VALUE: After Hours (S08 E39): Value Investing in Japan: Governance, Buybacks & Microcaps | Dan Rasmussen
Sunday, November 22, 2026 · john hopkins

During their recent episode, Taylor, Carlisle, and Dan Rasmussen discussed:
AI Stock Market Bubble: Why Dan Rasmussen Warns of a Major Correction
AI Infrastructure Spending: Are Nvidia and the Hyperscalers Overinvesting?
South Korea Value Investing: Cheap Stocks, Holding Companies and Governance Reform
Japan Value Investing: How Share Buybacks and Corporate Governance Are Changing the Market
Stock Market Crash Indicators: What High-Yield Credit Spreads Reveal About Risk
Nvidia's AI Financing Risks: Could Data Centre Overinvestment Trigger a Downturn?
TRANSCRIPT
[Tobias]
This is Value After Hours. I'm Tobias Carlisle. Jake is out today.
I'm joined by Dan Rasmussen. How are you, Dan? Good to see you.
[Dan]
Good, good to see you too, Toby.
[Tobias]
Oops, I've got my speed bet coming through. Wild times in the markets, Dan.
[Dan]
I know, maybe the AI trade is finally coming undone and your screen background of market crashes is going to become a reality.
[Tobias]
Well, I think that valuation has been pointing us there a little while and there's certainly been a lot of speculative fervour in the market. How do you see it from your sort of more traditional value perspective?
[Dan]
You know, the companies that value investors tend to like and tend to be invested in are sort of real world businesses, right? So I'm invested in a lot of manufacturing companies and consumer companies and businesses that are doing sort of real things every day. And when you think about this AI boom, at some point those normal companies have to become customers of the AI companies for the AI boom to really work out to the tune of something like 600 billion or more of revenue per year, which probably looks like almost a quarter of SG&A spend that most large companies needs to get turned over to AI for this trade to be a really profitable one.
And what I'm thinking about really is, is there any evidence that we're going to get there? Is AI in fact generating the productivity improvements, replacing the workers, doing the jobs that would fulfil that forecast? And right now I don't see any evidence that it is.
I see lots of commentary from tech people and bullish AI folks on how AI is going to take over the world if it doesn't kill us first. But I'm not seeing any evidence in the real world of significant enough profit improvements, significant enough cost savings, significant enough productivity gains to make the numbers work. And that's my big question.
On the one hand, you could say, well, AI is one of the fastest growing, most exciting, most innovative novel technologies. And so you're just, you're missing out on how rapidly things are going to change in the world. But I'd say, I think that the AI people are massively underestimating how much inertia there is in the world.
Even if the technology worked perfectly, how long it would take companies to actually find a use for it, implement it, feel comfortable running it in production, feel comfortable firing a quarter of their workforce to replace them with AI. I think that the timeline is going to be much, much slower. And the depreciation math on those chip investments means that delay works against them.
I mean, it's sort of the opposite of value investing. In value investing, you buy into something cheap and you wait for things, the positive surprises. In the AI trade, you need something good to happen soon.
Otherwise, it's a ticking time bomb of depreciation curves that works against you. And that's largely my view on where we are in this AI thing. I'm quite worried about it.
And I think that it could cause a major market correction if some of these more dire forecasts are correct.
[Tobias]
You're worried about it in the sense that the prices that are being paid for the equities is too high. Not in the sense that the AI itself is going to take away jobs or enslave us or something like that.
AI Stock Market Bubble: Why Dan Rasmussen Warns of a Major Correction
[Dan]
I sort of think that's sort of like a psyop. It's like, let's warn people that the AI- That's anthropic psyop.
[Tobias]
That's their propaganda.
[Dan]
It's going to kill you all. And you're like, well, if it could kill us all, it's probably really profitable. But no, I think it's about the equity valuations.
Because with growth stocks, you need a lot of growth and you need it fast. And I just don't see that materialising as fast as people expect it to. Because again, I think the real world use cases of AI are not yet working that well.
[Tobias]
Do you think that- It's been an unusual sort of- This is a topic that I've kind of revisited ad nauseum on the podcast. But it's been an unusual decade that has largely defied what the preceding hundred years did. And one of the clearest examples of that that I always give is the market capitalisation-weighted SPY that everybody's more familiar with, float-adjusted market capitalisation-weighted, relative to the equal weight version of it, which favours smaller companies.
So it's like a proxy for small versus large. Or you could think of OEF, which is the largest 100 versus the 500. So in this instance, the 500 is the small and the 100 is the large.
So you're using the same 500, but in two different ways. In one way, it's the large. In one way, it's the small.
And then if you track those two series against each other, there are these very clear peaks around technological booms, 2000 being a very clear peak. And then we're in this current one now. And there was this very clear peak, I thought, in 2020, when we were going through that meme stock mania, that sort of reverse course.
And then that now, from the lofty peaks that we've ascended to now, that's sort of a little foothill, really. And I just wonder whether it's that more than anything else that has sort of driven, everything else downstream of that has sort of suffered as a result, equal weight, values, small, cyclicals. How do you see that?
[Dan]
Right. I mean, it makes me think of that old quote about the wings of the butterfly that flap and change history because the wind blows, or World War I starting because somebody assassinated Archduke Ferdinand.
[Tobias]
Right.
[Dan]
And you think about how these tiny human things just affect the course of humanity. And you wonder if Jensen Huang's love affair with a Chinese spy might not be the thing that reversed a half decade of pain for the small cap sector value premium. But I think that's exactly the right thing to think about.
