The Twenty Minute VCWhy VC Today is Worse than 2021
EVERY SPOKEN WORD
100 min read · 19,685 words- 0:00 – 0:53
Intro
- RORory O’Driscoll
It looks like the easiest way to make money in 2025 is to take the very biggest companies and double down one more time.
- JLJason Lemkin
My gut tells me we're over-romanticizing verticals in the age of AI. Like, we're gonna hit the same TAM exhaustion, and it's gonna be worse because expectations are so high.
- RORory O’Driscoll
If any kind of deacceleration happens 'cause of, you know, any kind of saturation or slowdown, everyone's estimates on what's gonna happen here are wrong. Now, again, reminder, in a bull market, the most aggressive person will look the smartest just before the crash, because the more risk you've taken, the more money you've made. Frankly, it feels tough today, as tough as it's ever been.
- JLJason Lemkin
Ready to go? [upbeat music]
- HSHarry Stebbings
Team, it is so good to be back. It's slightly earlier for you.
- 0:53 – 2:16
Everett Randle joins Benchmark
- HSHarry Stebbings
Topic number one, Everett Randle joins Benchmark. Benchmark don't add partners very often. It's a big news announcement that he's joining. Joins Benchmark as their latest GP. He was with Kleiner and he was with Founders Fund before. Uh, Rory, I always think of you as Benchmark because you quote a fantastic statement being, uh, "Reports of my death have been greatly exaggerated." And I always think of Benchmark with this, with the portfolio, with great people like Everett.
- RORory O’Driscoll
Totally. Yeah, no, we were having that, you know, overall, "Oh my God, the world is ending," two months ago when, you know, a Benchmark partner opted to do his own thing. And I remember saying exactly that. They're gonna be totally fine. They have a great portfolio. They have a great tradition. And they did exactly what they sh- what they always do. They went and made... I, I can just see it there. You make a list of top firms that have good young people. You draw a list and you get on the phones. You say, "Who have we overlapped on a deal with?" And you go and you hire someone talented from one of the adjacent golden firms where the pitch is purely equal partner. You should do this. And mission accomplished, tick, done, and, um, on they go. So yeah, it's, it... [sighs] It's not that it's not a big... It's like they're just fine and Kleiner will be just fine. You know, Mamoon and Ilya, wildly talented people. There won't be any shortage of people if they need to fill that slot. Just, yeah.
- JLJason Lemkin
It reminded me of you, Harry.
- RORory O’Driscoll
Yes, just like you, Harry.
- JLJason Lemkin
Obviously, he's wildly talented, right? O- obviously. Work- you know, recruited by
- 2:16 – 4:15
Ambition, career moves & the VC shuffle
- JLJason Lemkin
everybody. I don't know Everett, but he may be as ambitious as Harry, and no one wants to screw around. It's like if you're ambitious today, you wanna go fast, right? You wanna go fast. Vista, Bond, Founders Fund, Kleiner, Benchmark. I mean, those are five good ones to get on your resume in eight years, aren't they?
- RORory O’Driscoll
Mm.
- JLJason Lemkin
No need to stop at any, uh, [laughs] any- anyone in the B tier. [laughs]
- RORory O’Driscoll
And that is probably a- a part of the message, which is, is that, you know, exactly if you've... This is a, you know, it's funny I say this is... There's just a lot of change going on if you're ambitious, if you're young. But everything's moving at fast velocity. You're getting fast velocity markups. You get to declare fast velocity success. You take that fast velocity. You want to rise up the organization. Yeah, you can just keep on moving up in a time when there's a lot of change, and there hasn't been as much change as this in the longest time.
- HSHarry Stebbings
I also know Benchmark are extremely generous in terms of like backdated carry, and being brought into this carry pool for this fund with Fireworks and with Makor and with Lagora and with Manas and with many others, that's a very attractive carry pool to be brought into.
- RORory O’Driscoll
Yeah, no, it's- it- it's the, the Godfather an offer I can't refuse moment, right? Yeah. No, I'm g- I'm g- good for all concerned. Capitalism is great. I mean, of c- I was laughing. I think about it preparing for this meeting and thinking, "Oh my God, it's a wonderful deal." And then you have to remember probably owning one tenth, one hundredth of what the best AI engineer is earning in Facebook Meta just to put all of us, us included, in our place. You know, it's, it's funny. It used to be the best economic gig in tech, and now we have to remember, no matter how wildly successful we VCs are, there are people vesting billion dollars over four years at Facebook Meta because they wisely did computer science and AI at school f- 10 years ago. Just, you know, in these kind of winner-take-all moments, the market for talent in every market, be it, you know, AI engineering, be it top-tier venture capital, it just becomes very ov- you know, very heated. Let's go with h- I was gonna say overall, but that's a judgment. Heated is definitely true.
- 4:15 – 6:25
Are VCs still the best paid?
- HSHarry Stebbings
Well, let's play that out, Rory, because I could have that debate with you where I could still argue that venture investors will be ending, ending up with carry better paid. If you're at a Thrive or you're at a GC or you're at a Lightspeed-
- RORory O’Driscoll
I disagree. There's no one gonna... I- if reports of the billion dollar plus package are true, very few people are gonna make four... $1 billion four-year vested in venture. So I actually argue with you, Harry. I don't think you can top that. Over 20 or 30 years, venture is a great career, right? But, so you look like you wanna disagree. Please, feel, feel free to disagree.
- HSHarry Stebbings
Well, I think if you are one of the top one to three carry participants in one of the large mega platforms, be it Andreessen, be it Thrive, be it GC, I would argue that you will have more than that in distributions in the next few decades. Uh, distributions-
- RORory O’Driscoll
Yeah, you see, you instantly, up until the last sentence, you were, you, you were wrong, but you were making your case. You, you admitted the truth in the last sentence. I mean, look, all these amazing funds, and we hope to have amazing funds too, over 10 years, you're gonna get a ton of money. From 2016 on, right, if you look at distributions in every one of these funds, congratulations, you own a shit ton of private stock that's worth a whole ton of money, right? Whereas, I, I'm going back to my comment, is that where the comp package for restricted stock in Facebook is congratulations over four year, and you have fully liquid stock. In terms of liquid stock cash payment, it appears to be still the best... It, it appears to be the best gig on the planet right now.
- HSHarry Stebbings
How long was it before you guys got your first carry check?
- RORory O’Driscoll
I think the interesting thing is the first carry check was relatively quick, but there was a 10 or 12 year period after that where, you know, it was very much the tail end of the '99, 2000 boom, and then there was a 10-year period of squat. Just not... I mean, yeah, when markets go down 80% as the Nasdaq did, and then they stay down and IPOs are postponed and you have a European waterfall, absolutely. Totally. Venture, as someone said to me years ago, it's the get rich slow program, and yeah, there can be 10-year periods of non-payment
- HSHarry Stebbings
Sorry, Jason, go there on that one
- JLJason Lemkin
No, I, I mean, um, yeah, Brian Halligan this morning was, um, quote tweeting Benchmark about Benchmark one coming up on 20 years and how it was a great vintage and, and wines, uh,
- 6:25 – 7:00
Carry payouts & delayed returns
- JLJason Lemkin
great wines age or whatever it was. I'm like, "But I'm not sure I wanna wait 20 years for my wine. [laughs] I'd like a few sips tonight." I do think it's complicated, but my 2017 fund, 2017 fund should hit 5X on paper by the end of this year, right? But that's like a lot of years already. Like, I- it could be 18... And listen, and, and do you really wanna sell your winners in today's world? Of course, you don't, right? And so, you know, I mean, I, I ho- ho- hopefully I'm not in a walker by the time I, I get my, get my, my, my, my distributions from it. It's, uh-
- HSHarry Stebbings
I mean, that
- 7:00 – 9:40
Revolut's $75B valuation explained
- HSHarry Stebbings
in- that, that increasing period of privatization, as you kind of mentioned there, ties in beautifully into Revolut's $3 billion fundraise at $75 billion, up from $45 billion in 2024. Uh, massively oversubscribed. I mean, everyone wanted this one, to be fair, in terms of large institutional platforms. Private markets win again and public's delayed. How did you read this?
