← Excess Returns 
Excess Returns
The Bubble Most Will Get Wrong | Aswath Damodaran on How He is Managing His Own Money in a World of AI
01:02:08
Loading summary
Transcript49 lines
- [00:00]
A
Investing is about buying something at the right price. It's not about buying great companies, it's not about buying superior management. It's about buying at the right price. And that animates almost every one of my choices. My prediction overall, I think AI is going to lower profit margins collectively across companies. There will be one or two players who make those investments. So the investment is going to pay off big time. They're going to, it's going to be a winner take all business and they're going to take hundreds of billions of dollars. Collectively though, if you take all of these companies that are spending tens of billions of dollars, they're going to step back and that collective investment is going to have a negative net present value. I'm wary about where we are at the market in terms of pricing, in terms of what we're building in. It seems while there's a pathway we can justify where we are today, there are also multiple pathways where things can go bad. So from that perspective, I want to move my money into things that are not correlated with equities.
- [01:02]
B
Watching Excess Returns, the channel that makes complex investing ideas simple enough to actually use, where we believe better questions lead to better decisions. I'm Matt Zigler. Kai Wu of Sparkline Capital is co hosting with me and our guest, the Dean Evaluation himself, Professor Azwath Debodaran Oswath. Welcome back to Access Returns.
- [01:21]
A
Thank you for having me.
- [01:23]
B
So last time we got pretty deep into the basics of how you actually build the stock portfolio. We also got some tips on booking Airbnbs in the home neighborhood for the family to come in, bouncing grandkids on knees and stuff like that too. I encourage people to check out the actual important life philosophy in that episode. But could you just as an overview, if somebody hasn't, hasn't heard it before, the stock portfolio philosophy, the number of stocks you hold, how you weight them. Initially, just some of the general philosophy.
- [01:55]
A
I think the first thing is there is, I mean, investing is very person specific. What works for me will not work for you. So I think the key thing to remember is rather than trying to replicate what I do, you got to look inward and say what makes me comfortable, what is right for me. So I'll talk about what's right for me. It might not be right for others. I love investing. I love, I like looking at companies. My life doesn't revolve around investing, but I do. And that's what leads me to pick stocks. So I don't start picking stocks because I think I can beat the market. In fact, I, I smile every time I see the title for your excess returns. Because we live in a world where that is perhaps the most difficult thing to deliver on an enduring basis. Right? It requires things that you can, that you need to bring to the table that are getting more and more difficult to do. So I start with the principle that I want to pick stocks. And then I want to build on that by saying I don't want to pick stocks in a way that I could do serious damage to myself, my lifestyle, and my family. The process of picking stocks, that risk aversion, that desire to preserve wealth, is what leads me to diversify. I know this is a contentious issue among investors. Should you concentrate and diversify? And there's all this talk about if you have conviction, you should concentrate. And you perhaps heard my thoughts on the word conviction, which I think is seriously overused. I don't have enough confidence, no matter how much work I put into an investment, to put my money in five stocks or four stocks because I think too much rides on it. So I spread my bets to how many stocks. For me, I need about 30 to 40 stocks because I tend to invest in younger companies as well as. If you invested only in mature companies, you could probably get away with 15 or 20 companies in your portfolio. But I invest in a wide array of companies, so I need more companies. So my portfolio is composed at any point in time between 30 and 45 stocks in them. And many of them are long standing. They stay in my portfolio for the long term. They're not in and out of my portfolio. My turnover is maybe three or four stocks a year that I sell and replace. So I have a fairly well spread out portfolio. And if 40 sounds like a large number, remember there are 45,000 plus publicly traded companies. 40 out of 45,000 is being incredibly picky. So I spread my bets because I don't want any individual company in my portfolio to have a day where I look at my portfolio and say, oh my God, what happened here? So on a day the market is down, my portfolio is going to be down with 40 companies. But I think that that is really what drives my spreading my bets. As to what goes into my portfolio. I don't have a particular frame of a company. It doesn't have to be a money making company. My rule of thumb is got to be an undervalued company. What does that even mean? I mean, that's where I think having a valuation framework that allows you to value young companies, money losing, losing companies, growing companies, is a good one because it opens the landscape for me that I can buy a young money losing company if it's priced right. And the key word is investing is about buying something at the right price. It's not about buying great companies, it's not about buying superior management, it's about buying at the right price. And that animates almost every one of my choices. So that's my overriding philosophy on investing. Buy undervalued companies, hold onto them as long as as they stay under or fairly valued. But be open to the possibility that something that was a great investment has run its course and it's time for you to leave and sell that position. So hopefully that encapsulates how I think about Investing.
