we go to 4 million in revenue per employee and 5 million in revenue per employee. And it's exhausting. Like I'm the most tired I've ever been in a sense, but my productivity rate is off the charts.

SPEAKER_00

You, you, you're right, Jason. What you're saying is if you're on this exponential curve that, you know, Dario and Sam keep talking about, then the knowledge to be frontier keeps growing every year and it's just super hard to keep up, right? In which case it does get pretty brutal. I mean, one of my favorite Max Planck's quote, Max Planck quotes is, you know, science advances funeral by

SPEAKER_01

funeral. In other words, people don't learn and change their views to just the people with the wrong views die. Right. And it's pretty brutal, but it's true. Right. You know, gradually all the people who believe the earth is flat died and the, you know, and retraining is hard. I mean, it is, which explains the unpopularity. It's uncomfortable to have to learn new stuff beyond a certain age. And, you know, and it must be pretty dystopian to have to do that or die. Don't worry, Roy, as a podcaster now, you're safe. The venture investors are screwed, but now you're a podcaster. Good to know, Harry.

SPEAKER_01

I got you. Don't worry. I look forward to seeing the past true income from your advertising dollars coming into my account. How far I've looked and they haven't hit, Harry. Oh, don't worry. Don't worry. They're coming soon. They're coming soon. Just like productivity gains. Okay. Yeah. Listen, S1. Let's start with OpenAI. Roy, you teased it out at the end of last episode, brilliantly as it kind of came breaking through. Confidentially files S1. $852 billion to a trillion dollar valuation for a Q4 listing. The main question that I had, I think probably is on the minds of everyone, is OpenAI forcing its way out before Anthropic? Let's start with that.

SPEAKER_01

Yes. They have to, and they should, and they must, and they got to go now. Because they've been lapped. They've been lapped. It's pretty grim. Last year, Anthropic Gap revenue, $4.5 billion. OpenAI Gap revenue, $13 billion. Q1, Anthropic Gap revenue, $5 billion. More than the entire last year, Q1, OpenAI revenue, $5.5 billion, $5.4 billion. So, still bigger than Anthropic, which is the point they were making by leaking that number. But if you look at it, it's only, what is it, 35% of last year's revenue. So, Anthropic has done as much in Q1 as all of last year, and OpenAI has done 30% of

SPEAKER_01

what they've done last year in Q1, and they're only slightly bigger than Anthropic. You play that out for a couple more quarters. We know Anthropic, but if they do anything close to 10, and then they flat for the rest of the year, let's say 10 in Q2, Q3, Q4, that's $35 billion in Gap for the year. One company is growing 10x year on year, and the other company is growing 2x, 3x year on year, and they're pretty close now. Within a couple of quarters, Anthropic, on the current trajectory, and it could change, will be visibly and obviously ahead. I mean, it could be something like,

SPEAKER_01

you know, high 20s Gap revenue for one, high 30s Gap revenue for Anthropic. That's not a good look, right? Especially when the one with the high 30s is profitable. Profitable, growing more quickly, and bigger. Dominant, that's Pareto dominant on all three vectors. And OpenAI could be smaller, growing less quickly, still unprofitable. That's a horrible strategic place to be, right? But that's a tough message to come out into market with. But to tell you this, would you prefer, yes, it is, but you've got two choices. You go first, and at least you can say, here's a chance to buy the first foundation model, go now, right? Which all, your chat GPT,

SPEAKER_02

everyone knows who you are. Are you wait, and are you waiting till the other guys go, and then you go right after them as, hey, we're like Anthropic, but smaller, and not as good? Or are you saying you wait two years and grow your way out of it? Because nothing's forever. You re-accelerate growth, and you go public in 2029. No, that's craziness, right? Unfortunately, the rule of life in all these markets is the number one gets to choose what happens, and the number two has to respond. OpenAI was number one, had that degrees of freedom. Now they could be number two. When you're number two, what you do not do is wait till the number

SPEAKER_02

one prices goes public, and then you come out as a pale version of that. So I think they're smart to go and they got to go. And this might be one, actually, where the risk is you exhaust some of the capital pool available, because these IPOs are, this is always something people would say, but it was never really true, right? But traditionally, folks that bought into the IPOs is a fairly slim segment of the market, because you can always buy the next day, right? Now we have, you know, Cerebus really was maximally oversubscribed, right? As Roy made to the point, it priced to the exact highest thing you could do without refiling, right? Without taking the risk

SPEAKER_02

of refiling with the SEC. But it's still a relatively thin investor pool, right? Even with Schwab and everyone coming in for SpaceX. So this might be the one time where it really does help to go just to not exhaust the capital pool. And tell the story. Yeah. And, you know, look, and given the trajectory of the last five years, they've turned title to tell the story. They did build a category, right? Yeah. And, you know, I would much prefer to go public as the first foundation model in the category, then go public, if in fact, the numbers that we've had leaked. And I always caveat

SPEAKER_02

this, because no one sees the numbers, right? If the numbers that have been leaked are correct, such that by the middle to latter part of this year, you're visibly number two, losing money and growing more slowly, then you do not want to wait until that becomes painfully obvious. If you're anthropic, do you change anything? That's a good question. Seeing open AI now force this out. And again, I'm going to repeat myself, because you're number one, you don't have to change on what the other guy does, right? As a total historical analogy, I remember when Britain ruled the waves,

SPEAKER_02

I like to do my little Britain digs here, right? When Britain ruled the waves and the British Navy ruled the world, right? Jackie Fisher, the Lord of the Admiralty said, why do you have the British Navy? He said, to go anywhere you damn well please, whatever force you want, right? It was, you can do whatever you want. That's the point of being number one, right? It was, this is what you want. You don't want to be sitting there agonizing about what the other guy does. It doesn't matter. Antropic could decide, oh my God, open AI is going to go public in September, maybe we go in November, and maybe they do bring it up a little.