- RORory O’Driscoll
The big picture story is it's another round where the public markets have seeded that business to the private market, right? And this company could clearly go public. It could have gone public years ago. It's doing, uh, what is it last year, 3 billion in revenues, making a billion, growing at 60%. Could go public anytime it wants.
- JLJason Lemkin
I'll tell you what I-- what it, what it made me think about a little bit is challenged one of my early tenets for many years, which is that, um, [smacks lips] the best founders figure out their TAM, right? A small market's okay. They figure it out. They add layers to the onion. Absolutely true with the best founders, right? I'm sure we could all come up with a story. But when I think about Revolut, and Fintech's gone in and out of fashion the last, uh, since we all met, right? It's, it's been hot and unhot, and then people don't like the margins and this and that. But man, the markets are big, right? And so this is my new heuristic. I, I, I-- if I could, I would like to invest in startups that 100 million ARR have 1% or less market share. 1% or less market share at 100 million. That's what I would like. Not... And not even fake market share, um, or... But really, because we're just-- if you look at the public markets in B2B, th- there's no-- there's almost no one except Palantir that's having an easy time north of a billion. [laughs] We can look at all of them, and there's even folks like Klaviyo that are crushing it and still trading at 6X north of a billion. And so I wanna believe that founders will figure this TAM thing out. Um, but now that we're staying private longer and that a billion dollars doesn't even count as an exit, right, on Monday afternoon. It just count, it just counts on a f- a few million decks to buy the house in Woodside, per Rory's earlier story. I'm worried. I, I, I'm much more worried about TAM than I was even 12 months ago. I'm much more worried about TAM exhaustion.
- RORory O’Driscoll
Wow. Three things. I, I think that's fair is that, yes, obviously, if the goal is to get to a billion dollars in revenue before you go public, and our job is to get these companies public, then you need a bigger market than if the goal was 100 million. So I, I totally agree with you on TAM. I don't agree, for what it's worth, on the 100 million 1%. I think the best deals... Because if you go into a big wide market where you, quote, "only need
- 9:40 – 12:30
The TAM myth in startups
- RORory O’Driscoll
1%," you're probably undifferentiated. I think the best wins are you start with this small market, and then as you succeed, your addressable market expands. So I would argue Revolut's early market, and they might say they have only had a couple of percent of it was, you know, not everyone banking, not every next generation bank, um, ba- um, banking customer in the, in Europe, but very much folks who were doing travel and had a lot of FX things. So you pick this pointy little niche, you get traction in it, you get good margins. And then the beautiful thing is if that expands out and you find yourself able to address more and more customers. 'Cause if from day one, I mean, if from day one they'd gone after everyone in Europe for all banking options, uh, uh, all consumer banking options, I think they've gotten spread out. So I, I agree with you on the TAM, Jason. I just think the best of all things is when, as it were, as you grow up as a company, your TAM grows up as well, right? And expands. And I think they've done that perfectly.
- JLJason Lemkin
I just don't believe it anymore in my heart. I, I, with-- I'm not chal- Like, we believe that, but if the A was at 25 post and the exit was at a billion, it all works out, right? And, and TAM exhaustion is someone else's problem because you've distributed 24 months after the IPO. Now I see TAM exhaustion across my portfolio, and I, I never used to... Even 18 months ago, I didn't think about it. I see TAM exhaustion everywhere, and you gotta run so fast as a founder to keep ahead of it, faster than maybe we used to think. We used to have more time. We used to have more time, Rory.
- RORory O’Driscoll
I, I agree on TAM exhaustion. Almost all these pr- high private market bets that, interestingly enough, even there you have a TAM question, not exhaustive TAM question, 'cause typically when you're paying $70 bill- you're buying the winner in a space, right? You're buying the undisputed winner, right? So for all these comp... And you know, you're probably paying a premium in terms of revenue multiple. So in fact, in all these cases, you're making some kinda, the TAM's even bigger than you think bet. Which, so it's interesting you're worried about it at your 25 pre-round, but if you were writing the Revolut memo at $75 billion, you'd be writing the same question, which is how big can this thing get? 'Cause look, for, for context, Revolut's got $75 billion market cap. The biggest bank in England, there's one at 110, and then I think Ba- Barclays is at 60 billion. So you're already as big as the biggest banks in the country you're domiciled at. So all these things are TAM bets at the kind of multiples people are paying
- HSHarry Stebbings
I go exactly to your statement though, which is the founder determines the TAM that they grow into. And, and Rory, you don't, you don't know, but like I'm super close to Daniel at Spotify. I'm super close to Alex at Deel. I'm-- I know Nick at Revolut very well. If you can be close to him, I'm close to him. All of them have expanded TAMs sequentially over time, opened up more new chapters in a way that has unlocked more and more
- 12:30 – 17:05
Why founders must expand their markets
- HSHarry Stebbings
enterprise value. The best founders unlock new TAMs.
- JLJason Lemkin
They do, but Harry just... And I was thinking about that as well. You have Deel. I mean, I know they're not directly competitive in every space, okay? But you've got Deel, Rippling, Gusto, even ones that are much smaller that are older, Justworks, they're all at nine, 10 figures in revenue. My point making is all of them have to start as point solutions for the most case, right?
- HSHarry Stebbings
Yeah.
- JLJason Lemkin
Unless you take the Rippling version. But the, the, but the notional TAM is huge. The best founders do it faster, right? They're not stuck wherever Deel was in 2018, but the notional TAM was large when Deel was started, even if everybody didn't see it.
- RORory O’Driscoll
I agree. So to pile onto Jason and be direct, I think you're wrong. I think-
- HSHarry Stebbings
You know why-
- RORory O’Driscoll
And the three-
- HSHarry Stebbings
I'm wrong?
- RORory O’Driscoll
Yes. And I'll tell you why, to be clear, right? I actually think Jason said it well. You named three companies. I'll just deal with the specific to those companies. Spotify, Revolut, and Deel, right?
- HSHarry Stebbings
Mm-hmm.
- RORory O’Driscoll
I think in all three cases it was pretty apparent that there was a potential very big market there, right? You have music consumption, and that's what they started with. That's what they're doing today. They expanded geographically, but that's the story. Second, Deel, payroll's one of the biggest markets. We'll talk about that in a second. Then Revolut, Fintech, obviously from day one, niche, big market. So I think, you know, where you, where you are correct, and I respect the founder comment, is all three of those founders threaded the needle to go from the entry point to a much bigger... To, to grab that TAM and kind of think of it closely adjacent empty space, right? That's how I think about TAM, is that you have initial small market, then closely adjacent empty space. 'Cause I think the... And but I think what you're missing here though-
- HSHarry Stebbings
To be fair though, I never said that they didn't start in big markets. I just said that they've unlocked more and more value where others wouldn't.
- RORory O’Driscoll
Yeah. Yeah. But my point is, you're im- the implicit statement you're making that I'm just explicitly disagreeing with here, there are lots of people who have started in much more circumscribed markets, and no matter how amazing the founder is, there's nothing they can do. But my point is, it's not just they succeeded 'cause... You were implicitly saying some version of they succeeded 'cause they were great and other people didn't. And again, this is me being a little pointed. I don't believe that. I don't believe entirely in the great man theory. There are lots of kind of thinly sliced SaaS markets that people invested in, in 2017, '18, '19, '20, and you just ran out of space, and you're never gonna be big, right? And it's not-
- HSHarry Stebbings
That, that, that... We, we mentioned more, we mentioned some alternatives to Deel in the exact same market who have grown much more slowly but with-
- RORory O’Driscoll
Agreed. But Harry, but Harry, that's the point. Yeah, the point is this. The market... I, I, you, you... I agree with that statement, right? But the way I describe it is this. You know, the market determines the size... The, the, the addressable market determines the size of the prize, and the skill of the CEO determines who gets the prize, right? But, you know, you, you can't say if you put the Spotify guy running, you know, workflow for back office banking that he's gonna turn up and get 20, you know, $100 billion market caps. Is that you have to start with big wide open spaces. That's my only point. I'm just being a little bit of a nerdy investor. And then you can say now in each of those spaces who won. Why did Spotify win versus... I mean, we looked at a bunch of the other music companies at the time. It was all very fun. I have a little arcane theory on why they won, but we can come to that. Um, what?
