# Anthropic at $2 Trillion? The AI Bubble, and Why the Best Startups Always Look Overvalued

Markdown copy of https://realalpha.app/v/K2OC5yKrbaM/ on Real Alpha, the best minutes from long interviews with the people building AI. Cite and link the page, not this copy.

Matthew Harris argues that venture power-law winners often look overpriced, while AI still leaves room for new services companies and undiscovered leaders.

- Show: How I Invest
- Hosts: [David Weisburd](https://realalpha.app/people/david-weisburd/)
- Guest: [Matthew Harris](https://realalpha.app/people/matthew-harris/) is a Partner at Bain Capital Ventures focused on financial services and technology.
- Published: Oct 26, 2026
- Length: 1 h 3 m

## The best minutes

- [1:58](https://realalpha.app/v/K2OC5yKrbaM/?t=118) [Valuation discipline can make venture firms miss power-law winners](https://realalpha.app/s/valuation-discipline-can-make-venture-firms-miss-power-law/) (5 min)
- [17:35](https://realalpha.app/v/K2OC5yKrbaM/?t=1055) [AI labs may avoid competing with their biggest customers](https://realalpha.app/s/ai-labs-may-avoid-competing-with-their-biggest-customers/) (6 min)
- [32:44](https://realalpha.app/v/K2OC5yKrbaM/?t=1964) [Anthropic's growth could support its reported $2 trillion valuation](https://realalpha.app/s/anthropics-growth-could-support-its-reported-2-trillion/) (4 min)

## Also worth watching

- [11:53](https://realalpha.app/v/K2OC5yKrbaM/?t=713) [Bain's AI strategy centers on robotics and AI-native services](https://realalpha.app/s/bains-ai-strategy-centers-on-robotics-and-ai-native-services/) (6 min)
- [8:51](https://realalpha.app/v/K2OC5yKrbaM/?t=531) [AI market leadership can shift despite consensus investing](https://realalpha.app/s/ai-market-leadership-can-shift-despite-consensus-investing/) (3 min)
- [27:59](https://realalpha.app/v/K2OC5yKrbaM/?t=1679) [OpenAI's open Codex strategy could win the application layer](https://realalpha.app/s/openais-open-codex-strategy-could-win-the-application-layer/) (5 min)
- [52:38](https://realalpha.app/v/K2OC5yKrbaM/?t=3158) [Steady venture pacing helps limit cycle-driven risk](https://realalpha.app/s/steady-venture-pacing-helps-limit-cycle-driven-risk/) (5 min)
- [39:48](https://realalpha.app/v/K2OC5yKrbaM/?t=2388) [AI agents could expand the market beyond human users](https://realalpha.app/s/ai-agents-could-expand-the-market-beyond-human-users/) (3 min)

Skip the roughly five-minute parenting discussion, the sponsor reads and the personal AI-use examples.

## Full analysis: Matthew Harris says Anthropic's $2 trillion valuation is reasonable

- Matthew Harris said a reported $2 trillion Anthropic valuation looks reasonable if growth continues.
- But he flagged uncertainty about continued growth and seller supply when lockups expire.
- For founders and investors, separate operating growth from liquidity and test the assumptions behind scale.

Its 6 sections, what it means for you and what to watch are on [the page](https://realalpha.app/v/K2OC5yKrbaM/#analysis), free with an account:

- [4:11](https://realalpha.app/v/K2OC5yKrbaM/?t=251) Valuation discipline can screen out power-law winners
- [11:20](https://realalpha.app/v/K2OC5yKrbaM/?t=680) AI leaders can lose consensus status
- [15:27](https://realalpha.app/v/K2OC5yKrbaM/?t=927) Service businesses hedge against lab competition
- [33:16](https://realalpha.app/v/K2OC5yKrbaM/?t=1996) The $2 trillion case rests on continued growth
- [40:33](https://realalpha.app/v/K2OC5yKrbaM/?t=2433) Agents widen demand while cheap models pressure pricing
- [56:01](https://realalpha.app/v/K2OC5yKrbaM/?t=3361) Pacing manages risk, not the bubble

## Summary

Matthew Harris is a Partner at Bain Capital Ventures, where he invests in financial services and technology companies. He says he ran his own venture firm from the late 1990s through 2011 or 2012, and discusses BCV’s investments in robotics and AI-native services.

Harris argues that valuation sensitivity can cause venture firms to miss power-law companies, since winners may look overpriced at every round. He says AI investing is increasingly consensus-driven, but leadership can shift, and describes BCV’s focus on robotics and AI-native services companies that perform industry work rather than just sell software. He expects compute access and model capabilities to shape competition, while arguing that foundation-model labs may hesitate to compete directly with major customers. He considers a reported $2 trillion Anthropic valuation reasonable if growth continues, but says liquidity and venture returns remain uncertain; he advocates steady investment pacing and personal agency in an AI-rich world.

### Minute by minute

- [0:00](https://realalpha.app/v/K2OC5yKrbaM/?t=0) **AI could be late in its cycle.** Harris says that if AI follows other technology waves, the market could be in the eighth inning. If AI is another SaaS-like chapter rather than a game changer, he says a downturn could be months or quarters away, though he does not claim to know the exact timing.
- [1:22](https://realalpha.app/v/K2OC5yKrbaM/?t=82) **Harris says BCV struggled to pay up.** Harris says BCV has unrealized investments such as Crusoe that he expects to return multiples of its fund. He suggests risk-management habits from Bain Capital’s other strategies may have influenced venture decisions, and identifies historical valuation sensitivity as the specific weakness.
- [4:33](https://realalpha.app/v/K2OC5yKrbaM/?t=273) **Power-law winners can look overpriced.** Harris says winners such as Databricks, Stripe and Ramp can look expensive at every round, even though their founders’ ambition and leadership were visible at the time. He argues that rejecting every deal deemed overpriced can systematically exclude power-law outcomes. He warns that treating a higher price as a signal of quality can push prices upward and inflate a bubble.
- [8:24](https://realalpha.app/v/K2OC5yKrbaM/?t=504) **AI leaders can still be challenged.** Harris agrees that venture is in a consensus market, where investors crowd into one or two anointed leaders, but rejects the idea that there are no diamonds in the rough. He says leadership remains dynamic, citing Ligura’s gains against Harvey in parts of legal AI. He argues investors need an open mind about how AI companies can overtake incumbents.
- [11:42](https://realalpha.app/v/K2OC5yKrbaM/?t=702) **AI reshapes the investment lens.** Harris says BCV evaluates investments on the premise that AGI is here or will be, and that abundant superintelligence could arrive in one to three years. He says physical-world businesses and AI-native services made up 75% of the firm’s investing over the prior 18 months, with robotics a longstanding area that includes BCV’s earlier Kiva Systems investment.
- [15:52](https://realalpha.app/v/K2OC5yKrbaM/?t=952) **Services firms can own the work.** Harris says AI-native companies can expand a market and build defensibility by doing the work in an industry, rather than only selling it technology. He describes Reserve, an AI-native insurance claims business incubated with a design partner, and says it was on track to do well over 100,000,000 in revenue in its second full year. BCV also backed law firms, an earthmoving company, a home health agency and other insurance services businesses.
- [18:08](https://realalpha.app/v/K2OC5yKrbaM/?t=1088) **Labs may avoid customers’ industries.** Harris thinks foundation-model labs may hesitate to start law firms or insurance companies because major companies in those industries are customers for their models. He compares this with Google and Facebook’s reluctance to challenge financial-services advertisers, and Amazon’s restraint in banking because JPMorgan Chase is valuable to AWS. He says labs already have other areas to pursue, including intelligence, applications and consumer products.
- [25:54](https://realalpha.app/v/K2OC5yKrbaM/?t=1554) **Compute could decide the AI race.** Harris says differential access to compute may become the AI labs’ prime battlefield over the next five years, replacing talent and research leadership as the key edge. He says faster model training could compound an advantage. He also praises OpenAI’s move to let Codex users use open-weight models, describing it as an open-platform strategy.
- [32:44](https://realalpha.app/v/K2OC5yKrbaM/?t=1964) **Harris says Anthropic’s $2 trillion price is reasonable.** Harris says Anthropic’s reported revenue and growth make a $2 trillion valuation reasonable if that growth continues, while stressing the uncertainty of forecasting a hypergrowth company. He says the eventual supply of shares may depend on how many investors seek liquidity when lockups end. He notes that staged releases and price targets can help manage that supply.
- [38:54](https://realalpha.app/v/K2OC5yKrbaM/?t=2334) **Agents could expand the user base.** Harris agrees adoption may still be early and argues that agents could expand the market beyond 8,000,000,000 humans to 80 or 800,000,000,000 new entities. He says cheaper open-weight and older models can make intelligence more accessible, while demand and supply both grow. He says the value of frontier models will depend on how much their capabilities advance.
- [44:50](https://realalpha.app/v/K2OC5yKrbaM/?t=2690) **New model capabilities shape demand.** Harris says older and open-weight models may handle tasks AI can already do, but the next frontier models’ capabilities will determine what new work becomes possible. If models can reliably complete two-month reasoning tasks, he says they could change fields such as drug discovery and insurance underwriting. He argues established companies could use AI to launch businesses they previously considered out of reach, such as insurance.
- [52:38](https://realalpha.app/v/K2OC5yKrbaM/?t=3158) **Steady pacing limits venture-cycle risk.** Harris says nobody knows whether AI investment is in a bubble and sees a greater than 10% chance that venture returns will be depressed in this vintage. He says a power-law winner can still make a fund, and that BCV invests the same 500,000,000 a year. He argues steady pacing helps guard against market cycles and investor behavior.
- [57:30](https://realalpha.app/v/K2OC5yKrbaM/?t=3450) **Agency matters as AI lowers barriers.** Harris says agency—the ability to choose what to do and act—is a key skill as AI makes tools and knowledge widely accessible. Drawing on raising six children, he recommends Montessori preschool and answering children’s questions with questions of your own, which he says encourages them to find answers. He also advises avoiding privileges, such as private aviation, that could become a child’s baseline.