SPEAKER_02

But to your point on exhausting capital supplies, we've got space. And let's worry about that, because look, to some extent it's a thing, but remember, if you're profitable and growing nicely, let's just say that, if you are the better company, right, and you have the more attractive metrics, the way this market is, and the other guy's going out three months ahead of you, everyone who's investing will think, I'm going to put some money in this, but I'm going to keep some money for the good one, especially if they have telegraphed when they're going out. But the truth is this, operational excellence gives you a fair amount

SPEAKER_02

of strategic independence. And it's, yeah, I mean, you've got way more degrees of freedom. Or maybe, maybe if Anthropic truly is profitable and truly has infinite demand for its shares, it doesn't need to IPO at all. Why would I, if I'm Dario, I've already pledged 90% of my shares to charity, right? I still have an altruistic mission at my heart. If I'm profitable and I have, and I can do infinite secondaries, maybe, maybe ironically, they stay like Stripe. They're like, listen, we crushed it. Why would I want to float a small amount of my shares and deal with the

SPEAKER_01

headaches? I am going to stay true to my mission. I'll tell you why. And I actually think that the dynamics of staying private longer are super different for people like Stripe than Anthropic, because at Stripe, you're in a fair degree of predictability. The business model is understood. The capital needs are light. You're kicking off cash. You're buying back shares. Those guys can actually stay private as long as they want. It's a payment processing company at scale with profits. It's a glorious thing. Who knows how it'll be valued? And that's still TBD in the public markets, but you can chunk on forever. I just think the capital needs for Anthropic are such that

SPEAKER_01

you would be wise to access the public markets when you can, right? Because unlike Stripe, where your capital needs are light, in the case of Anthropic, even though you're not bearing those capital needs today because you are lucky enough to have hyperscalers foolish enough to build on your behalf, for every dollar of revenue that you add, someone has to invest four or five dollars in CapEx up front. So if one day you want to do 100 billion in revenue, someone's got to invest 500 billion, and you probably want the operational freedom to get more of that money for yourself. And there's a dollar sum above which only the public markets can do it. Because even though,

SPEAKER_01

again, people talk about, oh my God, the private markets. SpaceX, as its prospectus points out so wisely, was pretty capital efficient. $23 billion raised, right? You know, it's not 100 billion, right? And it's entirely credible. It's entirely credible that if Jensen's correct, going back to the thing, and if CapEx is shock-houred, three trillion a year, and you're the leader in that industry, you might need to raise 300 billion because maybe two years from now, Microsoft doesn't want to, or Amazon doesn't want to be your hyperscaler provider. So I think they go public because the capital needs to dwarf anything we've seen. Or at least it may provide some

SPEAKER_01

independence, right? NVIDIA's got 200 billion of cash flow a year it needs to do something with, right? And the other thing, going back to the beginning of the conversation, that's why I don't think this circular revenue is as negative as you think. NVIDIA has so much cash, it has to recirculate this capital. It has no other use for it, right? Just going back to this, before we do SpaceX, if OpenAI were to go out, what do you think the reception would be? Do you think it will be overwhelmingly positive?

SPEAKER_02

In today's market, yes, because we're risk-on and people want to make an AI bet, and they're the leader. They are the leader, and they built the category. You go outside the valley and you utter the word Claude and people look at you. ChatGPT is the thing, right? So yeah, and they're going to have stellar revenue, albeit lost. And clearly, losses aren't a problem at the moment, I mean, when you look at where SpaceX is going to come out. So I think that it would be really good. You know, I mean, really good might mean not quite as good as Entropic, which is all that, plus profits too. But I think right now, the public doesn't have any pure-play AIs other than,

SPEAKER_02

you know, something like CoreWeave, now something like Cerebus. It's obvious that the top of the heap are the foundational model companies, and I think it'll be a great reception. That's why they'd be crazy not to do it. You've got to remember all this leaked. We're not going to be ready yet from the CFO, 2027, 2028. I would imagine we had, you know, a good internal session where, like, we're doing this, guys. Thank you for your input. But the day after the litigation settled, we are going for this thing. And I think they're right. I just think two things. One, we're going to get a little preview,

SPEAKER_02

which is if this retail demand for SpaceX is super high, then OpenAI can copy it, right? I mean, Elon's putting 30% of this massive IPO, the largest in history so far, into retail.

SPEAKER_01

It sounds like he's being democratic and meritocratic in Robinhood, and maybe he is, but I think it will benefit the stock price because they want to own Elon, not the jury in Oakland, but a lot of the country wants to own it. And if OpenAI clones that for its fan base, and it drives the stock up, they'll get a little preview of how well that strategy works. And the other thing is, it's fun to talk about this two-horse race on the 20 VC. But again, I don't think there's anything wrong with owning both as a public market equities investor. I think owning both is a good strategy. Why? There's so much change. Even if Anthropic is better,

SPEAKER_01

why not buy into the IPOs of both? Put in 50 million into each, 100 million, 250 million to each, and see how it plays out over the next years. The overall trends are strong. Not everyone needs to be a savant. One of the ways to think about this is, these are three CEOs trying to do something incredibly hard, which is stay on top of the most dynamic, evolving industry and tech trend we've ever seen. And no one, none of the three of them, and by that I mean SpaceX, Anthropic, and IBO is pitching a perfect game. Right? You can look at aspects of what any of the three of them have done and go,

SPEAKER_01

oh, that's not as good. Right? But the truth is, all three of them from a standing start have built companies worth plus or minus a trillion dollars, and Architon are riding this trend, and no one else is even close. I mean, so if you look at the report card, you know, yeah, you're right. Anthropic did an amazing job of building the model, which ironically meant that their CapEx forecasting was wrong because they underestimated the success of their own model. To some extent, OpenAI has been the other way around. They really nailed it on CapEx. They nailed it on compute, compute, but maybe should have spent a little more time, you know, on the enterprise features of

SPEAKER_01

their model. So they're behind there. And obviously, you know, if you look at the X.AI part of SpaceX, leaving aside the thing, just focusing on their AI business, is they're the world's best builders of, you know, fast data centers. But so they get the points for that. But on the other hand, they lose points because they weren't able to fill it with their own stuff. And you're right, it's not quite perhaps as modern or as SOC 2 compliant as some of the stuff OpenAI got. So yeah, you can make knocks on all three, but zooming out from a standing start, they've each created a trillion dollars in AI in the last five years. You know, what did you do on your summer vacation?