- HSHarry Stebbings
Yeah, I'd love to hear why you think they won.
- RORory O’Driscoll
All the other music startups based in the US got strangled at birth by lawyers, right? Because it was all about IP property rights, and little Spotify got going in a bunch of European countries that, let's be frank, your average big five record label didn't really focus on. So they got a much more attractive licensing deal, while in the US all these guys were wrestling with shitty gross margins and litigation with the music companies. So they got critical mass early on, built an excellent product, and then gradually increased their leverage versus the record companies. And it's one, you know, and if you look at, you know, Pandora was always slogging it out 'cause of the, the, the radio type license. All the other subscription companies in the US always struggled to get access to the music because the record companies were such a pain in the ass, and I really love it that Spotify has stuck it to the record companies. It's kind of 10 years late revenge. But again, combination of great execution and a little bit of serendipity to keep them away from the fray and get critical mass.
- JLJason Lemkin
For what it's worth, I, I think in
- 17:05 – 19:55
Are AI Verticals overhyped?
- JLJason Lemkin
the age of AI, we're making the same mistake again in our euphoria. We're very excited. I mean, oh my God, Replit and Lovable, but zero to 250 in, in 10 months, a billion at the end of the next year for... And that, you know, that's putting aside the Anthropic and OpenAI. This is the, you know. And, um, and so we're... There's so much froth, and there's so much greed and excitement in good ways that we are funding so many vertical AI plays that we magically think are, are, are massive, right? And we're, we're funding them, you know, you know, there's, there's no one better than Brett Taylor out there, right? But we're funding, we're funding Sierra at 50 million ARR at 10 billion, assuming... What's the assumption, uh, Rory? That, that that will hit 10 billion ARR in five years or something? I just don't... I get the upside, and I get people turning into, uh, you know, people budget turning to software. I'm already seeing it, right? We have four humans and 12 AI agents at Saastr. But, uh, my gut tells me we're over-romanticizing verticals in the age of AI. Like, we're gonna hit the same TAM exhaustion, and it's gonna be worse because expectations are so high. That's my take on Revolut, Harry. [laughs] It's TAM exhaust... And we're doing it again. We're doing it again in exuberance, and maybe it's fine. How many legal niche tools in AI do we really need?Like, they can grow like a weed, but how many can achieve the velocity at a billion in ARR we need on this 20-year journey to get these carry checks? Like, it- the bar is so effing high to accelerate at a billion, right?
- HSHarry Stebbings
I, I agree. I think, like, legal's actually a very good case though, because it's actually very enterprise, very sticky revenues with people who don't change tools much, and actually get-
- JLJason Lemkin
And a crappy TAM. And a crappy TAM that only looks good to-
- HSHarry Stebbings
Legal's a crappy TAM
- JLJason Lemkin
... it only looks good today, it only looks good today because everyone's in market. Here's a weird thing happening today in AI. It is blowing up our assumptions. We are, in our greed and our rush to make money, we are ignoring something that's happening, and I just got back from Dreamforce. Uh, in a way, it was the conversation of everyone a- at the C- CIO level. And everyone's at market in the first time forever. Every law firm. Listen, I've invested in legal and had a decent exit. It might be five or 10 years for someone to look at a tool, kick the tires, think about it, get nervous, wondering if it's c- Win31 compatible. Um, now because of AI, everyone's being yelled at and said, "Go find a tool," and they're buying. They're coming up with 50 grand, 100 grand, 150 grand, and it's nothing to buy a tool. The fact that everyone's in market instead of 5% of the market, which is like our traditional metric in B2B, is warping how we think about market size. It's like 2020 all over again when everyone was in market for a contact center or e-signature or a digital events tool, like a Hopin. Everyone was in market, and then they disappear the next year. They're not gonna be in market every year for an AI tool. This will be a window that will disappear.
- RORory O’Driscoll
I think this is a huge point,
- 19:55 – 24:00
The "Covid mistake" in AI investing
- RORory O’Driscoll
Jason. Seriously, and you... And by the way, you expressed it so crisply. It's been kind of re- running around in my head, but that was just super clear. Because if you think about SaaS, there was this 20-year period where you, it, it diffused gradually over everyw- you know, like, it, it, it... And you know, some people would market every year and, you know, your companies grew pretty consistently. You could lean in five or six-year growth rate. What you're postulating here, and if it's true, it's gonna be terrifying, is, is that because, because AI is on the front cover of literally every business magazine on the planet, you're saying, right, everyone's in the market. So your signal as an investor on what's going on in '23 and '24 and '25 might be entirely wrong. You're basically, I love it, you're making the Covid mistake. You know, when you looked at the g- growth rate for Zoom in 2021 and early 2022, people, what do you think was gonna happen? There's not a human being on the planet who didn't have a Zoom account by late '22, so the growth went to 10%.
- JLJason Lemkin
Or 4%.
- RORory O’Driscoll
And I don't know if it's happens like that here, but Jason, if any kind of deacceleration happens 'cause of, you know, any kind of saturation or slowdown, everyone's estimates on what's gonna happen here are wrong. Now, w- w- you're right, the counterargument that Harry's just dying to make is the eat the work argument. But at the very least, you, you, you articulate the downside ca- the buried downside case extremely well there.
- JLJason Lemkin
And that's why I bring up legal because, because no one would touch legal for years because the TAM was too small and, and all of a sudden we think it's huge. And don't get me wrong, we're, people are gonna make a ton of money here. There are markets that are utterly changed forever and will, will absorb massive capital. But I think in B2B, we're gonna make as m- more mistakes here than get them right, per Sam Altman's point. And as investors, I think we're being delusional. We're running that, that late 2020, early 2021 playbook again and assuming, not realizing the impacts of everyone being in m- no one's, people aren't gonna buy a legal AI tool every year. They're just not. They're exhausted. They're exhausted.
- HSHarry Stebbings
Well, so I w- uh, the comparison to Covid, I think, is, is not right because it, it was a temporary moment in time that did not sus- sustain. It was not enduring. Uh, to s- apply the same here would be to expect that we won't have AI continuing to have improvements in our productivity in the future, which I think we would all disagree with. So I don't think that is an apt analogy. In terms of the huge TAM in legals, I don't think that's why we're all getting so excited. I think it's because of the structure of data with legal that makes it so relevant for this current set of AI technologies-
- JLJason Lemkin
We'll see. We'll see
- HSHarry Stebbings
... and makes it so relevant. Um, and so I, I just don't really get that. And then Jason, I, I'm going back to something you-
- JLJason Lemkin
Well, hang on, just to step back. I- I- superficially, you're right. And, and listen, even before this week, I never would've said 2020, I don't even use the C word, was anything like today. Okay? It was so weird when everyone all of a sudden needed a contact center. They needed a, in a week, I had to buy everything, okay? Because of we were stuck at home, right? At least the, the, the privileged were stuck at home. The real people still had to work and go to go make your coffee. Um, but as we go f- deeper into this, and, and I, I don't think the analogy is all wrong. Okay? I think it wasn't a change to software. Software g- software was no better in 2021 than 2015. It made no sense. But, um, the fact that everyone is in market is similar, and I think it's worth just learning a few lessons for investing. Confusing ch- permanent changes from folks briefly, for exogenous reasons, being all in market, and there is a s- that's the only similarity, because software is nothing l- uh, the difference is software is radically better today. It's an exogenous reason that every CIO's neck is on the line. Every CMO is told, "Bring in an AI tool or you're gonna get fired." It, that will not last. It will not last, and deals will get harder because they're just gonna get harder.