## Transcript

How I Invest, hosted by David Weisburd, with Matthew Harris.

From the transcript How I Invest publishes with the episode.

[0:00](https://realalpha.app/v/K2OC5yKrbaM/?t=0) All it would take is one misstep to sort of blow the top off this bubble, like, you know, November 2021. Today, if I use analogy of a baseball game, what inning would you put us in terms of the bubble of AI? I'll answer one way, which is to say, if this is like other technology waves, we're in the eighth inning. If this is just like any... Like, the SaaS transformation, which I would say sort of started twenty twelve, twenty thirteen, and then ended in extraordinary fashion in November 2021, aided by COVID, aided by rates going down, all sorts of factors. If AI turns out to be like SaaS, which is to say, yet another chapter in the book of technology, but not a game changer per se, that probably won't be November 2026, but we are definitely, like, months and quarters away.

[0:51](https://realalpha.app/v/K2OC5yKrbaM/?t=51) How would you describe that that so many of the winners are, quote, unquote, overvalued? They have extraordinary founders. They have extraordinary ambition. Those elements that led to them being these hyperscale winners were present at the time. Multiple thoughtful investors saw those characteristics, the extraordinary nature of the Colisey brother leadership or Eric at RAMP and said, you know what? I will definitely pay 50% more on a multiple basis to back these individuals because of the scope of their ambition and the quality of their leadership. But we also have this phenomenon where the winners are So, Matt, when I looked at your performance at Bain Capital, you guys are perennially top quartile, but you don't seem to have any power law outcomes.

[1:42](https://realalpha.app/v/K2OC5yKrbaM/?t=102) How is that? I will say we we have unrealized companies now, Crusoe being one example where we led the pre seed and the seed that will surely return multiples of its fund. So plenty of examples where in the future, once they're liquid, we can talk about them. There's not so much what other returns, particularly in light of the fact that in recent years, have been fund returners just not yet liquid. But the why behind it. I imagine this is a pretty big issue. You guys must have discussed it internally. Why do you believe this has happened at BAND Capital? I think certainly going back in time, there were elements of risk management that were quite unconscious.

[2:23](https://realalpha.app/v/K2OC5yKrbaM/?t=143) I mean, gets into venture capital. Anyone who ends up being successful in venture capital with a risk seeking frame really is about actually looking for audacity, seeking out improbable founders and improbable outcomes and non consensus companies and markets and even the shape of the humans you're funding. And that certainly characterizes our partnership. But we were born out of Bain Capital, which of course in its other asset classes does have aspects of risk management which are good and proper. If you're running a private equity strategy, avoiding losses is prudent. If you're running a credit strategy, avoiding losses is critical.

[3:09](https://realalpha.app/v/K2OC5yKrbaM/?t=189) You can't really make up for a cluster of losses with vast returns elsewhere when you're in a fixed income world. So our question, I think, and I and I have to assume that there was some element in the early days and even the middle days of being capital ventures that sort of through osmosis carried some of that flavor of risk management unconsciously into which deals we brought forward and which deals passed investment committee. To be specific, though, I think it was really about valuation sensitivity. So it was actually kind of less about, you know, avoiding outlandish projects that ended up working because we did Crusoe and, you know, ten years ago, and that was when it started as an example of a company.

[3:52](https://realalpha.app/v/K2OC5yKrbaM/?t=232) You know, they were mining Bitcoin at the oil heads in, you know, Arctic Canada. So it was an outlandish proposition, to be sure. And those are actually the sort of thing we like. But where we struggled historically is sort of paying market price when market price got really high. And I think that that's a fundamental problem actually in a power law industry because price is a signal of a sort. It can obviously be wrong. But the fact is, like Databricks and Stripe and Ramp, I mean, the companies that end up being massively valuable are overvalued, quote unquote, at every round. And so if you take a view that you're gonna turn down things that are overvalued at the

[4:40](https://realalpha.app/v/K2OC5yKrbaM/?t=280) **Matthew Harris:** very moment in time they're being valued, you will systematically miss the power law companies on average.

[4:46](https://realalpha.app/v/K2OC5yKrbaM/?t=286) How would you describe that so many of the winners are, quote, unquote, overvalued? I mean, we're talking about semantics here. I mean, in retrospect, they were undervalued. Right? But But on on a multiple basis. Because they have extraordinary founders. They have extraordinary ambition. Those elements that led to them being these hyperscaled winners were present at the time. And thoughtful investors saw those. And multiple thoughtful investors saw those characteristics, the extraordinary nature of the Collison Brother leadership or Eric at Ramp and said, You know what? I will definitely pay 50% more on a multiple basis to back these individuals because of the scope of their ambition and the quality of their leadership.

[5:37](https://realalpha.app/v/K2OC5yKrbaM/?t=337) So it has to do with some subset of the market actually being perceptive. It leads to another problem in our asset class, though, which is that if you say to a venture capitalist and I'd be curious your perspective on this, if you say to venture capitalists, all the best deals are 50% too high on a multiple basis. Well, that's going to lead to some perverse behavior. Right? Because you brought a deal to committee. So your partner brought a deal to you and said this, you know, we we've looked at application layer AI companies and averagely 20 times revenue is sort of where the average, you know, application layer AI company is going.