SPEAKER_01

Well, and Elon founded both. Elon founded both of them. Oh my God. You have to remember that. I mean, yeah, the entire pivot. Plus Neuralink, plus he's controlling people's bodies with his brains, plus the Hyperloop,

SPEAKER_00

plus Tesla. Yeah, because we can, well, we can talk about that the entire, because the SpaceX,

SPEAKER_02

when we come to talk about SpaceX, the proportion of their future value they attribute to AI. It seems to me, this is Elon, you know, effectively losing OpenAI and says, God damn, I'm going to have one of my own and made it happen, realized it from whole plot. Well, let's discuss SpaceX directly then with the S1 dropping. Rory, given the fact that I didn't correctly identify the most important numbers, what did you think were the most important numbers when you looked at SpaceX's S1 dropping? Cynical answer, I don't think the numbers matter a damn, right? Because the point is this, everyone said the same thing. You run the sum of parts on those three businesses.

SPEAKER_02

There's a lowest growth, but tech enabling, amazing space, space business. There's a good growth, profitable, Starlink business that has much, you know, a bigger TAM than the simple rocket business and the launch business, right? But it's growing 30%, 40%. And then there's X.ai that at the time of filing was a great big gaping hole of 15 billion in CapEx and no obvious revenue. Since then, obviously, he's done two amazing deals to find the revenue, and now he's got a $15 billion run rate business. But you add it all up, right? And the truth is, you get to a launch business, you have a Starlink business, which you can value, and you have a core weave type business,

SPEAKER_02

because that's what the AI business has now become. And the sum of those parts is so far lower than the proposed valuation. And the only difference between the two is the Elon premium, and that's not a number that comes out of the prospectus. So I looked at the numbers, I read them, I understood them, but they offer me no value in trying to value this company, right? I mean, you know, other than saying utterly stupid things like the fundamental value of Tesla is 200 billion, and the trading price is six times that. So maybe the Elon premium is 6x value. I don't know. It could be the GeoCities deal of the AI era. We could look back on this. And listen, I'm all I'm

SPEAKER_01

on team Elon, okay? But 100 times trailing sales, we may look back and say these were some good companies. But 100 times, my God, my God. No, no, no, no, no, no, no, no, no, no, no, I'm going to defend. Like, no matter what happens, these are three amazing companies. Twitter's, Twitter's just, growth has fallen off, it's shrunk 50% since he bought it. Elon's the magician of revenue compression in Twitter. Revenue has fallen 50%.

SPEAKER_00

Yes, Twitter, agree. Twitter is, I mean, again, if you disaggregate the three businesses, the launch business is a good, stable, boring business with amazing technology. Barely growing.

SPEAKER_02

That probably would get tapped out at 10 to 20% growth rate. The Starlink business is a great business with a much bigger TAM, but 30, 40% growth rate, 10, well, think about 14 billion in revenue and EBITDA positive, right? So actually, I will give you the most interesting chart, Harry, and the number in the thing with the TAM analysis. And I want to say for memory,

SPEAKER_01

it was something like 28 trillion, largest TAM in history. That's actually not as interesting. The interesting part is, even though, let's say 10 years ago, this was a launch company, and eight and two years ago, this was a launch company with a communications business. 90% of the TAM that they identified is not in those two sectors. It's all about AI, right? So the interesting thing is the story they're trying to tell is a story that's 90% not focused on the two things where they're unique and differentiated, right? I think they're not unique and differentiated in the AI story because Grok as a foundation model has gone nowhere.

SPEAKER_02

They did make an enormously clever deal in the sense that they had Colossus. They built it quick. And again, the S1 is very clear on that. We think we build faster than anyone else. And they've sold that to Entropic. They've rented that to Entropic for 1.25 billion a month,

SPEAKER_01

right? With, I think, a 90-day cancellation clause on either side. So basically, it's $15 billion a year, right? In one sense, that's amazing because it's amazing to think that you had an $18 billion business built over 20 years, and then you just added $15 billion to that in one transaction on a product you only built a year and a half ago. So in one sense, it validates the comment that the AI business is so much more dynamic even than the launch or the comms business. If you're willing to spend the capital up front and invest $10 billion in a hole in the ground, and the compute demand is there, you can get to $15 billion in revenues pretty quickly.

SPEAKER_01

But what you are is a more efficient core wave. Now, they will say that- That's worth $2 trillion? That's my point. Renting out chips because Jensen's your buddy and you're a good customer, that's worth $2 trillion, $3 trillion? Or are the rockets growing 10% or the Twitter declining 50%? Which of them exactly is worth $2 to $3 trillion?

SPEAKER_02

And that's why I jumped on the Geocity thing. I love the S1, right? I think it's all madness. I wouldn't buy a share. I just love the optimism. I just love the fact that they've done this thing for 20 years. And some of the assertions of why they're great were totally true. We've been fairly capital efficient for people who put rockets in space. We build faster than anyone else. We have organizing principles around engineering that, you know, the algorithm to make things efficient. There's a lot of balls-out claims here that are justified on the basis of 20 years achievement from probably one of the most talented entrepreneurs everywhere. It was a great read.

SPEAKER_02

Go America that we can have this, right? So after I say all that, you're right, Jason, I still don't get, I'm so far from $2 trillion in DCF that my head hurts.

SPEAKER_00

Well, you know, it's almost worse. The 100x trailing revenues is like, it's hard to tie,

SPEAKER_02

right? But the real, and listen, I'm on Team Elon. I've bought five Teslas. I've got like three Starlink subscriptions across. I am from the very early days. But in some ways, this is just very cynical. This is financial engineering, which I have mad respect for. But my God, you're taking a bunch of disparate assets that most of them have no connection. You're throwing an S1 talking about the history of AI, which wasn't even your business a year ago. A little bit of financial engineering, I'm all for. But man, this is so much fun. This is bailing out my failed Twitter acquisition,

SPEAKER_02

my idea that I'll compete with open AI just by buying chips. That didn't work. I'm bailing both of them out, right? I had a quiet business, which I should have kept private with SpaceX, right? It was a great private business that blended, was doing great things. Now I'm bailing everybody out in this massive confab. It's SolarCity on steroids. And we're all here for it because we love AI, but it makes no sense. This conglomeration of friends of Elon makes no sense to anyone but the folks getting bailed out on Twitter. It just doesn't make any sense. The answer is the launch business enables Starlink, right? Yep.