- HSHarry Stebbings
Are you not making the case for why it's so important to own a market very quickly, a la Harvey then? They're all in market. They're all in market now. They won't be next year. You have to get it now 'cause they are spending, but they're all here now and they won't be tomorrow.
- JLJason Lemkin
You mean that kingmaker point you, you come back to?
- RORory O’Driscoll
No, it's not even kingmaker. I think kingmaker might
- JLJason Lemkin
[laughs]
- RORory O’Driscoll
... but I think, Harry, you accept that's the argument to make. What you're saying is, stepping back, what you're saying is what
- 24:00 – 28:10
The rush to buy AI tools before the window closes
- RORory O’Driscoll
should have been a steady progression of d- company by company decisions over five to seven years has been compressed into every company's gonna make a buying decision in the next one to two years, and then roll it out and stick with it for the following five. So you're right. Unfortunately, the game you have to play is you gotta be here now, 'cause showing up two years from now when, you know, take Harvey's example, 90% of American Law, um, top fi- top 500 American Law will have made a decision, it's just too late. So yeah, you're, you... But the point is, so it's not that Jason's saying it's crazy to be doing it right now. He's simply saying you could overextrapolate the growth nowAnd think it's gonna be like this for the next five years when in fact what you might find is everyone makes a decision and then you slow down quite a lot. The business doesn't go away. It's not Hop In, to invoke a bad memory of your COVID days, Harry. But it's like everyone in the, you know, the top law firms have all made a decision and now you're embarked on a three-year steady rollout. And maybe those growth rates go from, you know, this is back to the whole growth rate comment, the unprecedented, you know, 5X, 10X that you're seeing this year to a more prosaic two, three, 4X, which is still damn amazing. But if you've leaned in too much on valuation, you might be over your skis. And if you've done the number three, four or five player, you might be shit out of luck. I mean-
- JLJason Lemkin
Yeah, and it goes to Harry's point of, of if, listen, if you've got a, if you've got a winner, lean in. Both- b- because I don't think every law firm for the next five years is gonna be in market every year. They're not gonna be in, and they're gonna g- they're gonna settle on whoever they bought. When it doesn't work, they might switch once, but it's exhausting to switch vendors. Business process change is huge. We're ignoring business process change, right? One of the talks at Dreamforce this last week was CEOs were saying how bus- like, onboarding business process change was m- this is the highest it's ever been in their lifetimes. Like, they didn't budget for the cost of onboarding these AI apps. They got the, they got the price of the vendor right, but the soft costs, the training, the onboarding, the business process change, were the highest it's ever been in their lifetimes. They're not gonna do that every year. And so we may go back to 5% being in market in 24 months instead of 100%. And just get it right in your venture models and get it right as founders, to Harry's point. Like, like run, run like hell because they won't be in market again.
- HSHarry Stebbings
But I do think that we obviously just have to acknowledge the difference between enterprise and consumer, 'cause I think consumers will continuously be in market for a new generative AI tools to do videos, pictures, websites, you name it. So I think that, that's a big difference. And Jason, how do you square that away though with your statement before, which I always remember, which was, you know, you can see a future where you have so many more instances of Supabase and you need like 10X more, you just couldn't consume enough compute, combined with Benioff saying 0.1% of Salesforce has AI. What happens when 50% have it? How do you square away those two opposing thoughts?
- JLJason Lemkin
I don't know that there's ... It's a good que- I, like I, I feel like I'm becoming one of those curmudgeons that says you should only invest in trillion dollar markets. I'm, I'm agreeing with it because, I mean, Amazon just went down in part, I mean, whether it's DNS or whatever, it's database contention with Dynamo because so many folks needed, uh, databases. And Supabase hit Amazon issues because so many people need databases. So look, listen, every single app in the world needs a database, and what's changing is now folks might need 10 databases or 20 databases instead of one. So it, it actually is just to the point you should take more and more Supabase risk investing because the TAM is not only massive, but even bigger. And maybe do less vertical AI agents that you think makes, uh, that, that is like a small part of what ServiceTitan does, but it's, but it's amazing. But it also might increase the odds your portfolio comes up snake eyes [laughs] 'cause you're trying to do all the Supabases at pre-revenue at 200 or like, or like, uh, you know, throw, throw the dice at them at 5K a month in revenue, you're gonna have a high loss rate too.
- HSHarry Stebbings
Rory?
- RORory O’Driscoll
No, I'm, I'm, first of all, I'm laughing. This is the longest we've ever gone without actually starting on the agenda, so congratulations everyone. [laughs] I don't know if it's only do billion dollar markets. I think what you've got to be very wary of-
- JLJason Lemkin
Trillion
- RORory O’Driscoll
... I thi- sorry, trillion dollar markets. I apologize. I mean, 'cause I think there's
- 28:10 – 31:00
Will AI markets boom or deflate?
- RORory O’Driscoll
lots of different ways to play it. I mean, I think [laughs] from ... 'Cause I, I'm, I'm a big, you know, you know, the fox and the hedgehog. I'm very much a hedgehog in a ... Sorry, a fox in that there's not one rule from this, um, because, but what it does speak to is d- don't confuse 25 growth rates with long-term growth rates. Have a good handle on your, going back to something we said about Revolut, your TAM, right? And therefore valuation. The interesting style question, actually I'm processing through and we'll come here now. The interesting style question is there are probably lots of these good vertical markets where you can make money. There is gonna be adoption. You're gonna see five, six, you know, seven, $800 million revenue companies. The question per your thing is in a world where you don't go public until then, you know, h- how well do those investments do and how do you av- how do you avoid not overpaying for those extrap- you know, how do you think about valuation for those companies versus valuation for a company like Revolut, which is de facto already public and you're really just doing public investing at scale, right? That may be-
- HSHarry Stebbings
My, my question to, to you both is if this is 2020 and we are overestimating TAMs and adoption and we go back to that, does this pop or does this deflate? Like, is there a-
- RORory O’Driscoll
It's not 20- I thought you said ... It's not that it's 2020. I don't think Jason said that, so I'm gonna defend him. I think he merely said is when you're assessing trajectories, you know, there have been instances recently like 2020 where extrapolating on the last year was a mistake. It would ... If it's true in this case, it would be for very different reasons than 2020. I don't, you know, um, but it could, you know, it could be rate of diffusion slows down. I think the markets, you know, it, it, it may take ... Everyone may be in market to pick someone and then slow down for a year or two at process management to adopt it. So it's not like, it's not quite the same as 2020. It's kind of the history doesn't repeat, it rhymes comment, right?
- HSHarry Stebbings
Mm.
- RORory O’Driscoll
I think what Jason's saying that is why is this, if you over, if you take these growth rates and extrapolate them for the next four or five years, you could be catastro- which a mental model of SaaS, you know, slightly acceleration every year. You could be catastrophically wrong on growth rate. That's what I'm saying. I don't know. You look like you disagree.
- HSHarry Stebbings
No, no, I, I'm, honestly, it's, it's my thoughtful face. I, I learned from both of you.
- RORory O’Driscoll
Well, that's why I didn't recognize it. [laughs]
- HSHarry Stebbings
[laughs]
- JLJason Lemkin
The other, the related corollary just for investing, I think, in, in, in which this is the problem with being a solo GP, you only have so many people to talk about with the thesis. But I, I think that-If you get an M&A offer as a founder or an investor, and the founders make the decisions, the VCs don't make the decisions
- RORory O’Driscoll
Yeah, we make no decisions
- JLJason Lemkin
If your TAM isn't really accelerating, take it. This is my new learning to simplify all this stuff. Listen, if you're, if you get a great offer at 50 million ARR and you have half a percent market share, don't take it. This is the classic Paul Graham advice. Like, everyone
- 31:00 – 33:40
When to sell vs hold your startup
- JLJason Lemkin
regrets selling because the next year you're twice as big, and then you're four times as big. And that, but that can happen even if you're hitting TAM exhaustion. You can still keep growing, but your value doesn't, and we're seeing that in a lot of folks. The value stops, stops increasing. It's not all like Revolut, where every year you go from 40 to 70 to 140 to 280. And so I... This is my new learning when there's an M&A offer. If my, I, I, you know, I have a couple learnings, but one is, let's be honest, has our TAM grown faster than our revenue and are we at tiny market share penetration? And founder, do what you want, but if you're not, if your TAM isn't large and expanding, I, I'm t- I'm too worried the odds are against you that you're gonna hit a TAM headway. This is just my learning. So just sell. VC- we're gonna lose, VCs are gonna lose like 80% of their investments in AI B2B because they're ignoring these issues, and so be it. It's, it's okay as long as one to, one or two of them work out of 10, but they're gonna lose so much money. We're hyper-funding niches once again like we used to because of this in-market thing. We're hyper-funding niches we, we shouldn't be, right? We just don't need that many legal apps or veterinarians that only treat cats. There's just only so much demand.