[6:13](https://realalpha.app/v/K2OC5yKrbaM/?t=373) We've got a really interesting one. It's 20 times revenue so you know perfectly appropriately priced. Like by rights you should say look unless it's 30 times we shouldn't do it. Because my all my pattern recognition tells me that it unless it's 50 above market it's not a power law company so we should pass on anything that's not 50% too expensive. That wouldn't ever be like written down in an investment committee charter of course but the fact is like that mindset is what leads to bubbles in our asset class because people have a price quality relationship in their mind and therefore almost the higher the better.

[6:51](https://realalpha.app/v/K2OC5yKrbaM/?t=411) This episode is presented by Juniper Square, the operations partner for private markets. Early Bird, a firm firm in Europe, did did a study on this. They figured out that that this price persistence continues to persist. In other words, like, every round is overvalued versus the comps. Said another way is that the overvaluation is in relative to the competition set. But because in venture capital, the outcomes are so big and the markets are so big, even if another competitor could have gotten a 100 x on that return, but this competitor would get the 75 x if they win. It's smarter to go with the team that's likely to win than the team that would have theoretically given you the 100 x.

[7:37](https://realalpha.app/v/K2OC5yKrbaM/?t=457) Very clearly. This is... The physics of power law mean that you should be extremely price insensitive if you have a greater degree of confidence that you're backing the power law winner. Now that's true, but obviously everyone can't be right about that. So if everybody's playing that same game, but most people are wrong, and everyone thinks they're backing a power law winner, then every price goes up by 50%. And then the next time you play that game, the starting point is 50% higher and everyone goes up 50% from there because they've convinced themselves they have the power law winner. That's what inflates a bubble.

[8:15](https://realalpha.app/v/K2OC5yKrbaM/?t=495) When you look at what appears to you to be a bargain and your reaction instead of feeling greedy and excited is to feel suspicious and aversive, then you have a kind of a market where prices are going to trend up no matter what. I want to get back to this question on whether we're in a bubble, because I think it's an important question. But first, it seems undeniable that we are in a consensus market. Mark Andreessen says there are no more diamonds in the rough. There are only diamonds. Do you agree with that? I agree that we're in a consensus market. I disagree that there's no more diamonds in the rough.

[8:52](https://realalpha.app/v/K2OC5yKrbaM/?t=532) How do you differentiate those things? I observe that we're in a consensus market, which is to say a company is anointed. This phenomena has a name. It is king made. And maybe there's two, maybe there's one, but there's a leader anointed, often two, and then money piles in. And there are firms whose strategy is to be in the leader or number two with a chance in a sector, and they'll pay almost anything to get into one or two.

[9:23](https://realalpha.app/v/K2OC5yKrbaM/?t=563) **Matthew Harris:** And not for love or money would they invest in anyone who's not already anointed in the winner's circle.

[9:30](https://realalpha.app/v/K2OC5yKrbaM/?t=570) So I describe that as a consensus market, and that's where we are and actually would differentiate that from 2021, where it was much more of a thousand flowers blooming. I mean, were there were firms doing, you know, two two unicorn deals a week. So there were just a lot more. It was a breadth of the market that was getting sort of pumped up and valued at high levels. It wasn't that people were really radically disagreeing with each other, but it wasn't narrowing in each sector the way it is today where you have a sort of one, two, rarely a three. So I observe that it's consensus market. But Mark's statement goes further, which is to say that this is as it should be, that there's no point in doing anything other than backing the, at the moment, obvious winner.

[10:19](https://realalpha.app/v/K2OC5yKrbaM/?t=619) And we have seen things as more dynamic. We backed Ligura. One of our companies as well. Oh, excellent. You know, I... That was not really consensus. I would say the consensus when we invested, Shirley, when you invested, was that, well, Harvey had won. Yep. And that this European upstart had managed to land a few European logos. But really, that's just because Harvey hadn't gotten there yet. And and now what we observe is that Liguria is catching up and in certain ways is the leader in certain segments of the market and I think has a case for a kind of a second mover advantage. And we're seeing that in some other markets, financial analysis, instance, the companies that would have been anointed as obvious maybe are feeling less obvious now.

[11:11](https://realalpha.app/v/K2OC5yKrbaM/?t=671) So I think this idea of there are no diamonds in the rough, there are only diamonds takes a static view. And we observe this to be extremely dynamic to where leadership is actually not handed down from on high and protected forever, but is actually constantly under siege. So that's what I mean by there are I don't mean by diamonds in the rough that you should go to Des Moines, Iowa, and start looking for the next, you know, inference company. It's not really like a contrarian in the sense of geography or founder archetype, etcetera. It's more, I think, having an open mind about the dynamic nature of AI investing in particular.

[11:55](https://realalpha.app/v/K2OC5yKrbaM/?t=715) If you think about some new principles of investing that did not exist pre this AI market, what are some new principles that guide you as investor? The way we frame it, perhaps not unique to us, is in every one of our investments, stipulating that AGI is here or will be. That, in fact, AI is no longer a project or a premise or even a tool or an ingredient, but rather that in in various ways, they'll become obvious over time. Superintelligence will be here. Abundant superintelligence will be here. And so the question is, what wins under those conditions? What are the types of companies you want to back under those conditions?

[12:47](https://realalpha.app/v/K2OC5yKrbaM/?t=767) Not under the conditions that AI exhibited two years ago, where it was still, I think, question about not about the direction of travel, but about the speed of that travel and most importantly, whether it would plateau. This has been a really ongoing part of the narrative about AI is slightly less now. But if you go back and look through time, if you just take the New York Times as an example, you could find on different days of the same week articles that said AI is plateauing, is laughable, is a, you know, stochastic parrot, and clearly is not any good. And then the next day, A. I. Is dooming our civilization and will ultimately take over everything.

[13:34](https://realalpha.app/v/K2OC5yKrbaM/?t=814) And it's sort of like, well, which one of those do you really have conviction in? Or are you just being like sort of negative? And so this first category of critique, which has waned but persists, is the AI is a joke. It's a chatbot only. I mean, it's really harder and harder to argue that it's not an important technology, but there are still people seeking to relegate it to quite marginal status in terms of its importance. Our view and the view of most venture firms is is quite the opposite of that, that actually, if if every observer knew what was happening at the frontier, they would take the view that superintelligence is perhaps not comprehensively imminent, but is certainly inevitable.

[14:23](https://realalpha.app/v/K2OC5yKrbaM/?t=863) And in the next one to three years. So that that is our view. And so as we develop this perspective, call it eighteen months ago, we asked ourselves the question you just asked me, which is, Okay, so then what? And I'd say there are two themes that emerge that explain 75% of our investing in the past eighteen months. One is the physical world. So we've always invested heavily in robotics. We, twenty years ago, invested in Kiva Systems, which was the first sort of warehouse robotics company that Amazon bought for over $1,000,000,000 and consistently throughout in manufacturing and distribution robotics.

[15:03](https://realalpha.app/v/K2OC5yKrbaM/?t=903) And here in the past eighteen months, just another kind of Cambrian explosion of amazing robotics companies, including household robots and all sorts of really kind of unprecedented advances. And so that's, you know, maybe a third of what we do. And then we've made 20 investments starting in late twenty twenty two in what people now call AI native services companies. This whole question of like, if you can actually compete in a major industry instead of just selling technology to that industry, depending on what the industry is, it's just a massive TAM expander and allows you to have a more defensible business, you know, relative to the labs themselves.

[15:46](https://realalpha.app/v/K2OC5yKrbaM/?t=946) We worry about the Foundation Model Labs and the pressure they will face to build application layer functionality, horizontal and vertical. And one element of defensibility is to move out of application layer applied AI and into actually doing the work. And so we our first version of this is a company called Reserve, which was actually a fun exercise that's still worth doing. I think we, you know, I'm a financial services investor, so we mapped all the financial services and all the business models. We And asked ourselves, if you had a machine that could magically read and write, what are the activities within financial services that would be most vulnerable to this machine?