SPEAKER_02

And the launch business will enable data centers in space, which is how, and we've proven that we can build capacity on the ground. So therefore, we'll be able to build capacity in space and we'll be able to get 100 gigawatts of capacity a year, five years from now. Data centers in space is the thing that joins all the dots together, which is different than me saying, even for a second, I ascribe a high probability to that. There is a coherence to the narrative, but only if the next thing happens. And that's always the way. If you buy data centers in space will work, then it makes sense to have a launch business in the same company that has a data center business.

SPEAKER_02

And the tweets, where do the tweets fit into all this? The tweets don't fit in at all. Let's get real here. The tweets don't fit in at all. And the truth is that, you know, the 44 billion on Twitter was a miserable acquisition and the value of that company is less today than we bought it. And that was pretty obvious two years ago, right? He's chosen to bail everyone out by rolling it into X.ai. He's chosen to bail X.ai out by rolling it into SpaceX. And to be fair to X.ai, even though we just use the words bail it out, if they can keep the $15 billion a year, they're going to cut. That's a very high. I mean, the total

SPEAKER_02

CapEx in X.ai over the last two years, I think it was $12 billion and $7 billion, $19 billion, right? One year and a bit of Antropic revenue at $1.25 billion a month, and you covered your nut, which is a pretty high cash on cash return for a data center project. So look, I sat there thinking it's all crazy too. And then I'm like, he built something right at the time when capacity is at a premium and he's, you know, he's found a way to monetize it. I mean, you know, again, Jason, I repeat, I don't get to, I mean, even, so you apply the core weave multiple and you get to under $100 billion.

SPEAKER_02

I'm not sure what to make of that information, but it was definitely a get out of jail free card when he got the Antropic business. By 2030, what will that core business be? Will it be Starlink? Will it be data centers in space? Will it be NeoCloud? Good question. I think Starlink will be the vast portion of the value. The value, I don't think data centers in space will be a meaningful percentage of revenue. And I think their existing data center business will be a, you know, high revenue, relatively low return on equity business. How that compares to Starlink, TBD. But I think we'll

SPEAKER_01

discover that, as I say, it's a very, it's a, it's core weave with the world best engineering and construction team. So implicitly, I'm saying not a whole ton of data center in space revenue on top. And that will, that will only be worth anything like the current value if the narrative is tracking towards data centers, which again means it goes back to the first sentence of the show, which means if Jensen is right that you're spending $3 trillion a year and the ROE is there for the full $3 trillion, as Jason has pointed out, and if we can't build shishet anywhere else, then at some point you'll want

SPEAKER_01

to stuff them into space regardless of all the pain and suffering involved in doing that. And then Elon will be right. If any of those if clauses turns out not to be true, then it'll be a bridge too far.

SPEAKER_02

Or more correctly, a rocket too far. But who knows? I mean, if the one human being on the face of the planet who's earned the right to say, give me a trillion dollars, I'm going to make the bet is Elon. And if you want that bet, I'm glad it's going to be available for people. If you want to scratch that itch, go scratch the itch. Look, I think he made a compromise to get it all done. I think he's going to have to build a $50 billion core weave business to make the math tie. And I think he had to give up on going to Mars and go to the moon. Like he got, we reached a certain point in life where to achieve our goals, we have

SPEAKER_02

to be practical again. He's got into a practical phase. He's got to take this thing public. He's got to mash it all together. And he's got to, you know, the anthropic thing either has to die or keep going, right, to make this work. So he's got to commit to building a 50 plus billion dollar core weave and going to the moon, which is pointless, but it doesn't mean it won't get him there in the end.

SPEAKER_01

He could also, if the world changes, I mean, we're on such a intense trajectory, right? But if the world changes, he could also dump this whole core weave business and just move on, right? If it is possible, you can write off the chips, you can write off, you can write it all off. If the world goes a different way than we expect.

SPEAKER_02

And proving how little we all know at this point, how little you can know, let me get the other option is, if everyone else proves incredibly mediocre at building new data center capacity, and anthropic keeps continuing to grow, then, you know, even though they, I was thinking they did this like 1.25 billion a month, it's kind of like we all buy storage capacity. You know that when you buy kind of, you know, storage for your stuff, when you send it away to storage, you extra furniture, and you pay 100 bucks a month, and you think I'll be out of this thing in six months. And then five,

SPEAKER_02

five years later, you're still paying and they're now charging you 400 bucks a month, right? But basically, this is public storage, but for compute, and he could well be there for the next five years when anthropic just needing the capacity and making a ton of money. Consumer storage solutions, one of the best businesses. Yeah, yeah, exactly. This is like that. And you go in, I'm sure anthropic went in and thinking, we're only going to pay him 1.25 billion for four months. But they're relying on other people building data centers, and life is tricky. And not everyone will bulldoze a true like Elon did. And a year or two from now,

SPEAKER_02

because remember, this is very profitable revenue. You cannot take it away in his CapEx, in a shitty CapEx business where Core Weaver is a good company, but makes only okay profits, at 1.25 billion, it's a great deal, right? Relative to his CapEx. So his ROE on that is pretty damn high in a low ROI business, return on equity business. And don't forget, it's a poor analogy, but it's the same guy. There's no one in the US or Europe that can produce EVs remotely as efficiently as Elon, like the investments he's made. And I'm not saying it's the same. He can't control the whole supply chain, but the man has a history of doing things

SPEAKER_02

that are both at the cutting edge and radically more by investing hard, like Colossus and all of this, by doing things that are radically, maybe he does build his own massive fabs in the US and they work and we can take shots at it like we did when the Tesla Roadster came out, but no one can, outside of China, no one can approach what he's done in EVs. Maybe it happens in data centers. I mean, Core Weave has nothing. What do Nebius and Core Weave have? Nothing. No special technology, right? He could accelerate past everybody. He has the, he's got the capabilities, right? And we could

SPEAKER_01

turn around in five years and say, my God, he owns every data center in the galaxy. Because, I mean, if nothing else. Who else? You can't compete with a Model 3. A Model 3 you can lease, Harry, in the US for $299 a month. It's self-driving, it needs no fuel, and it's better than all its competition. And the only reason we don't all drive them in the US is 60% of the country hates them. Otherwise, there would be no other car in this country, right? He could, what if he does the same for,