And it's happened all up the cap range. I mean, if you look at small versus large, large versus very large, micro versus small, the smaller the stock, the worse it's done. And the larger the stock, the better it's done.
And it's the sort of most brilliant thing you could have done 15, 20 years ago is just say, hey, I want to overweight the largest stocks. But no historical evidence suggested that that was a good idea.
[Tobias]
Right.
[Dan]
And so I think one of the reasons that those things were foothills relative to what's going on with the AI thing is in some sense like humans are pattern recognising, like we update based on what we've experienced. And I think we look at, you know, you look at the SAS boom and, you know, we look at these previous tech things. And basically the growth tech people have just been right.
They've just been right. Like every new technology they've been right about, like it really did become super profitable and really big. And it did turn out to have a huge TAM.
And it did turn out to be very scalable at high margins. And so when they come in and say, trust us on AI, you're sort of like, well, you were right about the last three. I mean, everybody except Mark Zuckerberg.
Like, that's the funny, he's like the only exception where like, he was like the metaverse is next. And then like, yeah, it just didn't work. But everything else they've kind of been right about.
And I think that's why they have so much credibility now to go burn hundreds of billions of dollars on data centres and build this new massive AI infrastructure. But again, you know, I don't think that just because recent experience suggests something, I'm just not sure that the right forecast is that this is going to work the way that those previous things work. And certainly the capital intensity of it is much different.
Like SAS didn't require you to spend hundreds of billions of dollars on compute and inference. It was much more asset light. And now, you know, the big tech companies are more capital intensive than industrials.
And I think that's what's changed too, where there's sort of a law of CapEx, where high amounts of CapEx spend produce low future returns because people always overbuild and overestimate demand. And I think that cyclical CapEx cycle math is what's going to really dam these companies.
[Tobias]
One of the things that I've been watching is the relationship between the hyperscalers and the semis. I think that I've got a question in the chat here, which is, you know, Micron's on three times earnings, that's expensive. And you could say the same thing of NVIDIA, SK Hynix, all of these companies, which have historically been quite cyclical.
And so you've needed to take care when they're over-earning. And that's why they tend to be, when they're trading on low multiples, that's the market sort of treating them as if they're closer to the top of their business cycle and higher multiples when they're closer to the bottom of their business cycle. And they tend to stock perhaps a better bet when they are more optically expensive on a multiple basis than when they look cheaper on a multiple basis.
But I think that that's quite different to the hyperscalers, which that's a much more nuanced bet there that I'm not entirely sure how that plays out. I don't know that anybody is, but it's sort of this unprecedented level of investment where there's this absolute gold rush going on. And it's the hyperscaler spending and the semis sort of on the other side of that, collecting all of that money, which is, and the fact that it doesn't come out of free cashflow and instead it has to come out of equity and debt raising makes me think that they really do need to get to the other side and sort of start demonstrating that revenue in order to keep the party going.
Otherwise they will follow that sort of classic, I don't know if it's the Gartner hype scale, hype, whatever it is, hype curve, where it typically gets a little bit ahead of itself. I think it's funny that, in some senses like Zuck's comments about the metaverse were a little bit like Gates. He was talking about broadband when we were sort of months away from the internet web 1.0 kind of really kicking off in earnest. So he just had completely missed that. In the same way that Zuck's sort of looking at the metaverse and seems to have completely missed AI. AI is only a very recent trade.
How do you pass that? Is that a fair characterisation? Do you think that the hyperscalers really have to prove the business case?
AI Infrastructure Spending: Are Nvidia and the Hyperscalers Overinvesting?
[Dan]
Yeah. I mean, I think that, one sort of observation is that as long as your top line is growing very rapidly, the market gives you a lot of room to run. Investors will hold faith with you as long as your revenue numbers are really accelerating and high.
And in that sense, we've seen no slowdown in AI revenue. The revenues are up and to the right at a very crazy, crazy pace. And that's what's sustaining the boom.
And so you've got to see a slowdown or a deceleration, a meaningful noticeable one in the numbers in order to have any of the bloom come off the rose on the AI trade. And so the question of like, is this some sort of S curve where we're on the steep part of the S and like, where is it flattened out? That's sort of, I think, what everyone's sort of looking for and trying to figure out.
But I sort of think it's funny because when you hear the analogies people use, they're like, well, you have to understand that we've created a new essential element to our economy. It's like semiconductors are the new oil or something, right? And you're like, well, have you ever tried investing in energy stocks?
Or like it's a gold rush. It's like, well, try investing in Canadian gold miners for a few years. These are capital-intensive businesses.
[Tobias]
Be careful what you wish for.
[Dan]
Exactly, like capital-intensive businesses. And like they have a logic of their own, which is that everybody does competition neglect where they think the ROI on my data centre is X. And they don't consider the fact that all their competitors are building data centres too and doing the exact same math.
And so they built five of them and they only should have built one. And then the price collapses because they've overbuilt. And then you can't give away a thing, right?
I mean, it's just this happens in every capex-intensive industry. It's cyclical sort of on its own nature. And yes, it's concealed when you're in this S curve.
But at the end of that S curve, I think it's going to be a very different story. And I think the Chinese angle makes this even more interesting because it's a commodity with a serious low-cost competitor aggressively trying to win and destroy the economics.
[Tobias]
I've just released a white paper because I'm sort of interested in the business case behind AI. And I think it's real. I think that if you use the tools more than the chat, if you use work or codecs and you build through those things, it's clear that there is something special there and they are doing something a little bit extraordinary, I think almost verging on magic.