[16:28](https://realalpha.app/v/K2OC5yKrbaM/?t=988) And the answer in our mind was insurance claims, which is really just a human swivel chair multi screen. You get a notice of loss on a policy. You have a policy. You have to read those things. You have to synthesize them into a loss adjustment. It is like what version one generative AI, if you remember, that was what it was born to do. And so we started incubating this with a with a customer, a design partner customer and AI native insurance claims business that ended up being called Reserve.

[17:04](https://realalpha.app/v/K2OC5yKrbaM/?t=1024) **Matthew Harris:** And when we incubate companies, we don't take any common stock. We just put money in. We recruit a founder and then we give them all the common stock.

[17:10](https://realalpha.app/v/K2OC5yKrbaM/?t=1030) And so they really are the founder. And that company is they'll do well over 100,000,000 in this year in revenue and their second full year of life. And it's been an extraordinary growth curve. So we were heartened by that. And we went on, as I said, we've done 20 more of these, including two law firms and automated earthmoving company and a home health agency and six or seven more insurance related native services companies. And so we think those companies and actually so far, it's been very clear that they've benefited from advances made by the Foundation Model Labs, and we want to be levered to that, not fighting against that.

[17:49](https://realalpha.app/v/K2OC5yKrbaM/?t=1069) That's where I wanted to go, which is at some point, advanced intelligence gets so advanced that it's doing everything, and we're we're like that famous essay where we're just these... We're back in the jungle, and we have these overlords. But until that happens, why is going directly to services... Why is that a moat? And maybe you could double click a little bit about how your strategy relates to the large LM models. Well, it certainly calls for speculation. Think these companies have the theoretical ability to do anything. So what we can do is look at their incentives. Right now, a vast amount and fast growing percentage of the revenue for OpenAI and Anthropic is coming from large corporates who either directly or via API through partners are accessing their models to automate core business functions.

[18:52](https://realalpha.app/v/K2OC5yKrbaM/?t=1132) They will have to make a choice at some point along the way or at many different points along the way about which one of those industries they want to actually compete directly with. So right now, Anthropic is making a lot of money from the legal industry.

[19:07](https://realalpha.app/v/K2OC5yKrbaM/?t=1147) **Matthew Harris:** And our view, probabilistically, is that that will inhibit them in starting their own law firm.

[19:17](https://realalpha.app/v/K2OC5yKrbaM/?t=1157) And the same can be said of many other verticals. They make a lot of money from the insurance industry, and that will likely inhibit them from competing with Travelers, who's a massive customer of theirs, competing with Aon and Willis Towers Watson, who are massive customers of theirs. They feel like they have enough fish to fry selling, of course, just base intelligence. And then building vertical applications on top of that to sell to those corporate end users, not to mention consumer businesses, which they also have. So they may look at those three vectors, the intelligence vertical applications, consumer and horizontal applications like coding applications, for instance.

[20:02](https://realalpha.app/v/K2OC5yKrbaM/?t=1202) They may look at that and say, that's not enough. We also have to start a law firm or a home health agency or an earthmoving business. But our probabilistic view is that will be pretty long down the path and may never happen. Like, for instance, in fintech land where I've been living for a long time, there's always this question of, you know, if you look at Google and Amazon and Facebook and Apple, which one of those will move aggressively into consumer financial services? And the ones that have always had the most capability actually were Google and Facebook. They had incredibly broad. You could argue Apple, but for a variety of reasons, doesn't like to use data.

[20:50](https://realalpha.app/v/K2OC5yKrbaM/?t=1250) And so it's they've always moved very, very carefully in financial services. But the most capabilities resided at Facebook and Google and they never have. And they've made little forays here and there, particularly Google. They'll like announce something every two or three years and shut it down. And the reason is they're an advertising company. They make their money from the financial services industry in, you know, 20, 30% of their revenue. And so the incentive they have to go poke that industry in the eye by directly competing is a real problem. Is that because they've run this calculation, they've realized the best business model is advertising business?

[21:28](https://realalpha.app/v/K2OC5yKrbaM/?t=1288) Or is that because culturally, they have this whole ad team that has 1,500 employees, and they can't cannibalize their own business and culturally just wouldn't work. Which one of those reasons is driving that? I mean, they're related and inextricable. It is the cash cow, right, which means it's got this power base. And it is observably one of the best business models in history. It's 100% margin business. So those two things are like inextricable and together formidable and so far successful in keeping Google out of offering direct consumer financial services despite having they bought one of our companies so they could get into the business and then they shut it down because of this angst.

[22:08](https://realalpha.app/v/K2OC5yKrbaM/?t=1328) Now Apple has a credit card. As I said, I think they've done it in inelegant ways because of their own fear and inhibition. But they don't care about the credit card companies. They don't really focus on selling advertising. So they're more gloves off, I'd say, as it relates to competing in financial services because they're not wed to that entry. Yes, sure, they sell some MacBooks to venture capital firms, broadly speaking, they their butters, their bread is not buttered by the financial services industry. And Amazon is another fascinating example. If you go back ten years, they had massive plans in consumer payments.

[22:47](https://realalpha.app/v/K2OC5yKrbaM/?t=1367) They were going to be a proper disruptor in consumer financial services. And you know what killed that? AWS. All of a sudden, their most valuable business had financial services as a core customer. And so it wasn't the cash cow that advertising is at Google. But at the top of the house, you're starting to have really hard conversations about should we launch our own banking products? Should we get consumers to keep money with us? We'll pay them interest and they spend it on our products. Basically do what Starbucks did with their app. Amazon's never really done any of that, even though the plans are still on the shelf because JPMorgan Chase is just too valuable a customer if you're running AWS and if the person running AWS then becomes CEO of the company.

[23:38](https://realalpha.app/v/K2OC5yKrbaM/?t=1418) So these are the conflicts that we look to historically to give us some confidence about where to put our money today based on predictions about what the Foundation Model Labs are going to do and not do. And we, having thought about it, resolved that they are unlikely to start law firms, for example. One thing I've learned from talking to hundreds of investors is that great investment firms aren't built on investment returns alone. The firms under are great at the things most people don't see, their operations, their relationship with LPs, and the quality of information they use to make decisions.

[24:12](https://realalpha.app/v/K2OC5yKrbaM/?t=1452) And here's what AI has changed. Every firm now has access to the very same models, so the intelligence isn't the edge anymore. The edge is what you could feed it. Firm with its fund operations and data in one connected record can actually put AI to work. A firm running on disconnected systems simply can't. That's why thousands of GPs run their funds on Juniper Square. Juniper Square puts your fund operations, data, and administration together in one connected record. That means less time managing disconnected systems and more time investing, working with LPs, and building your firm. This episode is brought to you by Juniper Square, the operations partner for private market GPs.

[24:51](https://realalpha.app/v/K2OC5yKrbaM/?t=1491) Learn more at junipersquare.com/howiinvest. That's junipersquare.com/howiinvest. Everyone I talked on the show is chasing the same thing, an edge, and more and more of the edge comes down to your information, not just having it, but being able to trust it when the stakes are highest. AI is doing more of that information gathering for you every day, and most tools are very good at sounding right. The summary reads clean, but can you trace it back to filing the transcript, the specific passage that drove the answer, or are you just trusting the confidence of the output? For investors, that's not a minor concern.

[25:25](https://realalpha.app/v/K2OC5yKrbaM/?t=1525) A miss filing, miss weighted source, context that got lost somewhere in the retrieval chain, those aren't edge cases, they're how decisions go wrong. AlphaSense is the AI market intelligence platform built specifically for this. They own the content over 500,000,000 curated documents from broker research and expert transcripts to filings and earnings calls, and they own the retrieval layer on top of that. So every answer links back to an exact verifiable source because the answer is only as good as what's underneath it. And with AlphaSense, you know exactly what that is. The edge goes to whoever can trust their information and prove it.