SPEAKER_02

it might be with Colossus and friends and everything. He's actually doing this for real. Because remember, the one thing you need to be successful in a CapEx intent of data center

SPEAKER_01

business is a low cost of capital. And I think we can agree that a $2 trillion pre-money, that's a pretty damn low cost of capital. So yeah, no, he can sell. I mean, if he gets $75 billion, he can put $25 billion into another Colossus. Sell that for another $15 billion. Away you go. There are two private market stories that I want to touch on that we haven't touched on. One, AI Slop, spelled backwards is Pulsier. You're like we're being pumped. Listen, it is a company that enables a single person to build a business, and that business is then run by AI. It is bluntly not very popular. It's sent out a huge

SPEAKER_01

amount of emails unwanted, which is why a lot of people don't like it. The founder of OpenClaw, Peter Steinberger, even responded to the funding announcement saying how much he disliked these solicited emails. But they raised 30 to 40 million bucks. I can't remember the exact amount. I had a $250 million valuation, some pretty well-known firms in there. Jason, I'm intrigued. I think you'll have a thought on this one. I feel like I could do it. I feel like I could just call it AI Slop instead of Pulsier. I feel like they can raise the 250. I feel like the only question, if I did this, could I take the 40 million out of secondary? Is that okay? Do they care? I would

SPEAKER_01

do this if I could do it all. I mean, listen, I mean, it would be nice. Unfortunately, I wouldn't get QSBS on it, right? Because it would be very, that's the negative, but I wouldn't mind taking 40 million. I think I could do 10 million like this if I knew I could. I don't want to raise from VCs and be stuck with the moral obligation of turning it into 2.5 billion. But I don't know. Listen, I don't want to, just the fact that you have a startup named AI Slop raising it 250, if this is the peak, we're going to look back and make fun of this one. But I don't know enough about it in all fairness to be sure I'm wrong on this. It's just, I feel so punked by a startup

SPEAKER_01

called AI Slop raising it 250. I just feel like this is the, this is a bigger FU than Maltbook. Maltbook pretending these agents are talking to each other in a social network when it's all humans telling the agents to go onto Maltbook and talk to each other. I just don't know, is this the next Maltbook? I just, I literally don't know, but I would at least have asked them to please change the

SPEAKER_02

name. I did admire the tweak though. I mean, it is just such a two fingers salute to the great AI marketplace. I'll tell you what I don't like about it. There is, and I'm, listen, we can talk about what it's actually having just begun to build an app while we're on this Zoom or this, whatever. Yeah, it's pretty good. Like from a marketing perspective, it's pretty good. So I asked it to build an inbound AI SDR for me. I've already done this and bought and used a bunch of these tools. The way it qualified and spec'd out this product, pretty good. Pretty good, right? Certainly better than Replit or Lovable, which are horizontal tools. The journey it guided me

SPEAKER_02

through, it did what make couldn't do while we're here. It researched sastardoc.ai. It figured out what I did. It analyzed what tool I would likely want to build. It spec'd out the tool, did a great job. Like I give it a 10 out of 10 in terms of the journey so far. And then it immediately asked me to give a credit card and pay $49 a month before I got any value. My general experience in agentic tools like this is that when they ask for money before they deliver any value, it's not that good. Like the marketing exceeds the value of the tool. I can't think of an exception. I'm not saying they're

SPEAKER_02

out there, but every time I go through one of these, I immediately try to figure out where the cancel link is and it doesn't exist. So, but I give them a 10 out of 10 for the, for the, for the sort of before I put in my credit card journey, it's pretty, pretty good. Like they, they've definitely, they definitely make you feel like they can build something a pretty, pretty bad-ass for you. I have to say one thing that is impressive. I met them as part of the fundraise to be very candid.

SPEAKER_01

It is one founder. There's no team. There's no them. You're right. It is one guy, right? It's one guy, right? So it's pretty, I'm impressed with the aesthetics of it. I just, um, and I know a lot of founders disagree with me this day, a lot, and a lot of accelerators tell you to charge instantly for these genetic products, but I know the ones that I love, right. That I use every day, they give you value before you have to put in a credit card. And this is going back to the, the Y combator 2 million of tokens, invest $5 in me or

SPEAKER_02

a dollar of tokens in me or $2 to earn 49 bucks a month, which is the entry points for the, for the product. You can't invest two bucks in me. Um, disappointed. Right. Um, but I get, I get why a lot of young founders or others think it's not worth two bucks, but to me, I'm out. Boys, what other stories have I missed that you think we should cover? Yeah. Extra raised $250 at $2.2 billion to build the search engine for AI agents today. Open Router announced their round, which was, I think, $150 million at $1.3 billion led by capital G. Manus founders are trying to now buy back the company and quite a contorted deal.

SPEAKER_02

Yeah. I mean, I think, look, you know, what all these have in common and the first two have in common and the third one's obviously very different is look, super interesting kind of infrastructure companies all about the build out in AI. Right. And the big picture here is, you know, you know, enterprises are adopting AI that, you know, yes, they're using LLMs, but there's a coterie of other tools that developers and people need to kind of build great agents. And, you know, X is doing, I mean, you know, just what X does is around search, because if you think about it,

SPEAKER_01

if you remember back when ChatGPT first came out, it was dumb about anything other than what it, you know, what the model had been trained on. And then people like perplexity add internet search. So now when you do a ChatGPT query or a perplexity query or an entropic query, it queries the web at the same time. So it knows what's going on. Well, in the same way, if you're building as an enterprise, building an agent of your own, and you have access to the OpenAI API or the entropic API, what you don't have, what you also will want is some version of a infrastructure and web search, either across the whole web or across a set of kind of defined

SPEAKER_01

websites, so that when you build your own internal agent, it's a smart agent that knows what's going on and executes to that need. It's done really nicely themselves in parallel are the two companies doing that that have recently raised. And my point is merely, there's just a whole bunch of infrastructure going on one level below the foundation models that are just super interesting. You know, we talked about OpenRouter before, I think, which is a company that allows their customers to switch between various different models, both the foundation models and the open source models. They have, I think, 50 plus models hosted on there to allow, obviously, enterprises to