And so I can see that that technology really is revolutionary. And I do think that the effects on the economy will be profound, probably in the same, I think it will be at least as impactful as the internet itself, which has been transformative over the last 25 years. But the argument, I think, is that this sort of tech is unkind to everything that has gone before.
And I've called this little white paper The Last Investment, which is a play on the last invention, which has been this idea in science fiction that at some point you build this super intelligence that replicates itself, that builds every other invention. It's the last tool that humanity ever builds. If that's the case, do we need money anymore?
If the AI sort of takes care of us. And at the same time, there's this very clear population collapse going on over the next generation where developed countries like South Korea is going to go from 40 million to 8 million and China's population is going to fall a long way as well. So we'll have this super powerful AI as there are fewer of us around.
And maybe that's enough to sort of look after us. So I went through 200 years of this sort of these tech revolutions. And I think that they all don't follow the same curve necessarily.
But there is that, I looked at, you know, Carlotta Perez has this idea that there's an installation phase which is often attended by a bubble. And the Dotcom 1.0 is the sort of classic example of that where the fibre optic cable was hardly lit as it was built out. It was sort of overbuilt.
And we've continued to use it to this day because the useful life is 25 years plus. Yet all of those companies went bankrupt in the infrastructure build-out phase. And so it was companies that came along later that were able to build on that infrastructure layer or the users of that ultimate sort of the builder.
So that's the Google meta, those sort of businesses that were the profitable ones. And I just wonder if, the questions I have this time, does the same thing happen so that it's not so much the infrastructure layer that is the layer that gets sort of destroyed because they build out so much of it. And I don't know what you would characterise as the infrastructure layer, but that might be the semis.
And then there's that builder layer and the user layer. I guess the question is for Anthropic's IPO, which is coming up, is Anthropic a builder layer or is it ultimately a user layer where you have Anthropic motors, you have Anthropic pharmaceuticals, Anthropic just does everything. Or does somebody else come along and build on top of that?
And then what are the implications for investors? Have you thought through those problems at all?
[Dan]
Well, it's interesting. I mean, I love her a lot of press. I mean, she should be on every conference stage.
You know, her book is so relevant. Her books are so relevant. But I think, you know, if I were to walk down the hall to what we call our analyst layer, where our analysts are all working and you walked by their screens, all of them would have cloud code or cursor open, right?
So all of them are working fundamentally out of AI as their day-to-day interface. Like they're not using Excel. They're not using a command line.
They're not using Word. It is sort of a remarkable change, right? Because a year ago that was not the case.
It's totally shifted, right? Where this sort of day-to-day productivity app that people are using is cloud or cursor or one of these other agents. And I think that's very interesting, right?
And in some sense, that seems most threatening to Microsoft Office, right?
[Tobias]
Yes.
[Dan]
Like, ooh, I'd be nervous, right? Because they're not really using Excel or PowerPoint anymore.
[Tobias]
Or Google Search.
[Dan]
Or Google Search, that's a great point. Or Word. I'm like, I don't know, right?
I do really worry about that sort of productivity software. But on the other hand, it's an interesting dynamic whereby, let's say you're building, you see it in these sort of venture questions. Like you're building something that sits on top of Anthropic, like an application layer on top of Anthropic.
But your investor is like, why wouldn't Anthropic just build that? You probably built that with AI. Couldn't AI just build that again?
And if you're plugging into them as the engine, why wouldn't they? So it seems hard for someone to build on top of, like build a, you know what I mean? I wonder about the competition with the customers and where that sort of ends or where that line, like does Anthropic just become the super company that does everything?
Or does it find a way to provide a service to its customers without disintermediating them? I don't know, it's fascinating. I mean, I think it's totally unpredictable because it's such a revolutionary new technology.
And that's why I think you have to sort of zoom out and say, let's not get too into the weeds on the individual aspects of technology because who would have predicted Google or who would have predicted that like one of the killer apps for the internet, the killer use case for the internet would have been Facebook, right? Really? Like you invented the internet and then like the number one thing that made it really valuable was that?
That was really valuable was like a camera app or like Uber. I mean, you would have sort of been shocked by what ended up being really useful about the new technology. I think we're still at that phase where no one really knows.
But some of the things that people thought it would be instantly good at like replacing call centre employees turned out it can't really do that well yet. Or some of the most advanced things that it's really good at, like self-driving, I'm just, I'm not, I'm hugely bullish on the technology. I love using it.
It's very valuable. I love talking to it. But do I think it really gets rolled out into everyday businesses at scale fast?
No.
[Tobias]
Dan, I've just had a request to turn up your mic if you could do that somehow.
[Dan]
I'm trying to turn up my mic. Let me try.
[Tobias]
While you're doing that, I've launched... Certainly a little bit louder, yeah. You've launched some new strategies.
Do you want to tell us about those strategies that you've recently launched?
South Korea Value Investing: Cheap Stocks, Holding Companies and Governance Reform
[Dan]
Well, I'm very focused now on Korea, which is a value investor's paradise and a really fun market to learn about because it's so weird. So I'd say the sort of headline of what I've learned about Korea, first of all, you go to Korea and it's so cheap. I mean, I thought Japan was cheap and I'm buying stuff there at like 10 times earning and 0.6 times book. And I go to Korea and it's like 8 times earning and 0.4 times book and I'm just like, holy smokes, this is really cheap. And then you get into sort of the why it's cheap and that's where it gets even more fun because the why it's cheap was something totally novel to me. It was a sort of governance theme which is spreading out across Asia and it's had such a huge impact in Japan.