- RORory O’Driscoll
As, as an investor in vertical SaaS [chuckles] and with many vertical SaaS providers, I'm going, "Hmm."
- JLJason Lemkin
That's why I think about it, too. I, I've changed my mind. I've changed my mind. Um, I, when I s- you know, I, we, we j- it's just the stress at scale is so high now. I, I have... And you just, you really need ver- you know, you can say Toast is a vertical SaaS company, but it's the largest vertical there is in B2B is restaurants, right? It's the largest segment of SMB. Toast is 22 billion, so you're really gonna be worth 220 billion or 440 billion? I mean, maybe. But you bet your, your, your, your vertical better be bigger than restaurants, and you know what? None of them are. So you have to ask yourself, "Why will AI make me much better than Toast?" Otherwise, don't make the investment. And that's a tough question to answer at the partners meeting, isn't it? Why will this be much bigger than Toast? For any vertical SaaS. There's no vertical, I don't know about over there in the UK, but I think they eat a lot of chips and a lot of, uh, I, I'm sure it's true in the UK, too. It's the number one vertical that there is. Maybe you should only invest in restaurant ver- even though it's the worst vertical also, the lowest TAMs, the highest churn. But it's also-
- RORory O’Driscoll
And-
- JLJason Lemkin
But, but at least you can make it up at scale. [laughs]
- RORory O’Driscoll
I mean, I, I, I'm not torn on it because 22 billion is an incredible exit, but so is two billion, and two billion would return my funds several times over in certain cases.
- JLJason Lemkin
Yeah, but when you do that over-inflated investment, sure, if you got in really early, it's one thing. But when you did the A at 150 at 3 million ARR because everyone else wants to do it, Harry, and you beat the, and you're, you're patting yourself on the back for beating out Sequoia
- 33:40 – 40:10
The truth about vertical SaaS
- JLJason Lemkin
and Excel and, and, and Stride and all those guys. H- does the math pencil out on any of these deals? Is it really better than Toast, this, this vertical SaaS? I, I, I'm getting more worried as time goes by. [laughs]
- RORory O’Driscoll
Yeah. I think that's, that's, that's the nuanced reply, which is, you know, you look at these verticals and you kinda go, they're adding value. The i- you know, the product is better for the cons- for the c- business customer than the prior version of SaaS. They're either making the, the customer experience better, they're replacing labor. There's a business here. Typically, what we're seeing is in these verticals, it's a wedge product. Maybe it's document recognition or voice bot. And you can see, you can build a story down the line that TAM, you, you have TAM expansion within that vertical as you just do more and more, and you kinda take from that. And you kinda go, are you building value every day? You're damn right you are, right? And therefore, you're probably creating a valuable enterprise. Money should be available to fund those. They're at the cutting edge. It's AI. It's not trailing edge, plain vanilla SaaS. You've got this vertical. You may be only a smaller number of competitors. Y- you can build enterprise value here. I think the two things you're saying are, one, that's all very well, but if you go in at a p- at pre-money equal to your total TAM, you're never gonna make a dime, right? And I think you're exactly right on that. And, you know, assuming that every market is as big as the biggest market is the fatal error. And then the other thing is because the bar is so high for exit, right, as I said, there's really two games going on. There is the game of investing in companies at one million and trying to get them to two, 300 million, so they could go public. And frankly, that's the game that 25% of the, or 30% of the dollars are playing. And then I think, it's not our game, but I look at them and go, maybe the savvy money is playing a totally different game, which is these companies are already long past the point they can go public, but let's just keep pump, you know, let's do this winner take all, keep pumping and keep funding them in private space. And these little vertical companies aren't going to become that, right? So you've got to say to yourself, are you making a perfectly good product for a perfectly sensible world that no one gives a shit about, right? Because you're not Revolut, and the easiest way, it looks like the easiest way to make money in 2025 is to take the very biggest companies and double down one more time.
- JLJason Lemkin
Here's the bull case, right? The classic for me when I started investing, I, I, I stole this from a slide someone did at Emergence when I started. But for, for vertical, uh, uh, SaaS or vertical B2B that's, that's somewhat SMB, right? That's basically an ERP. It does everything. It does payroll. It does back end. It does, um, you, you... For the smallest customers, you want to get to $10,000 a year at least, okay? It's, it is what they run their business on. Get 10,000 of those, you got a hundred million dollar business, right? And that proved to be true, uh, again and again and again. It's just 100 million isn't enough today, right? So the question with AI, with replacing humans with software, can those same 10,000 spend 100,000If they do, you get a billion in it. You get a bill- you may still slow down at a billion, and that's the, you know, that's the question that I think we're, we're gonna have some wins on and some losses on. Can, can... Will people really spend 10... $100,000, small businesses, on your same vertical agent software and those same 10,000? Um, and we'll, we'll find out, right? We will find out.
- HSHarry Stebbings
That's where I think legal is attractive, though. Like, you know, we have this company Solve Intelligence that sells to IP law firms. Yeah, all of their contracts are over 100 grand, several hundred grand in a lot of cases.
- JLJason Lemkin
But so are the vendors you're competing with, so are the LexisNexis and others. It's not 10 times larger. I'm not saying it's not a great investment. I'm not saying it's not gonna return your fund. But it's gotta be an order... It's gotta be 10X higher for the math to pencil out in venture today, like, the, the deal sizes, um, not just the number of folks in market. That's where the confusion is. You could confuse the two, number of folks in market and is deal size 10X what it was 24 months ago, right? So I, I think that the, the, the lovables and the replits have massive deal sizes in a sense, right? It's so much TAM extraction away from crappy agencies and vendors. But if it's this... if it's just a little bit bigger TA- uh, deal size, we were gonna, w- we're gonna get crushed, right? I don't know. That, that's my simple math. Can you get 100K from a small business or a million from a mid-size enterprise like a, like a law... Will, will, will a plaintiff's law firm that used to spend 100 or 200K on just a couple pieces of software spend a million dollars on your software because they don't need humans anymore? If they do, it's golden.
- RORory O’Driscoll
And, and, and, you know, to, to, to, to get back from the dollars and cents, stepping back and giving the case for the, the defense as it were, right? You had legal software, horrible market for many years-
- JLJason Lemkin
[laughs]
- RORory O’Driscoll
... 'cause it was basically selling workflows to people who didn't care, right? To be fair, LLMs manipulate words. That's the core of what they do. And lawyers, it's the most LLM obvious market out there because it is the manip... So there is a lot of... You can definitely make the case in all these verticals. You know, we could talk about patent law, which is where you are. You can talk about Harvey in corporate law. You could talk about Even Up in plaintiff litigation. There's a... But you can definitely make the case that what came before is not predictive of what's happened now from a technology perspective. There is something really exciting going on in law because of LLMs. So I want to put that out there, right? And it will change the practice of law just because the technolo- again, at some level you have to be a techno determinist. The technology that we have invented or we... the world has invented, Sam Altman has invented, whomever has invented, is supremely good at ingesting, synthesizing, and spitting back out kind of word concepts, and that's what lawyers do. So if ever there's an industry that could be automated and changed, it's these guys. So that's, that's the case for the offa... That, that's the case for a lot of these companies can in fact swallow some much more of the... So I don't think the past is predictive in terms of the amount of dollars you can extract from these companies. I think it could be 10X, Jason, to your point. But I... But you are right in the sense of if, on the other hand, if you start slicing it, you know, you take the overall lawyer count and you slice it, how many are patent, how many are litiga- If you, you can... If the bar is 100 or 200 million, I don't think you hit TAM exhaustion. If the bar for an exit is a billion, then you could hit TAM exhaustion in some of these markets pretty quickly. So a lot of it, in my view, boils down to the healthiness of... Yeah, the ultimate return boils down to entry valuation and the healthiness of the exit market.