SPEAKER_01

access the lowest cost model. And these are all just interesting trends because, you know, I think the OpenRouter one is interesting because, going back to what Jason said, price cut. If people are spending $300 million in a premium product and having to lay off people as a result of that, there's going to be some interest in exploring cheaper costs. And I think companies like OpenRouter, the value proposition is, hey, not every query you need to make needs to be done on the most expensive foundation

SPEAKER_00

model. And there are other places to do that, right? So I think, again, it's just that the cliche, the picks and shovels of the agentic revolution is just a good place to be investing. And I think it will be going forward, like strong, interesting teams doing good stuff. You know, you can squint one way and say foundation models hoover it all up, but I don't think they do. I think there's a lot to be done around building your great next generation agents. I mean, Jason, do you guys use Parallels or

SPEAKER_01

Exa or what for WebSearch when you're building your agents? Yeah, and I'm a small investor in Exa. It's just a great product, right? It's another one where I was a user before an investor, right? An early user because, and the reason I think it's super interesting, even more interesting than OpenRouter, right? Which is a great one is OpenRouter is really

SPEAKER_00

interesting because it's a, well, first of all, they nailed the ability to dynamically and easily pick your LLM, right? They nailed something. And so it's a proxy for all these discussions, right? It's a great proxy insight. The world needs that product. Exa is more interesting and some of its peers because these are products that have no use without agents. And when we throw these term agents around, but agents have a lot of meaning, right? But literally, if you're not, at least as far as I'm aware of, if you're not building an

SPEAKER_01

agentic product, you have no use for a tool like Exa, right? Agents don't hop on Google and do search.

SPEAKER_00

They don't. And they don't create files on Dropbox and they don't hop on Zoom. Like agents have a whole, true agents have a different set of workflows and tools they use. And Exa nailed one of the core, it wasn't even obvious to me until I started. There's an obvious issue. Like your agents do need to find current information. Like it's just that simple. Just like we, Google was the killer use app for humans, right? Your agents need it, but they don't need Google. And putting it aside whether Google should have built it rather than promote Exa, which was a ball bouncing the right way,

SPEAKER_00

to see the type of growth, even at the early stage for something that agents need and humans directly don't, it's, it's a look into the future. And then it also says that a lot of stuff isn't BS, that a lot of this stuff is real. And that we, we may, the idea that we may all manage more agents than humans, isn't just X BS. It's real, it's real. And they needed a whole different set of tools.

SPEAKER_01

They don't need Zoom. They don't need a traditional CRM. They need tools like Exa. Maybe we should only be investing in those things and leave those human tools behind guys. Enough investing in human software. A very dangerous thing to do. Can you paint to me the upside case here at 2.2 billion? What is that upside case from here? Is this a a hundred billion dollar company, which powers the future of agents? And what is the market composition of this market? We've got Parag at Parallel doing it as well. Is this an Uber lift where one of the providers takes 90%? Is this an AWS Google Cloud? Just help me understand that.

SPEAKER_01

I think it's more the latter than the former. It's not a, it doesn't have, because it, you know, it's a developer tool. There's not huge network effects. So you can easily contemplate two companies being pretty successful here, right? Which slight, I mean, we did some work on this. We looked at Exa back in the last run, I think super team and good product. We, someone, someone bit higher and they won the deal and great name to benchmark. So what can you do? Right. I actually think it's a super good company. I like the team. I think they're smart. I think the market's there. And I don't

SPEAKER_01

think it's a winner take all. I think it'll be a couple of players because it just doesn't pay, you know, you're not going to have 10 because it's not economically efficient to have that. But at the same time, it's not like Google where the winner take all emerged because it was a consumer product. And then the ad network and the, the kind of structure of the market for ad sales kind of pointed to a winner take most. In this case, individual developers will be making decisions. So I can imagine comfortably two players here dividing, you know, reasonable market share, just like most developer tools. I think next level, how big can it be is the question. I think

SPEAKER_01

Jason's right. It's all about the more you believe in agents, the more you believe they're going to want to access structured information. And this thing's going to co-attach to all that. So that's the bet here, right? You are betting on lots of agents needing lots of information, doing lots of reasoning, you know, just doing lots of search. So it's a derivative bet on agents. And right now, that's been a good bet. You know, I think, you know, how you think about getting to a billion dollars in revenue, which is probably what you need to be in a normal market to be worth 10 billion in,

SPEAKER_01

you know, market cap. I'm just thinking about it on the fly here, because I remember we looked at the numbers a year and a half. It's not crazy in a world. Look, if we're spending, going back to, going back to first principles, if we're spending 3 trillion on CapEx, right? That's the Jensen comment. Then the second comment is, if you look at the, if you look at the SpaceX S1, remember I said that 90% of the market is AI? One of the next interesting thing is within that AI, they said 90% of that is enterprise. So they're basically saying the vast bulk of AI spend is going to be enterprise building

SPEAKER_01

shit. Right. And if that's the case, and they're building enterprises to frankly replace human work, then I think there's a pretty compelling market for the product that allows structured information to be accessible to that agent. I think you'll expand beyond just web search. I think you'll do curated lists. I think you'll do internal information, but there's totally a need for that product.

SPEAKER_00

That's fucking wild. Another fucking, another benchmark company. This fund is like going to be one of the

SPEAKER_01

best performing funds in history. I mean, seriously. Yes. Good for them. That's the job. That's the job. Every company, Jason, like the number of hits is extraordinary. Yeah. And like, and so they did exit 700, right. Or something like that. Right. I think, I think you're right. That's the, this is 20 VC. So, so it'll be an Epic fund. I think also going in early and realizing that investing in a smidge of traction and breaking your rules a little bit was the right bet for today. Arguably, I think was maybe another insight too, right. You know, Jack Altman just joined, right. And he just did Monaco, which I know, well, I'm a small shareholder.