And the sort of puzzle of Korea is that so much, so many of the problems in Korea, so much of the Korea discount is a result of family control. So Korea is a market that's still controlled by families and these families want to pass on these businesses to their heirs and they want to keep the economics in the families is they've created these very odd structures and the two odd structures that are most notable are first the holding company which there are lots of countries that have holding companies but in Korea there are tonnes of them. And so you'll have the credit ratings agency and then you'll have the hold co that owns the credit rating agency.
It owns say 30% of the shares in the credit ratings agency. Enough to control, it has a security stake and that credit ratings agency is going to pay money up to the hold co. The family is going to own all their money in the hold co and use that big stake to maintain control over the operating company.
And then they are going to take the capital up in the hold co and they might allocate it over to like a manufacturing business or a venture arm. And so the dividends that you're getting from the subsidiary, is going to be in the numbers of the hold co. And then the rest of the money from the hold co gets used for the family's own internal investments.
And so these hold cos often trade at a say 50% discount to the value of the underlying operating company. So you're buying something instead of at 8 times earnings and 0.4 times book you're owning it at like 5 times earnings and 0.3 or 0.2 times book because you're on the hold co Then the families make things even more interesting because they say well gee I'm still quite concerned about I only own 30% of this operating company I really want to control more of it what should I do? So they say well why don't we have the operating company issue Korean preferred shares.
And the preferred shares in Korea trade, they're just like common shares they have the same ownership they're not like bonds in any way they have two main differences one is that they don't have any control rights so you can't vote on the board you can't vote on sharehold it's no governance power and then second they have slightly higher dividends slightly higher and so they say get a little bit more dividend and let us keep control and these preferreds have now traded down in most cases to a 50% discount to the common to the top co at like a 40-50% discount so you can construct like a Korea it has to be small all the interesting stuff in small cap world as it always is but you can construct a Korea small cap portfolio that you're going into at like .3 times book and like 6 times earnings if you're willing to own holdcos and prefs and of course the operating businesses have all these weird structures and discounts and the reason that that's exciting right now is again there's corporate governance reform and so what's happening in Korea and if you've watched any Korean movies or Korean television you probably have seen this class tension class warfare story like squid games or Train to Busan there's this sense of it's sort of like the stock market and so what's happening in Korea is that there's a populist movement to essentially take economic control in the stock market away from the families and so all of these individuals they call themselves ants in Korea they vote and they want the government to pass regulations that will bind the hands of these families that will force them basically drive the price to book multiple and their holdcos up and close the discount stop doing some of these weird things like preferred share issuance so there's just this huge momentum that's coming from this populist movement that's getting stuff passed in the legislature that's introducing things like for example major reform that board members have a fiduciary duty to them so there are big big changes to the legal structure in Korea and for example people are saying well what does that mean for the pref shares right I mean does that mean that they have to be treated more fairly because now you have fiduciary duty to minorities so can the board vote on say doing a buyback of the common but not the pref or buying right like there are all these interesting things that I think are going to be playing out over the next few years that are going to drive hopefully not only a revaluation of the Korean market in general but a specific revaluation of these weird structures because the only purpose of these structures has been essentially to dodge taxes and maintain family control and that's exactly what these populist ants don't want
[Tobias]
and how do you feel about the prospects of that sort of becoming a more globalised sort of open arrangement
[Dan]
yeah well I think there's sort of two ways to look at that I'd say the first is again this corporate governance wave is sort of spreading across Asia it started in Japan it's been a huge success in Japan and so now you're seeing all these other Asian markets adopt these same principles and say like we want some of that action let's copy paste on what the Tokyo Stock Exchange has done and so there's that element to it and then there's also the element of Korea opening up right so Korea has in the past had a lot of restrictions on foreign capital so if you wanted to trade Korean stocks you had to get like a licence and a domestic broker it was a total nightmare since I think April of this year interactive brokers you can trade Korea like it so I think it was January of this year that they relaxed the rules on foreign brokers so it's like this stuff is happening and it's happening now and I think they really would like to see Korea included in the MSCI developed market index which it should be based on per capita GDP and it isn't because of these weird like weird rules like not allowing offshore trading of the won or these sort of licences for people to trade Korean stocks and so I think as they get rid of more and more of these weird rules and as they solve these governance things you're going to see more and more money flow into Korea I think you're starting to see it's just like Japan but a few years ago you're starting to see increased private equity activity private equity firms opening offices in Korea you're seeing a lot of interest in the Korean activists and then you're seeing foreign investors like me say wow like I can trade this now and it's actually pretty interesting and and it's not picked over right it's like a pretty inefficient market like who are my competitors there right like who else is trading this stuff it's funny we even like we've just been we've been doing a lot of research on Korea and we just sent an email to our data provider with about 70 data errors in their Korean data like we like we just pooled them all together and like sent them in mass over and I'm like the fact that we could find 70 data errors and they're like I mean like clearly we're among the first people to look at some of this data otherwise other people would have found this sooner and I think it's just it's to me that just tells me that I found an interesting opportunity that other people aren't playing in
[Tobias]
Speaking of Japan I had the opportunity to visit with your analyst in Japan as he made a visit to one of the companies and I Jake and I were together when we did that in April last year and I think that we were incredibly hot on the opportunity in Japan speaking to private equity and some of the activists and then we got to see when we visited with the company with the analyst why the problem why the opportunity existed let's say but I understand that that's that's opened up quite considerably in the sort of 18 months since can you talk a little bit about the experience