- 40:10 – 41:34
AI is changing the legal industry
- RORory O’Driscoll
I mean, but I think it's the right place. I mean, you should... Stepping back, you should be investing in the area where the technology's having the greatest impact, which is... I mean, it totally makes sense we're looking at these spaces, because what else are you gonna do?
- HSHarry Stebbings
For me, there's like three areas where I'd be investing if, if I had infinite capital sources. Like the absolute winners in this space, your OpenAIs, your Anthropic, the absolute anointed winners. Your absolute winners with great economics, which is your Revolut, it's your deals of the world. And then it's your really early. And I think that is three great pillars. What I don't want to be doing is respectfully, and I'm... I don't mean this horribly, this, this is a different game, but your Mira Murati's 2 billion at 10 billion. Your 300 million into Periodic Lab. So it's like this is a huge amount of money into a very still questionable early asset. That's where I'm like, "Ah, I don't know."
- RORory O’Driscoll
And it's, it's interesting that the fact that the two out of the three... I mean, I think the fact that two out of the three Harry chosen spaces are effectively post-pub, post public eligible anointed winners speaks to where the market overall is. Most of the dollars are going there, right? And you're all of... Basically, what you're saying, Harry, is w- one third of your money you'd like to do venture capital, and with two thirds of your money you'd like to do public style investing with a two and 20 com structure 'cause they're still private. And I think you're exactly right. [laughs] And, yeah, and the market seems to agree with you that that's a good way to make money.
- 41:34 – 46:44
OpenAI's billion dollar cloud play
- HSHarry Stebbings
There we go. Uh, listen, we mentioned the, the anointed winners there. I'm loving this conversation. Um, OpenAI ha- have said that they will spend potentially more with Oracle than with Microsoft. I find this relationship fascinating in how it's developing, the OpenAI and Microsoft. Um, how did you guys read the OpenAI spending more with Oracle than Microsoft and what it means for the power dynamics and that relationship?
- RORory O’Driscoll
I think Microsoft didn't want to make an e- didn't want to spend money economically irrationally, and Oracle wanted to be in the game, and OpenAI seems to be extraordinarily good at divining other people's needs and wants and desires and taking advantage of them, right? I think when all is said and... As I said it before, on Microsoft's side, when all is said and done, the shareholders should award medals to Satya, their CFO, and their GC, and they should hire someone else to do their technology 'cause they haven't shipped, right? They have cut a brilliant deal with OpenAI, and now they're gradually stepping back as the hype comes in and they're saying, "We're just not gonna make economically irrational investments."I think, I think they're smart
- JLJason Lemkin
Yeah, I mean, th- there's a lot... I mean, obviously in the end, ChatGPT... I mean, OpenAI needed much more capital than they thought when they started. Microsoft bailed them out in buying 49% of the company. Now they need much more capital than Microsoft thought. [laughs] Or probably two orders of magnitude more than Microsoft's, uh, b- uh, you know, uh, high-end model of how much capital OpenAI required. So Microsoft, by, by de-acquiring it, right, in, in essence spinning it out for 30% ownership of what they get, but not having to, to, to fund it, um, they're getting, they're getting, um, the folks that, that, that, that can tolerate a much lower margin, right, and can somehow get a market benefit out of this in Oracle. So it, it's kind of crazy that Oracle comes out of here and replaces Microsoft, but, um, but Microsoft also gets out of a may- maybe even what could have been an awkward situation, right? If they were somehow stuck funding, funding their subsidiary, I mean, uh, it might, that might be more than nickels and dimes if they had to fund OpenAI for eternity.
- HSHarry Stebbings
And you said about economic rationality and Oracle kind of stepping in and being that capital provider in a lot of ways. Their debt-to-equity ratio now is like 4.6x. It's high. Are Oracle out over their skis, or am I being overly cautious?
- RORory O’Driscoll
Well, we said two weeks ago we think they're over their skis, and since then the stock's down, so I think we can claim an attaboy on that one, right? I think, yeah, look, they're, uh, you know, you'll only know when they play the game. If the demand for AI compute is as high as OpenAI appears to think and Oracle appears to think, then all this, and, and, and then all... And they can bring this investment in on time, then they will be rewarded with a perfectly good business at decent growth margins, not as bad as they currently are because I think there's some startup costs. So it will have paid off, and their current market cap will be validated. I just look at the risk-return profile and say, "It's no accident that Microsoft said, 'You know, that's an interesting risk-return profile, but I don't need that bet.'" And Oracle said, "I'm a wannabe in this space. I'll take it." Right? I mean, one of the interesting things, going back to what I said, is I saw a dumb tweet that was like, "Oh, OpenAI is gonna go bust," you know, like, uh, because you had the whole capacity, AI is not gonna take, get those quickly, and a really dumb OpenAI is gonna be in trouble. No, they're not. They have brilliantly palmed off all the risk on everyone else. I mean, if you step back and say, you know... 'Cause we're gonna talk about Poolside in a second. OpenAI said, "Yeah, we, we, OpenAI, need gigawatts and gigawatts of data centers. We need gig- you know, gazillions of chips. We need all this stuff. You all should do it. Go team. And yeah, we've signed commitments, and if we need them, we'll actually pay you, Shakara, one fine day with money we don't yet have." But they're not taking on huge amounts of leverage. They're not taking on huge amounts of building. They're just like, "We're in the market to buy this stuff. You should invest on our behalf." It's brilliant. And they've offloaded a lot of the balance sheet risk to everyone else, and all these other people seems to be happily taking it on right now, and we'll see. We'll see how that, that, that strikes me as a lot of risk to take, especially when in the end, if it all works, OpenAI gets the upside. Your best case is you're, you know, the commodity compute provider to someone who is very rational and gonna be able to grind you down at scale. Your worst case is you put billions of dollars into, you know, fixed assets that aren't gonna return. So I think that, you know, again, the OpenAI corporate, pardon me, the OpenAI corporate development deal-making machine is second to none. They have a rootless instance for weakness and take advantage of other people.
- HSHarry Stebbings
I totally agree with you. Uh, I'm pleased you mentioned Poolside there, and I, I do want to go to it now actually, because it is a super relevant one tied to that, which is, like, on the vertical ownership side, Poolside announced building its own two gigawatt AI center, um, which is a big announcement. Also, Poolside have not released a product to the
- 46:44 – 50:19
Poolsuite builds its own data center
- HSHarry Stebbings
public. They have customers, and they do have usage, but they haven't officially launched a product to the public. And for those that don't know, Rory, how would you, or Jason, how would you describe it?