SPEAKER_01

And, you know, they marked up the deal very quickly, almost two X from founders fund. But when founders fund invested, it had no revenue. So he got a much better deal for benchmark by investing, you know, the, the, the markups a couple of months later, but the, but going from pre-revenue to blowing up, right. That's if you can get into the moment an AI leader blows up that, that hour, like that's when you want to invest. Right. And so then they've got, they've tripled their, their eggs evaluation. And I just think this going back to seed is for suckers. Like you want to invest the hour it blows up. The minute it blows up, you want to get the DM and just wire the money

SPEAKER_01

in AI. That's the play. Pushing on that a little. Cause first of all, you got to give huge credit then to light speed for doing the seed in XR. Yeah. For suckers. Well, I mean, if a sucker means you won a lot of percentage at low dollars. Yeah, no, but I think what you're saying is actually, so first of all, I just wanted to give that credit. But second, I think what you're saying is correct,

SPEAKER_02

which is there's a quantum reduction in risk when you go from no revenue to revenue. And then there's a linear reduction in risk thereafter. So the sweet spot, I always think on the risk return continuum is early product market fit, which is why we focus on it. Now you're right, Jason, the, the, the, the amazing thing about these AI companies is that moment used to be a year, year and a half, maybe two years of, you know, early product market. Now what you're seeing is you go from early product market fit to, Oh my God, it's incredibly obvious. Just given the take rate, literally sometimes in, you know, weeks, if not months and definitely not years. I mean, you know,

SPEAKER_02

you know, the lovable and replet numbers. I think the same thing kind of happens here. You got to find that moment and yeah, you got to pay. I mean, you know, in retrospect, I was wrong. You should have, you shouldn't have bid what I bid. You should have bid 750. Yeah. But also you have to abstract away. Not only does it happen so fast, but the counter, the counter side is you have to abstract away some of the competitive risks. You have to not worry that Google could, why didn't Google do it? Right. Because otherwise you'll do none of these deals, even if Google wouldn't do it,

SPEAKER_02

the company didn't even exist a year ago in its current form. So what are the renewals going to look like? Like we don't know any of these data points, right? I mean, I agree. This is actually

SPEAKER_01

an interesting discussion about our job as this thing from being public market pontificators. You're right. You have to decide that you're willing to pay more on less information than the SaaS era, simply because the market and traction that you're going to see if you're right are so much, come so much quicker and is so much bigger. That is the core of figuring out what's going on right now. And that's been the bet that's worked. Right. And I look back on the stuff we did that I feel right about. Those are the criteria. And then obviously you have to get the traction. If you don't, you feel

SPEAKER_01

like a moron. Right. And on the stuff you missed, you look back and you go, you should have leaned in even more. You should have been more willing to give conviction on the traction, even though the data points were sparse because the stuff that's working, that that's in tune with what people are trying to build is just, you know, it's the universality of the propensity to adopt right now. There's no 10 year cloud journey where some people decide now and more people decide two years from now. It goes back to the first thing we said, corporate America flipped, flipped the switch a year ago and

SPEAKER_01

said, thou shall do AI in 2026. And if you're selling anything on the picks and shovels business to help corporate America do AI, you get revenue this year. You don't get the luxury of waiting.

SPEAKER_00

Oh, I'll wait to see one year's renewals before I bid. There's a thread, which is, can you make, if you have a gentic products, can you make up in volume? So there is a software

SPEAKER_01

explosion. The number of apps we are building, it may not be benefiting Dropbox, right? Where Drew just stepped down, but overall the number of apps is exploding at a rate we've never seen before. So apps are being built everywhere. Workflows are being built everywhere. Agents are being built. If you consider an agent, a piece of software, it's exponential, right? At Databricks, Neon, which is their Supabase competitor, over 90% of the databases are built by agents, not by humans, okay? But they don't pay as much as they pay Mongo. So the question is, and I don't think Exa makes anywhere near what Google makes on a search, right? And so where I

SPEAKER_01

think we're going to get caught in a lot of these tools is where there aren't orders of magnitude, more volume. And the reason Supabase and Exa and these groups are neat is because you don't need them for every app you build or every use case, but you sure need them for a lot. You sure need them for a lot. And we're going to make mistakes as investors. I know I've already made one or two and they're not fatal mistakes, but where it all, it's an agentic workflow. Agents need it, but they don't need it all the time. And so I can't make it up in volume. Like we need vast volume for these plays to work, vast.

SPEAKER_01

You're right, Jason, because you're right. The dollars per search of the whale. And basically what you're saying is, you know, what is the, what is the, call it the harness, call it what, what are the core primitives that it takes to build an agent? And what are the other things that, that pretty much most enterprise developers will figure out, oh, I need a, this, a, that, and the other to make my agent. And if you pick the thing that only 10% of developers need, because it's corner case, you'll pay up and you'll have a tiny market. You'll be sad. But there probably is going to be five or six things like a database, like a search engine,

SPEAKER_01

like some kind of observability where, you know, over the next two or three years, revealed preference will say, you know, the vast majority of developers realize they need this. It doesn't come in the box, in the harness from Antropic or OpenAI, and therefore standalone companies will be built doing it. You're right. And we just actually had that discussion last week. What are the five or six things that you need to own? Right. And I think it's credible argument that agent search, which is XI. And actually, I think I could be wrong on the open router thing is that kind of a model switching layer could be two of those things that every, when you look back five

SPEAKER_01

years from now, every app company, either building a third party app or building in house said, oh,

SPEAKER_02

I need that product. So they end up with, you know, mass developer adoption. That's the best. Okay. We're going to do a final round, which as always is rage bait, but real. I just tweeted and me and Jason were talking about it. I just interviewed a CEO who said three things. We were number one, we replaced our 600 K Salesforce contract with a vibe coded CRM, which we built in three weeks. Number two, we will get rid of 80% of the SAS we use internally. And number three, if anthropic doubled pricing, we would not change our usage. Which do you think is the most rage bait statement there?