[25:57](https://realalpha.app/v/K2OC5yKrbaM/?t=1557) See it for yourself. Start your free trial at alpha-sense.com/howiinvest. That's alpha-sense.com/howiinvest. Many of the smartest people in the world actually think that industries are not the bottleneck for the large language models. It's actually gonna be compute. Even Elon has said this, which is why he's he's building the TerraFab and doing, data centers in space. Do you not think that compute itself is the bottleneck? They're separate questions. I think that you have a focus question within a lab like Anthropic about given a fixed level of compute, where are they gonna spend their time? It could be that the level of compute is too low for them to even entertain starting a law firm.

[26:43](https://realalpha.app/v/K2OC5yKrbaM/?t=1603) I don't think that will be the reason they don't start a law firm. But I do acknowledge that the differential access to compute is going to be perhaps the prime battlefield in the next five years. I would argue the prime battlefield in the last five was talent and leadership. That it was about having the best researchers, having the best research agenda and the best leadership that separated the wheat from the chaff. But I think it gets a little bit more mechanical here forward if you have the compute to train the most impressive model in the shortest period of time, I think that's going to create compounding advantage.

[27:24](https://realalpha.app/v/K2OC5yKrbaM/?t=1644) And so the chess move here about one of the three big ones getting ahead and then getting lapped by another. And by the way, here comes Meta and Google and their resources, I think, among the more fascinating things I've ever seen, particularly when you nuance it by all the external players, the private credit industry, infrastructure investors, insurance company balance sheet and what they're funding, the chip companies and how they're picking favorites or not picking favorites or buying companies to secretly compete. I mean, it it really... Wish I could buy the book that'll be written three years from now.

[28:01](https://realalpha.app/v/K2OC5yKrbaM/?t=1681) There's a lot of layers of complexity there, and I'm tempted to double click on a bunch of them. But let's just keep it to OpenAI in Anthropic today. What's the game theory between them and how they're looking at making these massive investments in compute? There are a lot of games being played. The original game, I think at both companies actually is about advancing the state of artificial intelligence and actually achieving generalized artificial intelligence. And that has become a little muddier as capitalism has kicked in, both for the individuals running these companies, more so in the case of Anthropic than OpenAI, but also just their competition itself and the human desire to win and not be second place.

[28:51](https://realalpha.app/v/K2OC5yKrbaM/?t=1731) And then these massive cap tables who have unambiguous goals, which are just to make as much money as possible. So I think the original game of being first to artificial general intelligence has been subsumed in a revenue and profits game, or at least complemented by it. And in doing so, the agendas have definitely shifted. OpenAI was more aggressive out of the gates in terms of being having audacity around funding. They raced ahead. Anthropic caught up both with the product and now more recently with their ability to go get capacity. I think Elon is the best of the three at this at that particular game, both of building data centers specifically, but also about sort of just, you know, world building in people's minds and attracting vast capital to vast projects.

[29:40](https://realalpha.app/v/K2OC5yKrbaM/?t=1780) So of the three, think on the data center game, I actually think the advantage sits with Elon. OpenAI, Verseanthropic, They have a lot of business puts and takes right now. Think one of one one announcement here recently that I think was underreported on was OpenAI opening up model usage within Codex. So OpenAI wants to be a bunch of different things. They want to sell the most tokens and their most expensive models, but they also want to be the most popular, you know, coding tool. And they made a really strategic decision, which is to let people who are using codecs to code more efficiently use open source, open weight model tokens in addition to just OpenAI token.

[30:25](https://realalpha.app/v/K2OC5yKrbaM/?t=1825) And that is a STARK departure from the historical, you know, cursor sells to Elon and the models shut off access to their models to cursor. That's where we were quite recently. And now in a few weeks, we're in a place where at least OpenAI has said we want to win at the horizontal application layer, which means we have to go multi model architecturally. So I really like that move.

[30:55](https://realalpha.app/v/K2OC5yKrbaM/?t=1855) **Matthew Harris:** I think it's genius, actually, and I think they're still going to sell a lot of OpenAI tokens into codex in the same way that ostensibly and actually Microsoft was open

[31:09](https://realalpha.app/v/K2OC5yKrbaM/?t=1869) architecture with its OS. You could use WordPerfect. You could use Lotus one, two, three. You could use freelance graphics. But over time, the market share shifted to Microsoft Word, Excel and PowerPoint because they own the playground, they own the operating system. And that was just a massive advantage, even if they weren't overtly bundling them in a way that triggered the regulators and in a way that actually forced lock We'll see what Anthropic does, but these are the kind of chess moves that are happening right now that I think make this such a fascinating place to watch. I mean, there's a rumor going around the valley right now that both companies are going to shut off API access to vertical apps.

[31:52](https://realalpha.app/v/K2OC5yKrbaM/?t=1912) That's a paranoid rumor that I don't think makes any sense because the API businesses are their biggest businesses. So how would that work? What's incentive? Well, the incentive would be you can only get the frontier models from applications they build. And so if you're Goldman Sachs and you're buying an AI application layer tool and you can go to a you know, you can go to Hebia, but you know, they can't get anthropic tokens or you can use anthropic for finance and get, you know, those sweet, sweet anthropic tokens. That would be the logic. Again, I'm I'm skeptical. I think that's it. It would be a bad business decision.

[32:36](https://realalpha.app/v/K2OC5yKrbaM/?t=1956) And OpenAI has just shown they're actually going in the other direction, much more open architecture versus closed. But these are the kind of theoreticals that are getting mooted right now. And I think I just as a student of both business history and military history, it's like, you know, catnip. I didn't wanna disclose, my investment in Anthropic because I wanted to hear hear unfiltered unfiltered opinion, but let's assume that goes public at reported $2,000,000,000,000. I don't have any insider information on that. How do you think about the valuation, and what will be the drivers on whether that's overvalued or undervalued?

[33:16](https://realalpha.app/v/K2OC5yKrbaM/?t=1996) I think the valuation based on the reported revenue numbers and revenue growth is actually reasonable. You know, these are unprecedented. Is that just because the sticker shock on $2,000,000,000,000 is just so great they had to price it like that? To be clear, there is sticker shock about 2,000,000,000,000. But I think if you look at the underlying business momentum, other than the sticker shock of that being a kind of an unprecedented number, It's actually reasonable. Like if you took three zeros off of all the numbers, you would say, well, actually, that makes sense.

[33:49](https://realalpha.app/v/K2OC5yKrbaM/?t=2029) **Matthew Harris:** If that growth rate can continue or anything like it can continue, they will grow into and through that valuation in a very reasonable period of time.

[33:59](https://realalpha.app/v/K2OC5yKrbaM/?t=2039) So I don't think it's unreasonable at all. You know, we have a hard time as investors pricing companies that are in hyper growth because it's really hard to do any kind of forward. Goes back to where we were talking before. Correct. Seems overvalued. It might see overvalued at a moment. But again, if hypergrowth continues, it'll quickly seem undervalued. So I think we just you know, we just don't know. And so I don't think it's unreasonable. And I am sure there's demand. I think what will be interesting to see technically is how much of the cap table is in strong hands versus weak hands. And therefore, what the supply demand picture looks like when the lockups start to come unlocked.

[34:40](https://realalpha.app/v/K2OC5yKrbaM/?t=2080) You know, SpaceX has done well so far. But I just don't know enough about the detail of the anthropic cap table to know how many, you know, venture firms and other investors who haven't produced a lot of liquidity in the last, you know, five years are really eager to produce some liquidity. Another set of incentives. And as a small shareholder, not not so small outcome, I I could see if you invested 50,000,000 or a 100,000,000, that number is very difficult to walk away from even if you think it's undervalued. Yeah. I think that... And you're the LP and you're sort of saying, alright. It's a lot of TVPI, but we haven't had distributions from this asset class.