- RORory O’Driscoll
The, what, what, would you mean it, the product, or it-
- HSHarry Stebbings
Yeah, the, the product as in like-
- RORory O’Driscoll
I understand. Yeah, and again, you know, I've met Jason in the past. I think he's a wildly smart guy, very talented and very successful. My understanding is it's kind of an enterprise-focused, uh, they're building a core LLM to do enterprise-focused coding and software development. So some version of that, right? And provide an entire runtime environment for these models. So big enterprise idea. So not knowing the traction, stepping back, if they're right or if they're wrong, either way, it's terrifying. 'Cause the conclusion they're basically saying, and they're very smart people, is in order to compete at the software layer, you have to not only build your own LLM, but now, goddammit, you got to build your own data center, right? So what they're basically saying is this game that you thought was a software game is now a fixed asset at scale game, right? And they're basically saying... I mean, they're not doing it because they're saying, "Hey, I'd really love to own a data center 'cause nothing says fun like fixed assets." They're presumably doing it because they can see no other way of doing it. And what that means is these smart people have concluded that that's what it takes to win in this space. I don't know if I agree, but I haven't looked at the specifics. But again, if they're right, if they're 100% smart and 100% right, what they mean is do all the other, do all the other companies trying to build models have to do the same thing? Is there a conclusion here for the capital intensity of safe super intelligence, for the capital intensity of thinking machines? Do you really have to own your own damn data center if you want to build an LLM? It's, it's, it's, it's a interesting and big ass conclusion, right? And I looked it up. It's not even like they're doing the Altman thing of having someone else build it. You know, Core- they're partnering with CoreWeave, but, uh, part of the documents I read is that Poolside's gonna be developed. But I might have guessed it would have been one... You know the way big corporations often do a build to, to lease, where they say, "I'm a software company. Mr. Developer, build this building and I'll lease it from you for 10 years." So I might have thought they'd have said to CoreWeave-Mr. Data Center guy, build this data center and I'll lease it from you. But in fact, they're actually stepping up and being the prime on it. So I think it's a big escalation in capital intensity. I think they must have been driven to that, not by choice, but by necessity. And it just speaks to this, the business of playing in this space has become even more and more high stakes.
- JLJason Lemkin
But what is clear is the competition's gone way up over that time, right? The competition, everything from, from Claude Code to GPT 5 Codex to whatever now, no one wants to manage a massive data center, but there's no way to achieve their goals otherwise. It's probably, and going to Rory, I think I got it wrong. It's not about cost, right? It's not, there's no way they can do this cheaper, right? It's not about cost. It's about the fact that the bar has gone up to, to compete, um, with horizontal applications and, and it's just much, much, much bigger than when they started this journey.
- RORory O’Driscoll
'Cause if, I mean, if you think about it, if someone came in to you... Just think how different the economic intensity here is. If someone came in to you when you're building a SaaS app and said, "I'm building this great SaaS app, but by the way, we're not gonna use AWS, we're gonna need our own infrastructure layer." And you said, "Hmm, that's interesting." And then they said, "Oh, and by the way, we're not even gonna use someone else's data center. We're gonna build our own data center, right? And we're gonna do all this so we can have really great stuff." You'd be like, "Get out of my office," right? But that's where we are in this market, right?
- JLJason Lemkin
It's good to sneak some of these things up on your VCs, isn't it? You don't wanna, you don't wanna let, you don't wanna-
- RORory O’Driscoll
It totally is
- JLJason Lemkin
... you don't wanna scare them in the first or second check
- 50:19 – 1:33:58
The rising cost of competing in AI
- JLJason Lemkin
on things like this. [laughs]
- RORory O’Driscoll
And, and, and you know, it's a cynical comment, Jeff, but you're exactly right. What's happened here, it, it is the boiled frog of capital intensity, right? And I think this, again, I'm gonna say it, is where OpenAI and maybe G- they have made a game of cap- they have made it a game of capital intensity, where they're clearly winning, right? And it's kind of making it harder and harder for people to emerge and compete, right? And again, I don't know what drove, what, what level of compute they felt they needed and therefore what they had to do. But a- again, I repeat what I'm saying, assuming smart people making intelligent decisions based on the facts they have today, it's a terrifying conclusion about capital intensity for people who wanna play in this space. And you're right, you, you said it, it sneaks up on you as a VC. You think you're in a business that needs $500 million to cash flow breakeven, and suddenly you're in a business that needs $5 billion to cash flow breakeven, and you own bulldozers digging a hole in somewhere in Texas. Oh my God, what the fuck just happened? Excuse me.
- HSHarry Stebbings
So I, I was, I was actually like one of the first investors. I don't know if you guys knew this. Um, I, I invested in, uh, Iso, the founder of Poolside's business-
- RORory O’Driscoll
Cool
- HSHarry Stebbings
... which piv- which pivoted into Poolside. And so I got rolled in very luckily into, like, the first round, which is great. I'm very grateful for it. It's like a 50X. Thank you, Iso. Uh, I'm just trying to understand though the rationality, like, for all the providers who are building models, what do you think Poolside are seeing that they are not? Like, ju- I, I feel very curious.
- RORory O’Driscoll
I'm, I, I'm gonna make a really pointy distinction here. You used the word rationality. Let's agree that the word rationale is not the same as the word rationality. Rationale is why you think you're doing this. Rationality is whether you're right, right? And I think the rationale is pretty clear here, right? Oh my God, I need this compute, right? That's the rationale, and we'll know in five years was that rational or not. Does that... I'm being-
- HSHarry Stebbings
Well, no, but that, that's very helpful. Thank you for the, that English lesson. Um, but-
- RORory O’Driscoll
No, it wasn't meant... Genuinely, it was a little bit of a sn- it wasn't actually meant to be snarky, though I can see why you often think I am. Um, and it was just really trying to distinguish carefully between why you think you're doing something, which can make a ton of sense on the day on the assumptions and whether in fact you're correct in the end.
- HSHarry Stebbings
But oh my gosh, I need the compute when no one else who is building their own models shares that opinion.
- RORory O’Driscoll
Well, OpenAI does, and Anthropic. I mean, need the compu- remember, need the, need access to the compute. It may well be what you're seeing here is that because OpenAI and Anthropic have a s- for all their faults, and the hyperscalers, have sucked up all the capacity. It may well be it's as simple as these guys realize, "I need X gigawatt of data center capacity, and I just can't buy it today. So if I can't buy it, I gotta build it." I mean, they're not doing it 'cause they want... They're doing... Literally, it speaks to, you know, all this CapEx is, you know, sucking up all the capacity there. And, you know, even though going back to the now versus the future, I might be skeptical of the ultimate return on this marginal CapEx, and I could be right or wrong on that, you'll know in five years. It is probably an objective fact today that if you woke up and were trying to build your business and needed that scale of compute, you simply couldn't get it because you'd ring CoreWeave and they'd say, "Look, I promised 22 billion to OpenAI. I promised 10 billion to Anthropic. I promised 5 billion to Microsoft. I got nothing for you." And then you're left going, "I either p- give up my dream and say I can't do this, or I can't put my company on pause until '27 when I think all this shit is cheaper. So I gotta play the game now." And they said, "In that case, I gotta go build it." I can totally see how you get to that point, which is different than saying you won't regret it in two years. You just literally go... I mean, I assume they're doing it 'cause they rang and said, "Will you sell me two gigawatt of data center capacity?" Whatever megawatt. Um, gigawatt, sorry. Yes, it's gigawatt, yes. And they couldn't find anyone to sell it to them at scale because it's all been taken up by people with bigger balance sheets.
- HSHarry Stebbings
Thank you for explaining the rationale. That makes total sense now. Uh, and it really does. I, I was struggling to actually understand the logic behind that thinking. Um, it also to me though indicates, like, their expectations on future abilities to fundraise. It is a bet the boats decision to have this permanent investment, and they clearly think they'll be able to raise a huge amount more.
- JLJason Lemkin
Well, probably when they started, they didn't... I'm sure their, their p- their slides looked great, but deep down, I don't think they thought that, um, you know, uh, Cloud Code would be at a billion, that Cursor would be at a billion, that Replit would be coming up on a billion. I mean, this is... I, I think they believe the notional TAM, going back to the conversation, was huge. I just don't think they thought we'd be in the billions of, of revenue already. And so now, I mean, going to Rory's point, there's probably no other way to get 40,000, uh, NVIDIA GPUs and the p- like, it's just not possible otherwise, 'cause you're not the leader. But also it's f- but going to your point, it's probably fundable today because they d- this is much, this is much bigger than they probably... They knew they would be huge, they just probably didn't think it'd be this big in, uh, uh, you know, Q4 of 2025. And so now they can raise $5 billion or $10 billion, I don't know, whatever the number is, which was probably impossible when the, when they started. So go for it.