Japan Value Investing: How Share Buybacks and Corporate Governance Are Changing the Market
[Dan]
Yeah so Japan you know we talked about the discounts in Korea being related to these governance and family control problems in Japan the problems are totally different the problem is overcapitalised balance sheets so Japanese companies basically are extraordinarily conservative and so for 20 plus years every time they've earned net income their willingness to dividend that out and certainly their willingness to any buybacks has been very very low and instead they've been put money in the bank or they've bought these cross share holdings in other Japanese companies or they bought real estate and so they've built these sort of fortress balance sheets which were in their mind designed like oh if we ever have a bad time we've got two years of revenue in the bank and part of that is related to keeping full time employment right we want our employees to always be employed so we need to have a rainy day fund and the rainy day fund just grew and grew and grew and grew and you see these management teams who have no interest in or particular knowledge of our facility with the idea of like cost of capital or return on investment it's just like a foreign language to them and so you're trying to explain like well you know you're building a new factory and your stock trades like you know 10 times earnings and so you're going to put all this money in and then it's going to be valued at x and so the like return on investment on this is like 5% but your stock trades at .5 times book with a 5% dividend like you could buy your own stock it would be clearly a better economic decision than building this new factory like giving your return on capital and it just doesn't make any sense to them they're like well what if I bought this stock then what would happen to the cash I just like spent the cash and it would be gone and like why would that benefit anybody like you're like well the stock price would go up but wouldn't it just go right back down after we bought it you're like no no and you're sort of like walking through some of these basic maths of how sort of we've been taught to think about capital allocation and I think that that's why when you go to these company meetings you're sort of like well holy smokes like this doesn't make any sense to me what are these people smoking but the more time you spend with it I'd say the more I've learned that actually these things are changing there is much bigger openness to thinking about cost of capital thinking about what the right size of the balance sheet should be and I think that the sort of Japanese way is that this thing is going very slowly it's very conservative but it is happening like you look at buybacks in Japan boom like dividends in Japan boom like reductions in cross share holdings boom right 23 it starts and it's just like dramatic change but to get from here to where it needs to be it's going to take a few more years like we're still early in the Japan governance reform trade in the sense that like these companies are adopt they're dipping their toes in the water but they're not whole hog adopting jack welch's approach to management overnight
[Tobias]
yeah I thought it was eye opening but also that their views were entirely genuine and they were concerned about the customers and their suppliers it wasn't I didn't feel that they were pushing back for the sake of stubbornness or that's the way it's always been I felt it came from genuine concern and that was the thing that made me think it would be harder to shift than unwillingness to do it just while we're having a quick break I'll shout out to all the folks who are listening I'll shout out to folks who are listening I'll shout out to all the folks who are listening I'll shout folks who are listening I'll shout out to all the folks who are listening I'll shout out to all the folks who are listening I'll shout out to all the folks who are listening out to all the I'll shout out to all the folks who are listening I'll shout out to all the folks who are listening I'll shout out to all the folks who are listening I'll shout out to I think the case for taking this seriously is real.
In November, an evaluation group, METR, found AI helping one lab, Anthropic, improve its models at about twice its earlier pace. That is the first measurement of goods loop at work. Eight days later, the economists Alex Imus and Luis Garacano read the same evidence differently.
Research spending rose about 6.2 times to buy that doubling, so each dollar bought less progress, not more. It's too early to know which reading is right. That matters, because about 60% of a company's value comes from beyond the next 10 years, and those are years that we can't forecast.
If the builders don't keep the value, who does? William Nordhaus estimated that innovators in the post-war US economy kept about 2.2% of the value they created. The other 98% went to everyone else.
His favourite example is light. In 1800, a fixed amount of light cost more than five hours of work. By 1992, it cost less than half a second.
AI is doing to the price of thinking what electricity did to the price of light. The profits that survive don't vanish. They move.
Clayton Christensen showed that they gather wherever the industry is still held back. IBM built the defining personal computer of the 1980s, and profits went to Intel and Microsoft. Today the bottleneck is computing power.
NVIDIA earned a 75% gross margin in the quarter ending July. When chips stop being the constraint, those profits will move again to power and land, or to whoever owns the computer. Price decides whether you take any share of it.
US stocks returned 6.6% a year after inflation from 1900 to 2025, crashes included. But the Dow fell 89% after 1929 and took 25 years to get back. The NASDAQ fell 78% after 2000 and took 15.
The technology kept spreading the whole time. The investors who lost were the ones who paid too much. Today, the cyclically adjusted P on the S&P 500 is about 41.6. In 1999 at the peak, it was 44. No one knows which future arrives. Explosive growth, expensive progress, capital taking everything or redistribution. Each rewards different assets.
So I don't bet on a forecast. Own the economy, not the builders. Mind the price, follow the bottleneck, but check what its price already assumes.
Hedge the scenarios and avoid borrowing that forces a sale at the bottom. For most people, the biggest asset is the ability to earn a living, which is exactly what AI may reprice. Earning a share of the capital that might replace your work is the hedge.
So is the last invention the last investment? Probably not. A machine that can make every invention makes invention itself cheap.
History says the value of cheap things flows to the people who use them and to the owners of whatever stays scarce. Energy, land, chips and trust. No single bet is likely to capture it.
Good sentence carried a proviso. The machine will be the last invention, provided that the machine is docile enough to tell us how to keep it under control. That is the one risk no portfolio can hedge.
If we get it right, the biggest winners from the last invention won't be the people who built it. They'll be everyone who uses it. If you go to thelastinvestment.com, you'll see the white paper in its entirety. And you can sign up to a daily email that sends out some of the big moves and whether it benefits the builders, the infrastructure layer, the users, sort of to give us an idea where we're going. It's for allocators and for investors. Dan, how is the U.S. value situation looking from your perspective?