[35:24](https://realalpha.app/v/K2OC5yKrbaM/?t=2124) So let's get going with this. And this is how this works. You give us money back and we give it back to you, but you need to give it back. So again, that doesn't make me pessimistic. I just don't know the math around this week. It's more of a short term volatility. Correct. But these things can have a... Actually, I'd say medium term effect if you have a cap table full of weekends that need liquidity, particularly when the numbers get that big. Because how do you replace, you know, in a theoretical, say you had half of the company were liquidity seeking owners in a period of two years, you know, between IPO and two years after IPO.

[36:02](https://realalpha.app/v/K2OC5yKrbaM/?t=2162) That means you got to go. If the stock stays flat, you got to go source another trillion dollars of liquidity. That's not just a valuation question. That's just a like cash. Where's that coming from? Probably not coming from T bills. So it's coming from somewhere. So there's some like physics of the market that but by the way, that 50% was like total speculation. I'm I have no no intuition that half the cap table is a seller in the first two years. But, you know, we're an unprecedented territory, and so we just don't... We don't really know. I'm still shocked that the standard of no liquidity for VCs and then at six months, everybody gets unlocked.

[36:42](https://realalpha.app/v/K2OC5yKrbaM/?t=2202) That is the dumbest model I've ever heard that that has persisted. Couple companies recently had... Have changed that. Cerrebris, one of our other companies that it did this unlocking over over many tranches. Some tranches were as small as 6% Great. Of the holdings. SpaceX also did tranche. Do you think that this trend could continue, and do you see any of your portfolio companies speculating? For sure. Yeah. I think not only tranching, but price targets. We did an IPO in many IPOs, but one in particular in 2021, where we create a very thoughtful ladder of price targets and liquidity only comes with price targets and price targets that are hit with some duration and some level of volume.

[37:25](https://realalpha.app/v/K2OC5yKrbaM/?t=2245) And so you you can actually really mitigate the risk of drawdowns when you do that kind of careful titration of how stock gets fed into the market. And there's no doubt that feeding stock in sooner in a titrated, careful way is way better than not feeding it at all and letting it build up. So I think it's actually the underwriters can be a little hard to deal with on this, that the investment banks, you know, who run these IPOs actually, from a regulatory perspective, they are the underwriters. And so that that word has some meaning still, and they have to be very careful about this. They're often They're the counterparties to some of these transactions.

[38:07](https://realalpha.app/v/K2OC5yKrbaM/?t=2287) They're the ones who'll get sued if something novel gets tried and doesn't work. So that's actually I could see how the conversation went with Elon and his team and the banks. Well, this is what we have to do. Well, this is what you're gonna do. Okay. Yes. And if Dario wants to play But sometimes that's how standards are set. And then everyone's like, I want I want Elon's deal. And then the industry changes like magic. And I think that'll... I mean, it happened for every run of the mill issuer, but surely either Anthropic or OpenAI or both could could get a similar deal. Going back to Anthropic and Anthropic's evaluation, not just because I selfishly am focused on this.

[38:48](https://realalpha.app/v/K2OC5yKrbaM/?t=2328) Not because I I need to do some tax planning and portfolio planning, but also because I had... I previously had Lucas Swisher who runs growth for Co2, extremely smart investor. And one of the things that he said that I'm still processing, it blew my mind, is that despite these revenue numbers at these large labs, OpenAI, Anthropic, and others, he still believes we're extremely early. And the reason he believes that is because the majority of the revenue today is still coming from developers and engineers. And the penetration of those engineers is just so tiny, even in The US. So you're having a small part part of a small market, and that's leading the revenue today.

[39:37](https://realalpha.app/v/K2OC5yKrbaM/?t=2377) Now what happens when just all of engineers are unlocked and then all of other knowledge workers? I think there's only, like, 50,000,000 or so engineers on the planet out of 8,000,000,000. So you just have a small portion of them really doing a lot of token spend. What do you think about that? Could it could it really be that we're really that early in terms of token usage and AI adoption? I'd say the two countervailing things in my mind when I hear that one is vehement agreement.

[40:07](https://realalpha.app/v/K2OC5yKrbaM/?t=2407) **Matthew Harris:** And at a certain point, increasingly in our companies, there aren't even human users of AI. There are just agents doing stuff.

[40:16](https://realalpha.app/v/K2OC5yKrbaM/?t=2416) And so it's not 8,000,000,000 that you're going, you know, that 8,000,000,000 humans getting more productive. It's actually 80 or 800,000,000,000 new entities being created. And then they're doing deals with each other. And they're doing and they're using tools and creating more agents. Yeah, right. So I actually think that, you know, this this artificial cap of 8,000,000,000 as we think about market size is for the first time lifted because of agents that have proven they can do long running tasks with great efficacy. So I agree with Lucas and I would amplify it. And then the one caveat is open way models.

[40:56](https://realalpha.app/v/K2OC5yKrbaM/?t=2456) So intelligence is becoming more abundant. It's also becoming cheaper. Those who are related, it's abundance and people's desire and ability to access it gets greater as the price comes down. So there is this beautiful expansionary quality to open weight models and even just the token cost of old anthropic models being so much less expensive than new anthropic models. But that mix shift and exactly how that mix shift happens for what applications is the $2,000,000,000,000 question. I think there's no doubt that demand is going to expand and that will have a benefit to pricing as demand expands.

[41:39](https://realalpha.app/v/K2OC5yKrbaM/?t=2499) It helps pricing, but supply is massively expanding as well. And so the value of a frontier model is what we don't know. If the twenty twenty seven frontier models are as transformational relative to the 26 models as the '26 models were to 2025, then I think 2,000,000,000,000 is surely just a starting point. We basically went in '25 until about November, December 2025. It's still chatbot territory and kind of like finish your line of code but not independent long running reasoning agents doing their own coding. That was really sort of winter twenty five, twenty six. And that led to SaaSpocalypse and that led to all of these advances in agentic workflows the course of 2026.

[42:36](https://realalpha.app/v/K2OC5yKrbaM/?t=2556) So I don't know what the next leg up is. I don't know when it comes and if it's as dramatic as chatbot to agent. But if it is, then open weight models that don't have the frontier capabilities yet will be a pale substitute for those high priced tokens. It's interesting because we live in such a meme world. So many people, even with podcasts, they just consume this fifteen second clip, the the thirty second clip. So everyone tries to reduce these paradigms into these very simple binaries. But in reality, there's so many factors. Some are negative. Some are positive. Some play on each other. Some counteract each other if they start to compound to a certain level.

[43:20](https://realalpha.app/v/K2OC5yKrbaM/?t=2600) And then there's all these actors and all their game theory. And as one person makes a move, another person makes a move. Almost feel like you need an AI model to tell you what everybody's going to do and what all the AI models will do. It becomes complex quite quickly. Just today, was talking with Claude about data center CapEx. And the question I had was how much of the call it $5,000,000,000,000, which I think is sort of consensus CapEx spend next five years, how much of that, by the way, it could be 7. How much of that will be spent on expanding access? So more and more people using more and more A.

[43:58](https://realalpha.app/v/K2OC5yKrbaM/?t=2638) I. Versus expanding actual functionality. So training models so that A. I. Could do things that at this point simply can't do. At least Claude and it's funny in the caveat to the answer, says, you have to remember, I was made by Anthropic and that surely influences my answer. So that was a brilliant- It's now disclaimers. Yeah, disclaiming very specifically. I think in the IPO quiet period. That's exactly right. The chatbots in the in the quiet period. But it was it came down squarely in the on the functionality side of things. It said that, you know, it's a mix, but that the preponderance, the reason the number is 5,000,000,000,000 and not 2,000,000,000,000 or some other lower number is actually because of the intensity of training models to do unprecedented things, which was my intuition that this isn't just about the spread of A.