SPEAKER_02

I think the first one's the first one is Clark ask race rage bait. Yeah. And it's unhelpful. It's unhelpful. If you want to build your own CRM on top of Postgres, go for it. Go for it. We, we moved to headless Salesforce. Okay. We don't log into Salesforce anymore. We could, we're a small enough organization. We could swap that out for our own database. It's not worth it. It's not worth maintaining that database. It's not worth fixing the connectors

SPEAKER_01

to our 10 external agents, which all natively built, built into Salesforce. It's not worth giving up agent force, but if we needed no third party apps and we had time to burn a sure, because we already have our own layer, right? We already have our own autonomous agents running on top of Salesforce, but this is such a waste of time talking about vibe coding away Salesforce. And we have more important threats to our existing portfolio than this one, or more important things to invest in. So I just think it's, it's a dated 2025 take. Even if there's truth in it, even if there's truth in it, his statement was actually that the functionality was so much better

SPEAKER_01

in what they were able to custom build when the integrations that they needed, given their healthcare focus company, that actually it was worth it to build their own and that they were able to do so much more given that. It might be if you have extreme vertical focus and willing to have an internal CRM that is not collaborative and you have a relatively small team using it and the benefits of knowing of a traditional CRM interface and the thousands of apps that integrate with them are zero to you and you have time because you don't save any money. This isn't worth it for the money. If all those things

SPEAKER_01

are true, vibe it away, dude, vibe it away. But this is of all the threats there are to Salesforce and HubSpot and Monday and Atlassian. This ain't in the top 10, right? That's why I think it's rage bait because it matters, but it's not, it's not, it's not important in today's world. And it's not even the SaaS, it doesn't even explain the SaaSpocalypse, right? Other things explain the SaaSpocalypse, right? I like that answer on the first and agree. Yes. If you can get better functionality for knock yourself out, but vibe coding, that's a silly. I actually want to take the third one and

SPEAKER_01

actually put it back on the CEO. What he basically said is if Anthropic doubled their prices, I wouldn't mind, right? Then you should be ashamed of yourself. You know why? Because what it basically says is the ROI on this AI is so good, I could pay twice as much. Well, in that case, let me tell you what you should do right now. You should go in and tell your people, use twice as much because logically there is some kind of, there's an ROI on every token you spend and the early ones are really great. And eventually you probably hit the marginal cost is exactly equal to marginal advantage. And then you're overspending. And if you're token maxing,

SPEAKER_01

you're overspending, right? If you think you are willing to pay twice, if the projects that you're doing right now with AI really, not just rage bait, but really have a 2x return, have a return that's 2x better that could support twice as much the price, then what you should say to your engineers is you're not using enough, go do more, right? Because the two things have to be true at the same time, right? If the, if you, if, if Anthropic could double your prices and this project is still economically viable, then what's the next project on your to-do list? Go do it. I'm with you. But the company Harry's talking about is at 2 million in revenue per employee,

SPEAKER_01

so they can afford it. Let's be clear. They can afford to spend more on tokens. You're right. Of course you're right, Rory. If you're that efficient and you don't care what it costs, spend more, right? You haven't reached that. But I will tell you because we're there, I don't care what we spend on Anthropic either. Okay. Our AI VP of marketing customer success costs $257 a month. I don't care if it's 500, what's 257? We're getting toward the end of the pod. It doesn't matter. It doesn't matter. But here's the, here's the interesting thing.

SPEAKER_02

This is why I actually think it's interesting is we're all, if you're really good at this stuff, the issue isn't more money. The issue is idle, idleness. Our agents are idle. We don't have enough brain sight. Like this is why these, these, what's the, what's the term for folks that can't sleep in San Francisco? They're, they're, they're coding all night. They, they have a, what's, what's the, what's the term? I forget what it is. I could be cynical and say incels. No, there's this term where you're like addicted to it. Like your, your, your brain, your, your brain's been rewired. What the problem is with like, listen, there's the Uber issue where we spent too much

SPEAKER_02

on slop. Okay. Or we don't know what to measure. There's a different issue coming, which is coming for

SPEAKER_00

more of us in tech, which is that our humans just can't process all the output. It's not even about the money anymore. I, if you, if you wake up and build me 50 features every single day, how many features can I even qualify overnight? Because we wake up every day, Rory, and our AI VP of marketing nags the hell out of us with three ideas we should be doing every day. It's exhausting. You can't

SPEAKER_01

implement 21 ideas a week, right? Every day, 7, 13 AM, three ideas of the day, 10 K comes up with and pushes. You didn't get to my idea yesterday. Did you now you're behind, right? But stop. You're right. I agree with what you're saying. I just still go back to my comment though, is that it's a, I mean, that, that, that paper I cited is that, you know, you can increase the complexity of the task and spend more money because he's implicitly saying he's getting, he's getting way more value than he's paying for it. Right? So when you look and you're, you're saying the same a 250 bucks, which is nothing. You're saying you get three great ideas a day and you can't process

SPEAKER_01

them. Correct? Yeah. We get more than that, but in addition, we get nagged. We got three things we need to do. We're going to end on a note of optimism. In that case, you need to hire another human. Oh, I agree. Because if the AI is giving more good ideas than the humans can process that, and this is kind of a whole nation of compliments and substitutes. If the AI is now giving more ideas than one human can process and they're good ideas, hire a second human who will be now more effective and the return on human capital will go up. But you know, you're right, but you know what the

SPEAKER_01

problem is to tie it all together. I desperately need a human for those ideas, but a 125 K laid off click upper is worth nothing to me. I need a million dollar person to process these ideas and they need to turn that into 5 million and I will hire them today for a million bucks. No joke, but you got to be the million dollar person. Not think, not want to be that person. You got to be that person. The odd thing is every single discussion here has gone back to, you know, what's the optimal span between humans and tokens and you know, what are the consequences of that configuration? That is the same question the whole time.

SPEAKER_01

Jason also does not like other humans. And so that would be I do. I don't know why people say this. I'm a people person. That's what they all say. I'm a people person. As long as I don't have to interact with too many on any given day. Jason, I hate to be with you. If you have to say you're a people person, you're not a people person. I know. I know you're right. You're right.

SPEAKER_02

If you have to say you're right. Do you want me to be honest? I am not. Whatever the two by two is, Harry and I are at the opposite ends of whatever the two by two is. Okay. The extroversion, people person, everyone's Harry's best friend. I never get that. Why don't I have to say on social media? I was just in London meeting with my best friend, Harry Stebbins. I see like 11 of these on Twitter a day. Because you're not insincere like Harry. I see. That could be part of it. That might be part of it. Yeah. Yeah. Because I'm my big man. When I'm in, when people are in town, they come see me. But they're always my best friend. How many best friends can you have? 300?

SPEAKER_02

Oh, many. The Dunbar number is 160, just for the record. Best friends? Oh, dear.

SPEAKER_02

Yo, boy. Yo, boy.

SPEAKER_01

! Yo, boy.