[Dan]
You know, I confess that I don't do much in U.S. value. I like playing in less competitive markets. I've always, I think I've told you this story before, but when I was little, my dad wanted us to all get into good colleges, and so he thought we should all play niche sports.
So my sister was a female ice hockey player, and my brother was a male figure skater. I was too fat to jump high enough on the ice to do enough turns to be that good at figure skating, so I got switched to rowing, which leaned me out. And then my little brother did rowing.
And so I've always thought, you know, I didn't want to do basketball, right? Like, too many good basketball players. Like, find the sport that nobody plays.
And so that's what's led me to Korea and Japan. And like, not that there aren't competitors there, but relative to the U.S. markets, like, the U.S. market's a thunderdome of the most, like, vicious, aggressive capitalists in all mankind competing. And so I've just not been as eager to put all my chips in that basket.
So when I look at the U.S. That said, when I look at what's cheap in the U.S., what, when we do invest there, you know, I've had a decent exposure. I've found some pretty interesting opportunities in health care. There's a lot of cheap health care stocks these days.
I'd say another area that I've found pretty fertile is energy. A lot of cheap U.S. energy. And that's sort of differentiated.
It's hard to find small cap energy stocks outside of the U.S. So if you want cheap energy exposure, resources exposure, you've got to kind of find it in the U.S. So those have been the sectors that have probably been most interesting to me in the U.S. But again, I've spent, you know, five, 10 percent of my time on thinking about U.S. value relative to spending a huge percentage on Japan and recently Korea.
[Tobias]
When you look at Korea, the two names that really stand out there are SK Hynix and Samsung. Do you have any views on those? Do you need to think about those if you're invested in Korea or can you sort of avoid them?
[Dan]
Well, those two stocks are half the Korea benchmark. So the minute you start trading Korea, you become very cognisant of how those two stocks are doing, because as they go, it goes the benchmark. And that's been a big contributor to Korea's sort of meteoric rise.
So I'm sort of my in my best case scenario is that those stocks become sort of an advertisement for Korean equities, where people say, oh, wow, I didn't know anything about Korean equities until I heard about, you know, I learned about them through Samsung and Hynix, and then I went a step deeper to find other things. You know, that's sort of the best case scenario. The negative scenario, worst case scenario is that when the when or if AI has a rough patch, those two stocks drag down the entire market, which I don't necessarily think is going to happen, because the other stuff is not not not not not as correlated to those two stocks.
And then the other, you know, potential path is that Samsung and Hynix continue. And, you know, that AI is the last investment. And, gee, the only thing worth owning was Hynix and Samsung, because those are the only two big things exposed to it.
But those interestingly, those those also have some pref and holdco structures around those two entities. So even if you like SK Hynix or you like Samsung, you know, you can find pretty interesting structures than the Korean market to bet on those at a discount.
[Tobias]
It's been an appreciation for perhaps the degeneracy of Korean investors on social media. Largely, I'm guessing speculation in those two names.
[Dan]
I'm disappointed when you read your list of names that there is no Seoul, Korea. Maybe it's just timing is wrong. I guess maybe for the hours, we got to get your Korean viewership up.
[Tobias]
It's early morning in Australia, so I'm not sure. I don't know what wrong language, perhaps. Yeah.
What do you do? Did you did you do you think that that brought a lot more people into the market? Is that part of the reason why it's opening up?
Is it a good thing for the Korean market?
[Dan]
I think so. I think it's like it's advertising, right? It's all of a sudden someone knows that there are stocks in Korea, you know, and I think that's a good thing.
You know, I'd be hard pressed to like think of what's like what's the marquee stock in India? Like, I don't know. Like, you know, I can't quite think of it.
But even Europe, perhaps. Yeah, exactly. That's the only one I can kind of think of.
But even if you're like, what about what are the best companies in Europe? Like I can kind of think of them. But in Korea, it's just so obvious or Taiwan.
It's so obvious. And that's both a good thing is that it's putting them on the map. And a bad thing is that you want people's attention to go beyond that.
[Tobias]
I think the U.S. is unique in sort of having enough of a spread through those sort of businesses. Even though the index here is incredibly concentrated in just a handful of names. Even that incredible concentration is still like five or ten names rather than one or two.
[Dan]
Yeah. And it's funny because in some sense, this is like it was designed that way. Right.
In Korea, the sort of miracle on the Han River that led to the country's economic growth. It was based on this idea of forming national champion companies. They basically said like let's focus our resources on a small set of national champion companies that we want to be globally competitive and that will win in the global market.
And we'll provide them cheap financing and help these families win. And like now you look at it and it's like, well, they have Samsung and Hynix and Hyundai. And shipbuilding and whatever.
They've got five, six of these major industries that were all sort of the direct product of that. So it's not a surprise. It was intentionally concentrated.
Whereas the U.S. and maybe the U.K. and Australia have this entrepreneurial spirit where they're trying to create a very broad array of companies that will, through innovation and creative destruction, grow to form a more diverse and rich economic material.
[Tobias]
Yeah. I think Australia is a little bit like Canada. It tends to be it's a third basic materials and it's a half finance, which is banks and insurers.
And I think Canada is the same. There's not a big consumer discretionary technology. Those sectors don't exist to the same extent.
I think the U.S. is almost unique. China, I think, is the only other market that really has that very broad distribution of sort of other interesting businesses. I don't know why.
Perhaps not a big enough population to support it.
[Dan]
Right. Right.
[Tobias]
There's a question here. Interactive Brokers only allows trading on about 1,800 of the 3,000 listed companies in Korea. What percentage of those fall into Verdad's universe?