[44:51](https://realalpha.app/v/K2OC5yKrbaM/?t=2691) I. And the scope of A. I. It is actually about what A. I. Could do. And that's what's behind these CapEx budgets. And that's scary and just incredibly exciting. I want to bring up a contrarian thesis. There's now this meme in the market that AI has advanced so much that you could use old models to do virtually all of tasks. All the important... All the day to day tasks. A lot of people are are are saying this meme. You said it earlier that even the old anthropic models could do a lot of these tasks. I hear that. And I think also open source has a great role to play there. But at the same time, it conflicts with what you previously said and what is true about capitalism, which is this winner take all markets.

[45:43](https://realalpha.app/v/K2OC5yKrbaM/?t=2743) The number one player gets 70%. Number two player gets 20%. So are we overestimating these old models' capacity, do these tasks, or is now all these industries game theory and that if Hertz is using Anthropic 7.0, now Avis needs to use Anthropic 7.0 as well, and now the old models no... Are no longer have value. It really depends on this question we were just discussing, which is what can the new models do? So I don't think it's as simple as, you know, tit for tat with competitors having to keep up with the Joneses. I think it actually has to do with the functionality. So if you told me that Anthropics 2027 family of models could do two month long reasoning tasks reliably, then I don't think you have a choice for many of the things that are being attempted right now.

[46:53](https://realalpha.app/v/K2OC5yKrbaM/?t=2813) I don't think you have a choice but to use those models because they can accomplish things in drug discovery and in insurance underwriting and in all sorts of other historically human fields of endeavor that an open weight model will struggle to do in smaller chunks and will still be, you know, arming the human to do better versus actually replacing what has historically been exclusively human tasks. So if you told me that the leap forward in 2027 was massive in terms of functionality, then I would completely agree that for anything AI is doing today, a year from now, they'll be using cheaper, you know, open weight or old models to do.

[47:42](https://realalpha.app/v/K2OC5yKrbaM/?t=2862) But what will AI enable us to do tomorrow? And that tomorrow task will always be a frontier by definition, always be the frontier models. And so what's that mix of the value? Have we plateaued? Have we gotten the most out of it that we're going to get and it's all incremental from here? That's true. 2,000,000,000,000 is way too high. If the pace doesn't even have to steepen, if it just continues on its trajectory, the twenty five to twenty six trajectory, if that pace continues, then we're... Then absolutely both hurts. And even the car rental companies with their relatively modest reasoning tat.

[48:18](https://realalpha.app/v/K2OC5yKrbaM/?t=2898) I use those two companies because it's not the leading cutting edge companies, but they have a lot of data. They have insurance and all those factors that they are data intensive. Yeah. And I actually think the the wrong question is like, how can Hertz and Avis run their business better? The broader question we see and we see this on the private equity side of bank capital, and there's this beautiful interplay between venture and private equity now around it isn't merely how can you run a relatively old school company better. It's what can you now do? Like the idea that you could build a beautiful software product as Hertz or the idea that you could actually launch your own insurance subsidiary versus partnering with AXA, sort of laughable like that.

[49:07](https://realalpha.app/v/K2OC5yKrbaM/?t=2947) We don't we don't make software and we don't manufacture insurance products. We don't do any of those things. We don't know how to deal with the regulators and all the paper they require us to have. We don't have the actuaries in house to do that. We don't know how to price insurance. But then, you know, somebody like you would walk in that room and say like, well, don't you? Have there ever been accidents in Hertz's cars that you don't know about? Or do you have a comprehensive data set of every accident that's ever happened in a Hertz car? Oh yeah, we do. We actually know all that. And we know to the penny how much it cost to fix it.

[49:38](https://realalpha.app/v/K2OC5yKrbaM/?t=2978) We know who the driver was and how long the booking was and who that person worked for. We know all that. It's like, aren't you perfectly situated to launch your own insurance business? And just the fact that you have to file reports with an insurance commissioner in 50 states and another 30 countries, that's like trivially easy with AI. Also could be solved. Yeah. These things. It's like just take your blinders off and say, like, what are we unfairly positioned to do amazingly well? And relax the constraints that you've historically had about what is it feasible for us to do. And so I think this actually, particularly for kind of old school companies, is it's a really wonderful moment.

[50:22](https://realalpha.app/v/K2OC5yKrbaM/?t=3022) And it actually gives me confidence that when you pick Hertz and Avis, you did so. I think it's a it's a telling example. At first analysis, you'd say like, right now they're using AI to like figure out where the cars should be and manage their customer care costs. And their relatively few programmers are getting more efficient or they're pulling things from a BPO and whatever. So, they could do that with open weight models like probably today. But as the models get more performant and more capable, if they have talented management, they'll start quickly thinking, how do we use these models to do things that we never even contemplated?

[51:02](https://realalpha.app/v/K2OC5yKrbaM/?t=3062) Or asking their models what business should we be in? Correct. All of it. This is what people I mean, I remember going back quite a ways when I was learning how to use these agents and just trying to like by the quite a ways meaning like Q1 of this year. But I was like pushing myself in the deep end to like get good at this. And the guy I was working with would always... Every time I asked him a question, he's like, Why are you asking me the question? Ask Claude the question. Like, I am a crutch for you. Like, ask everything that you need to know is in the model. And anytime you pick your head up and ask me a question, it's just a mistake.

[51:41](https://realalpha.app/v/K2OC5yKrbaM/?t=3101) You don't know how to connect this to the box drive? Ask it how to do it. And it will tell you. And it was it's just an incredibly liberating thing to realize that you don't need a coach. You don't need advice. You don't need training. You actually have it all there. You watch, like, a 15 year old use it, and you can see. They just never pick their head up from it because that is everything they need. Going back to the 50,000,000 or so engineers in the world, up until 2026, coding was in code. Today, it's in every language. So in theory, up to 8,000,000,000 people can now code versus a year ago, 50,000,000 people could be code.

[52:23](https://realalpha.app/v/K2OC5yKrbaM/?t=3143) That's another factor that no one's really factored in, which is what happens when the number of engineers go up by a factor of a 100 or 500 x. We're gonna get a lot of useless software, but we're gonna get a lot of useful software too. I think it's amazing. I mean, this is why I like the dystopian stuff. I just don't have time for it. I I like nothing I can do about it. And I prefer to revel in the possibilities, because I think they're so abundant, frankly. Speaking of dystopian possibilities, we earlier talked about this concept of a bubble and how if you keep on overpaying and everybody keeps on chasing momentum, at some point, you're overpaying.

[53:04](https://realalpha.app/v/K2OC5yKrbaM/?t=3184) And at some point, the bubble bursts. Today, if I use analogy of a baseball game, what inning would you put us in terms of the bubble of AI? It's probably a disappointing answer for podcast, but I just have to plead ignorance. And I actually go further and say nobody knows. No one has ever seen anything like this. I'll I'll answer one way, which is to say if this is like other technology waves, we're in the eighth inning. And all it would take is one misstep to sort of blow the top off this bubble, like, you know, November 2021. If this is just like any like the SaaS transformation, which I would say sort of started twenty twelve, twenty thirteen and then ended in extraordinary fashion in November 2021, aided by COVID, by rates going down, all sorts of factors.

[54:00](https://realalpha.app/v/K2OC5yKrbaM/?t=3240) And then it was an excruciating eighteen months after that. If AI turns out to be like SAS, which is to say yet another chapter in the book of technology, but not a game changer per se. That it probably won't be November 2026, but we are definitely like months and quarters away, you know, in terms of my, like, pattern recognition about What do you put the odds of that at? And percent? I mean, I I just think that's How how would it even be possible? Because unlike other types of technologies, I feel like... And this is a... Another question I ask people, which is if everyone's getting so much more productive, why hasn't it showed up in GDP?