- RORory O’Driscoll
The, the thing is, everyone's aspirations and their risk appetite have been walked up. You're exactly right. No one had a plan back in 2016 for OpenAI or 2000 and whatever, '22 or '23 for these to say, "I'm gonna need $5 billion to even play." You think you get there on 50, you think you can get there on 500. The stakes have gone up. The signal is strong because the returns are there in terms of market adoption, and you just... Everyone's risk appetite increased. Now, at some point, could that perspective change and would that be pretty painful? Yeah. But that's, that's how every boom goes. That's what it feels like. That's what it feels like when you're trying to buy memory chips, when there's a memory chip shortage, and you've got no choice but to sign up with five different distributors and commit to buying them 'cause you can't get any capacity. And then one fine day, capacity comes online, demand diminishes slightly, and oh my God, these things go down 25, 30% in value, right? That's what the boom and the bust cycle is like. And at some point, that'll happen here. And, you know, you don't want to-
- HSHarry Stebbings
I, I think that's a great way to put it. It's really interesting that the boom and the bust cycle makes me think of a bubble, and I was looking at definitions of a bubble last night because, you know, I have far too much free time on my hands. Uh, and, you know, it really was two things. One is a more than proportionate drop in, you know, the value of assets, more than proportionate being more than 20%. And then pr- and the second is productive capital leaving a market for more than three years.
- RORory O’Driscoll
And that's a bust, to be clear. That's not the bubble definition. That's the bust definition.
- HSHarry Stebbings
That's a, that's a bust definition. And I thought, wow, that's not where we are today. Like, we will not have productive capital leave AI and data centers for more than three years. And so everyone that's like, "Oh, we're in an AI bubble, we're in an AI bubble, and it's gonna bust," I don't think so. Because if we're in a bubble and you're anticipating the bust, you're suggesting that those two elements will happen.
- RORory O’Driscoll
No, I don't buy any of that. I mean, I think your description of a bust is actually correct. That's what, that's what it feels like on the, on the downturn, which is different than saying it's gonna happen. I think it might, but i- the definition is correct. If it goes wrong, let's go with the if. If it goes wrong, I, you know, I don't think it'll be because none of this stuff works. It'll just because, oh my God, to Jason's point, we over-extrapolated on one year's adoption and we thought everyone's gonna buy this in three years and we're gonna need X gazillion dollars of capacity. It turns out that growth next year slows more than we thought. It's still a dominant long-term trend, but the diffusion of this technology is gonna take 10 years, not two, and we've over-invested in capacity. The marginal player cuts back on their purchases, and then pretty soon, instead of having a shortage of data center capacity, you have a mild glut. Then the price goes down, and that's how it unravels. That's what happened in the bandwidth bust. It was, it was a boom in 1996, probably about till '99, 2000. And then there was five years where no one invested in more bandwidth, 'cause you wouldn't. Because once there is existing assets available for sale at less than the price it takes to build new assets, then no one rationally builds new shit, right? And you could imagine a world, I'm not saying it's gonna happen yet, um, we can t- discuss that separately, but the way it goes wrong is if people don't need the marginal cent- data center that they built for $2 billion, they have to sell it. And if the only offer is a billion, that's what they take. And if that's the case, no one's gonna build another data center for $2 billion. That's what the unravel would look like. Now, separate question, is it gonna happen?
- HSHarry Stebbings
Is it gonna happen?
- RORory O’Driscoll
Of course, 'cause if I had certainty on that, you think I'd be wasting my time talking to you, Harry? I'd be trading as we speak, right? I mean, it's super hard to call the time. I mean, I, I find it plausible that... I find it almost inevitable that at some point you will over-invest, 'cause that's the nature of the beast.
- JLJason Lemkin
We could talk more about it in subsequent... I'm making an investment this week, a B2B AI investment, that, uh, it's early-ish, it's early, but it uses more inference than anything I've invested in yet. It uses far more... And in fact, what they wanna do soon is to use 24/7 inference, to be running massive amounts of compute for a, a relatively common B2B use case 24, 24/7. And so the math here is, now you've got to... There's, there's work to be done. They can't, you can't afford it, right? But it's a sign of the future that smart folks are gonna figure out already how to use 1,000 times more inference and compute than we're using today. Because instead of running a little one-off thing, right? Or even, or even using Lovable for an hour and then letting the servers with no load go, this is 24 hours, 7 days a week, 365 days a year, running, like, 20 different passes through the cloud API, wanting to go as quickly as possible. We're gonna have more apps like that, uh, at all levels. We'll... That's three orders of magnitude more inference than you really wanna use today, right? And if it was available today cost effectively, they would consume all of it. They would consume all of it today if they could. They have demand from their end customers.
- RORory O’Driscoll
Yeah, and that was the key sentence, cost effectively. And then on today's price, they probably can't afford that to do all, all the time, and the bet you're taking is as price comes down, that will get used up, correct?
- JLJason Lemkin
Historically, folks have gotten smart at this, right? Like, an ear-early bet I made is a company called Opus Pro, Opus Clip, that Harry knows, right? And they, they made clips from videos. The truth is, it didn't need as much compute as you thought, but they got really good at it. For example, in the early days, they'd only show you the first couple clips, and you'd have to request the rest because there was no point in giving you 30 clips when the 30th was never as good as the first one. [laughs] And they, they got better at a million things. Um, now we're at the age, I, I don't know. You know, if you're running, if you're running massive inference constantly, it's, it's not that simple. But it does augur well for the build-out, right? We've just... Forget about where's the apps. We've just started in the amount of inference these apps can use. We, we've just started, right? Maybe, maybe the next legal app, the next Harvey, or maybe Harvey does this, shouldn't just do what you want on demand 24 hours a day. It should be figuring out what you want, and you wake up in the morning and it's done all your legal work for you all night long.
- RORory O’Driscoll
If you believe that, should we not just be plowing money into NVIDIA?
- JLJason Lemkin
Yes. I mean, it's, it's where all the m- ... We all, we all are. It's all of our 401 [k] s. We're already deep. We're already d- [laughs] All of our QQQ and 401 [k] we- we're already, we're already long NVIDIA. [laughs] Depends, just depends how much more you wanna concentrate.
- RORory O’Driscoll
You know, I, I, I think-
- JLJason Lemkin
But yeah
- RORory O’Driscoll
... w- one-dimensional sentences aren't useful, right? Um, because, you know, [sighs] 'cause a couple of things. One is, almost inevitably with a trend this amazing and impressive as the AI technol- what, what, uh, what the technology, in fact, the trend is undersells it. The technology of AI, with something that powerful in terms of a, a powerful economic impact, you will get over-investment, 'cause it's just the nature of the beast. People will keep leaning in until it hurts. So at s- it's, it's, it's inevitable that at some point in time people will find themself overextended and there will be a retrenchment, because that's just the way markets work. They keep- 'Cause if, if it works at 10X growth, go 20. If it works at 20, go 30. And the only thing that stops you doing is when it hurts. So of course there's gonna be a correction. So sitting here and going, "Is there gonna be a correction sometime?" Isn't that useful, right? The challenge you all face as investors, we all face as investors is, you can't sit it out and say, "I'm gonna wait for the crash." That's not a useful thing. How do you make sure you get enough, you, you, you're there to take the upside and still be survivable when the shakeout comes? So that's, uh, so it's not a one-dimensional comment. It's not, is there a crash? Yes, no. It's more a question of, you wanna take all the advantage of this amazing technology, but you wanna run your business, time your investments, um, do temporal diversification such that at any point in time, and you don't know when, the whole market's gonna find itself overextended. You can survive that overextension and keep on, you know, lean into the trend afterwards, right? And where, where temporal diversification-
- JLJason Lemkin
But no one's doing temporal diversification now, are they?
- RORory O’Driscoll
No.
- JLJason Lemkin
Everyone's just ri- rai- raising a fund every 20, 18 to 24 months. There's no divert- temporal diversification.
Episode duration: 1:34:11
Install uListen for AI-powered chat & search across the full episode — Get Full Transcript
Transcript of episode CJpFFKXMNQ4