[Dan]
That's interesting. We're able to trade everything we want to trade, so it hadn't really bothered me at all. I've got to look into what the ones that we can't trade are.
[Tobias]
You had a, I don't want to call it a crash indicator, but for want of a better word, a crash indicator. Can you give us an update? Perhaps what's in it and what are the readings look like now?
Stock Market Crash Indicators: What High-Yield Credit Spreads Reveal About Risk
[Dan]
Yeah, so we've long been very focused on the high yield spread, which is the spread between the cost of junk bonds and treasuries. And we view that as an excellent economic indicator. And when that spread blows out, it tells you that you're in a crisis.
And so what's been weird about the high yield spread is that it's just been extraordinarily low for a very, very long period of time. Like the market essentially has not been pricing in almost any default risk. And so you look right now and it's elevated.
So, you know, this summer, the spreads were like below three, which is historically below 300 basis points, which is historically very, very unusually low reading. Now they're back up to like 330 or so. The long term, you know, median is probably around 400.
And we would say like 600 is crisis level. And we haven't gotten to 600 since March of 2020. So it's been a long, a long period since we've had not even 22.
[Tobias]
We didn't we didn't see any of that.
[Dan]
No, right around Silicon Valley Bank. You got into the high 500s, but that was the closest we got. So it's just been this very long, placid period in the high yield spread.
And I sort of interpret that it's a little bit like a forest that hasn't had any forest fires. Right. You can see the kindling building up.
You know, you can see the ground cover and, you know, when the spark comes, there's going to be a lot of tinder. But but you don't know the high yield spread doesn't tell you what the spark when the spark is coming or what the spark will be. It just tells you how much risk is building up or how long things have been cheap.
And I think there's I think there's a lot of a lot of weak hands that have been supported by tight spreads. And the minute spreads blow out, it'll be painful.
[Tobias]
That's that Minsky idea that stability breeds instability.
[Dan]
Yeah.
[Tobias]
Is it the option adjusted spread? Is that is that what you're looking at for that? So that seems to spike.
That's big. That doesn't really move until right in the teeth of the drawdown. It's not really an early warning signal.
[Dan]
No, no, no. But it is it is it's self-reinforcing. So once spreads blow out, it's that itself becomes a driver of the crisis accelerating, which is what makes it valuable.
[Tobias]
I don't know if you watch, but that the B's and the C's have been fairly tight as well. I think everything's just been that there's been a lack of fear. But they do seem to have been blowing out over the last few months.
[Dan]
Yes. So one of the issues is that the high yield spread has become higher quality because all the bad debt has migrated into private credit. So if you look at like the triple C rated spread, it has blown out quite a bit.
So if you look at like this summer, that spread was that like, you know, not 900 in the 900s. It's now over 1300. So that's been a pretty big blowout in triple C's.
And you've seen that there was a credit hedge fund called Arini that blew up recently in Europe. And that was like the Europe triple C spreads blew out. And that's a blew out a lot of these distressed debt funds.
So it's been a very interesting time in the sort of distressed debt market because that lower end is finally giving out. And that's sort of an interesting sign. Like, will that be sort of the leading indicator of something to come?
Or is it just contained among the weakest issuers?
[Tobias]
I asked that question on Twitter, and the response was that basically I was just trying to stir something up. The suggestion was that it was only the hyperscalers borrowing. And that was what was causing the spreads.
I didn't really understand the response, honestly, because I thought it was a fair question to ask. Like when you see spreads blowing out in high yield, that's a classic warning sign of something coming. But the response was that, no, that this is perfectly acceptable.
Honestly, I didn't really understand the argument, but you don't have any insight to add there?
[Dan]
I'm on the side of when you see spreads blowing out, be nervous. Be very nervous. But again, we're still at 330 for the high yield spread.
Like we're still in the market not seeing a material amount of risk right now.
[Tobias]
But it's not really forward looking particularly. It sort of seems to it'll react after that. Or in concert with the stock market, it'll move with the market.
So I think it's a good indicator in the sense that it does tell you that the drawdown in the market, there's some fundamental reason for the drawdown in the market rather than just the yearly. I mean, that must turn up somewhere. We've seen all of the near clouds and there's a lot of, it's sort of hidden leverage there.
It's a little bit like the WeWorks where they're leasing on one side and then sub-leasing on the other side, which is a great business because it's incredibly leveraged. So it's on the way up. It grows really, really quickly.
And on the way down. It cuts both ways. Do you have any view?
Because I think that NVIDIA has sort of been flattish for quite a long period of time and now down quite a lot. And there was an Australian near cloud that tried to go public, Firmus, and couldn't get across the line to this. There does seem to be, despite the fact that there is this absolute tidal wave of money going to this sector, we're still seeing sort of trouble already at this point.
Does that give you any pause?
[Dan]
Yeah. I mean, I think that would all be sort of consistent with the same story, that at the lower credit quality end of the spectrum, there are a lot of jitters right now. And we don't know yet if that spreads to the broader high yield market.
And some of that's perhaps tied to just the sheer volume of issuance. But a lot of things, you know, Oracle's trading very wide. And I think that's an A-rated bond that's trading at like 9% on Oracle bonds.
Right. That's a little scary. Right.
It's like, I mean, and they're a big funder of this, of OpenAI. And so we're actually doing a deep dive on Oracle internally to try to understand what Oracle's bonds indicate about the fate and health of the AI economy. Because it's a good case study.
[Tobias]
What do you make of the circular financing that NVIDIA seems to engage in?
Nvidia's AI Financing Risks: Could Data Centre Overinvestment Trigger a Downturn?