[54:42](https://realalpha.app/v/K2OC5yKrbaM/?t=3282) So we'll we'll handle that in a second. But assuming that I'm not myself hallucinating, How is it possible that I'm getting all these productivity gains? I'm getting better, doing more episodes, doing more investments, not using as many resources in terms of human capital, and and at the same time, we're in a bubble? How's that even humanly cost? I don't think it is. That's why I meant, you know, this tiny percentage. On the high end. Yeah. I think... I mean, now are people systematically overpaying because of this power law chasing dynamic in a... So it might be a bubble from an investment side, not necessarily from adoption.

[55:21](https://realalpha.app/v/K2OC5yKrbaM/?t=3321) Correct. So I do think are there's a chance that venture returns are depressed in this vintage? Yes, that's greater than 10% chance. But we also have this phenomena where the winners are arriving at a scale that no one's ever seen before. So in a power law industry, just because you make goofy decisions against, you know, 49 companies in your portfolio and you systematically overpay if you get one of them right and it ends up being 25 of an anthropic, you're still good and your LPs are still good, even more importantly. So I can't even say with great certainty how venture returns look like in this vintage.

[56:01](https://realalpha.app/v/K2OC5yKrbaM/?t=3361) **Matthew Harris:** I do know from our perspective, we try to keep pacing even. So we invest the same amount 500,000,000 a year.

[56:07](https://realalpha.app/v/K2OC5yKrbaM/?t=3367) I don't think this is 2021, but I do think you never know. And you do generally know that when you wanna do more, everyone else is having the same animal. It's a Stanley Drunken Miller, which is nothing looks as cheap as after it's gone up 40%. Right. It's exactly paradox. It's a huge paradox. And so... And the opposite is true too. Something goes down 80% even if there's fundamental value, it looks expensive and you want to sell. Right. So those human factors act on us and all we can do is recognize them. And the conclusion we've come to is you just got to pace. Just give up some of the potential upside for staying in the game.

[56:42](https://realalpha.app/v/K2OC5yKrbaM/?t=3402) Yeah. And avoiding some of the massive... The other the human factors of of investing a ton in a given ventures that ends up being a bad one is really significant. Like, there are firms out there who tripled their pace between '20... The industry tripled its pace between 2020 and 2021. So that just means there's like legions of partners sitting on boards that they massively overpaid for, and they're probably terrible SaaS companies that aren't gonna be worth anything. Like, the human factors of that actually go well beyond the returns of the vintage. They can actually cripple. It's the other side of the survivorship bias.

[57:17](https://realalpha.app/v/K2OC5yKrbaM/?t=3437) Yeah. You only hear from the survivors. Well, how do you know you weren't lucky? You don't think about the the people that did not survive. Yeah. Exact. So I think, anyway, this is what we've resolved to as a kind of a behavioral, you know, thing that we can put in place to discipline ourselves. And it's... It is challenging. Going from the existential dread of not knowing where AI is going to the existential dread, you have, I think, six kids 18 or younger. True. So you have to deal with this this topic all the time. I'm having my first first kid on the way. What skill set survives in terms of...

[57:55](https://realalpha.app/v/K2OC5yKrbaM/?t=3475) In a post AI world for somebody that's growing up today? The one thing I see in my kids, because I have 18, 15, 13, nine, nine, seven. So I've got kids that are doing stuff. I certainly have kids who have this sort of ineffable quality that people in the valley refer to as agency. And just because it's commonly said doesn't mean it's not important. I find it to be actually the defining. Do you get credit for that or were they just born that way? That's more like their mom, I would say. Oh, in terms of our parenting. I think Montessori school is a good idea. So at a Montessori school, they show up and they can choose whatever they want to work with.

[58:36](https://realalpha.app/v/K2OC5yKrbaM/?t=3516) There's no program for the day. That's the characteristic of Montessori preschools. And I think that our kids all directionally then have this instinct upon a new situation of like, I get to choose. And you can see it's varying degrees. But I would say, you know, when you have more and more kids, every every belief you have gets shaken out of you over time because you're like, oh, this is true for one kid. Oh, it's not true for the second kid or whatever. And there's very few things that survive all six kids because sort of one by one, all of your priors get knocked out by the kid number four.

[59:12](https://realalpha.app/v/K2OC5yKrbaM/?t=3552) Turns out that rule isn't true, etcetera. And the one in terms of parenting? Correct. In terms of human nature in general, but certainly in terms of parenting. The two rules that actually survive. One is Montessori preschool is just good. Just do that. They're actually very, very consistent. Wherever you live, you can find one. Send your kids there. Two is don't answer their questions. Kids ask questions all the time. And as a good parent, you want to be like, oh, my my daughter wants to know why the sky is blue or whatever. It is not useful for you to answer their questions. They will never remember the answers to, you know, just say, why do you think what do you think?

[59:50](https://realalpha.app/v/K2OC5yKrbaM/?t=3590) Why do you think it's blue? And so our kids find it incredibly annoying, particularly as they get older. They just want to know, like, what subway stops to get off. But that I would just I just commend that to you as a new dad. And congratulations, by the way. Just you have no obligation to answer their questions and you're not doing them any favors. And that builds in them these processes to answer their own questions. Correct. And it's all about this agency. I mean, this is what the world demands right now is when faced with a blank screen. What are you going to do? And you have the ability of everything at your fingertips to do anything you want, anything you don't know how to do, you just ask it and it tells you.

[1:00:30](https://realalpha.app/v/K2OC5yKrbaM/?t=3630) So if you have agency, you start a business at age 12 or do whatever you want, write a college level thesis at age 12. But most people don't have that kind of agency. I don't hold myself out as a great parent. I certainly, though, have a pretty big sample size, and those are the only two things. You're an experienced parent, at least. Those are the only two principles that have survived the ringer of six kids. I'm sure your kids have grown up in quite a privileged background. Is there a way to account for this in terms of how you parent them? It used to be that choosing the venture capital asset class was a good way to keep them from being too privileged.

[1:01:08](https://realalpha.app/v/K2OC5yKrbaM/?t=3668) I mean, you know, I ran my own venture firm from, the late 90s through to 2011, 2012. And those pretty lean years. So I think at least the older kids understand what it's like to grow up in New York and feel like dad doesn't run a hedge fund or mom doesn't run a hedge fund. I think it's a little harder with venture capital having been a more productive. But at the same time, it's kind of an easy question. You remember how you were brought up. I remember a public school kid with two professor parents. And and I think if you think about it, you can remember what lessons you learned from whatever degree of denial and austerity versus whatever degree of privilege you were fortunate enough to have.

[1:01:53](https://realalpha.app/v/K2OC5yKrbaM/?t=3713) And it doesn't mean that you have to pretend you don't have means. It's the conclusion we've come to. It just means that you have to, you know, build a life that makes sense and could reasonably be available to them in the future, even as they stand on their own. And so that's set the standard too high. Yeah, you can't. In our view, you can't indulge in private aviation. That's like again, not that we would anyway likely, but that's surely something that once a kid experiences and feels is normal, that without it for the rest of their life, it's just going to be it's going to be weird. And there are a lot of things in that category around how you spend vacations and what kind of home or homes you have, etcetera.

[1:02:40](https://realalpha.app/v/K2OC5yKrbaM/?t=3760) So I actually think this is one where common sense is like pretty available and people just need to make, you know, good Well, on that note, this has been an absolute master class. Thanks so much for stopping by. That's great to be with you today. Congrats on getting to the end of my conversation with Matt. If you enjoy that conversation, you're gonna love the conversation right here.