[Dan]
I mean, all this stuff, again, it's just like you described, good on the way up, terrible on the way down. Right. Like, all well and good.
But, gee, if the market takes a turn for the, like, that stuff is going to blow up in their faces. It seems like a very high risk decision. And, you know, you do wonder when they elevated the CFO to be the head of NVIDIA when Jensen stepped, you know, Jensen is out, the CFO gets promoted.
You're like, maybe they know something we don't know about the risk in some of these structures, you know?
[Tobias]
Yeah. The risk is in the finance rather than the tech. Yeah.
I don't know. I think it's an incredibly difficult problem to pass. I think if you look at history, it's a classic gold rush overbill in the short term.
Just overestimate the demand in the short term. There's so much inertia in the economy. It takes so long for everybody to adjust and to fully start using the capacity that's available.
And then on the other hand, I see reports of, you know, the chips, the older chips, they're being rented out at very high rates because there's just such a scarcity of these chips. And I don't know how quickly you can turn that machine on to sort of print more. But it does feel like that's a bottleneck that goes away at some point.
[Dan]
It's got to. It's got to go with capacity overbill, right? And, you know, when the demand stops, it'll stop all at once because everyone's doing the same thing.
[Tobias]
What about the political risk? It seems that I looked at the – there were 41 propositions against data centres in the last election, and 40 of those were supported by folks in favour of restraint on data centre building. Do you think it's a political risk as much as a sort of financing build-out risk?
[Dan]
Yeah, it's interesting. I mean, I saw there was some graphic that showed, like, the number – the two topics that have dominated, like, city council hearings across America in the past year. And there are two of them.
It's like over – I don't know if it's 60 percent of the conversation are those two issues, pickleball courts and data centres.
[Tobias]
Build one on top of the other.
[Dan]
Exactly. But I thought that was just so funny. It's like, you know, like for all the various political issues that, like, we debate in this country, the ones that people actually care about in their own neighbourhoods are pickleball courts and data centres.
But people really don't seem to like the data centres. I mean, to me, I don't totally get it. It just seems like another big building that goes up, so who cares what's in it.
[Tobias]
I mean, compared to the other kinds of factories that you could have built near you, it's reasonably inoffensive, just an air-conditioned room with some computers in it. The concern that I think is legitimate is folks say that the fans from the air conditioners emit this low-level hum that you can hear from a little way around. So that would mean that it is a local council kind of issue.
But every factory produces noise. Surely that's resolvable by sticking them in an industrial area.
[Dan]
Right. Yeah, one would think. I don't get it.
But it seems to be a big issue for people. So I wish I understood it better. I've got to try to walk in their shoes.
But I do think the political issue is real, and it seems to have motivated people in the midterm elections.
[Tobias]
So you don't subscribe to that view that Anthropic is winner-take-all, and if you miss out on the IPO, you'll be part of the underclass?
[Dan]
No, I'll take the under on that bet. But it is funny. I mean, I have a lot of friends in Silicon Valley, and you hear them talk, and they've all become like prophets of doom.
I mean, it's like, holy smokes. I mean, they really think they're enlightened and see the future in a very clear way. And the key to understanding the future is understanding the future of this AI technology.
And I'm like, I don't know. Here in Boston, I'm not seeing the same thing. But, you know, they've been right about a lot.
So we'll see.
[Tobias]
I mean, it seems to spell doom for software as a service.
[Dan]
Yes.
[Tobias]
I don't know if you saw that. There was a tweet doing the rounds yesterday that DocuSign had quoted a 15-person firm like $8,700 a year so they could collect signatures. And so the guys reverse engineered it.
And he calls it SignUDoc. And it's not perfect. It doesn't have some of the security features that DocuSign has.
But there is some like prisoner's dilemma where you're going to escape. You're going to confess at some price level. And it seems like that's a pretty big number.
So I do wonder if it's software that sort of feels the pain first. And so that's probably why San Francisco is so concerned.
[Dan]
Right. And it's what AI is best at right now. Right.
I mean, that and maybe self-driving. Right. Those are the two most tip of the spear tools.
Right. So they're seeing the future in their view. They're riding Waymos around San Francisco and their other agents do the coding.
[Tobias]
I think that folks who are coders seem to be the ones who are most at risk. Because if you're sort of entrepreneurial or business-minded, you don't need a technical co-founder anymore. Because Claude, I think, can do some of that technical work for you.
[Dan]
Yes. Yeah. And, gee, I mean, here in our firm, it's doing a lot of work.
It's very impressive. Right. I mean, I'm sort of in love with it.
It's a massive productivity enhancer for us. But we haven't fired anyone because they've been replaced by AI. Right.
In fact, we're hiring more people because we're getting so much more done. So I just don't know where those sort of things land.
[Tobias]
Well, Dan, we're coming up on time. Very grateful for the amount of time that you've given to us today. If folks want to get in touch with you or follow along with what you're doing, how do they go about doing that?
[Dan]
Yeah. I'm on X at VerdadCap. And you can sign up for my weekly research, which goes out every Monday.
[Tobias]
Which is excellent.
[Dan]
So thank you for that. So that's the best way to reach me.
[Tobias]
And, folks, thelastinvestment.com has my white paper, if you can stand it. And there's also a little slot where you can sign up for an email where we send you a 47-second recap of the day's capital and AI news. So check it out.
Thanks, Dan. Thanks, folks. We'll see you next week.
Jake should be back then. Bye, everybody.
Medical Device Stocks: Why Medtech Could Be Undervalued
The 10 Cheapest Large-Cap Stocks This Week
Value Options: Nvidia Without Jensen: Does the Price Now Make Sense?