- [05:51]
C
You mentioned 40,000 global stocks and it sounds like you're pretty bottoms up, like go anywhere. Are there certain sectors or countries or maybe even cap ranges that you try to avoid? The Buffett circle of competence idea. Do you use screens to kind of narrow down the universe? Because you're doing, you know, a lot of in depth work on each single name. 40,000 screens would be a lot of Excel tabs, right?
- [06:17]
A
I mean, I don't think of valuation in terms of Excel. I think in valuation in terms of stories. I will buy a great story if I think the story is not being priced right. So the story might show up in the numbers. I'm looking for great stories to invest in that are not recognized as great stories. It opens the door for me for a wide array of businesses. I've never believed that you truly have to completely. I mean, I don't know what a chip looks like, a computer chip looks like, but I bought Nvidia in 2018. If you make this about I need to know what a product is, you're going to end up in the Buffet trap, which is you're going to own a lot of Coca Cola and read a lot of the Washington Post. And the older you get, the more disconnected you're going to get from the market because you just don't relate to TikTok. So that does, you know. So to me, I don't have to completely understand a product, but I have to understand the business. I know the business a company is in. The fact that I don't quite understand its product will not stop me. The one group of companies I try to avoid are companies where politics drives the value more than business. Right? And the old days, this used to be government connected or connected connected companies in emerging markets. Increasingly the world is getting more and more government connected. I Mean, there was a point in time where you asked me, is it a US Company? Would you worry about it? They said no. Now, I think you're the subset of US Companies where you worry about government connections driving value. It's not because I don't like the companies, but because I'm not good at politics. It's part of the reason I avoided the green energy space for a long time, because so much of that space is driven by are you in the right place to take advantage of government subsidies and benefits? And that's a political assessment, and I'm not good at that. So from my perspective, there's a subset of companies I avoid. That subset includes companies where I have to forecast who win the next election to decide whether to buy a stock. And I don't want to be in that space.
- [08:22]
B
I want to ask you about your watch list process on two sides, because I've heard you talk about this before, and I think it's enormously useful. Both the watch list for the existing portfolio stuff that might exit at some point, for all the reasons you can discuss those, and likewise stuff to enter. Because the story aspect of the framework that you just brought up, I think is what makes your way of doing the watch list exercise so much different.
- [08:46]
A
Than others, I think. I mean, I'll give you an example. There's a company called Mercado Libre. It's a Latin American company I would love. I mean, people call it Latin America, is Amazon, but it's far more than Amazon because it's a financial service company that actually increasingly is displacing portfolio managers and banks in Latin America. I love the company and it's been on my watch list now for a year and a half. I have a valuation of it that I revisit. It got really close sometime during March or April of last year, and I wish I'd pulled the trigger when it was within 10%, but wishes and $2 will get me a cup of coffee. It's not going to do much. It's on my watch list, as is Palantir. I mean, I think much as Palantir gets this political connection issue, it is selling a product. It is a company that I think is one of the few that's actually converting the promise of AI into delivery in terms of products and services. There are two companies that I don't have in my portfolio right now that I would like to buy if the price got to the right point. And I keep a running tab of companies like those companies I see described by and said that Sounds interesting. And I take a look at the company and I look at the management, I look at the financials, I look at the business it's in, I look at the competition, I say, this is a very good company. But at the price today, I'm not that interested. But it's going to go on my watch list. It's how I've ended up with six of the FA7. Now five of the MAG7 stocks in my portfolio is because they went on a list at some point in time and it's, you know, at some point further down the road, the price dropped enough that I could jump in and buy those stocks. So I think that it is, if you, if you find a company that you find fascinating, that you love the management, love the company, love the story, even if it's overpriced today, don't just let it go, keep it in your on your radar, track it, follow it. And you can't go back to your original valuation. You got to revisit that valuation, saying one of these days, AI is incredible. It can teach you how to fry an egg and even write a poem pirate style, but it knows nothing about your work. Slackbot is different. It doesn't just know the facts, it knows your schedule. It can turn a brainstorm into a brief. And it doesn't need to be taught because Slackbot isn't just another AI. It's AI that knows your work as well as you do. Visit slack.com meetslackbot to learn more. New year, same extra value meals at McDonald's. So now get two snack wraps plus fries and a medium soft drink for.
- [11:35]
B
Just $8 for a limited time only, prices and participation may vary. Prices may be higher in Hawaii, Alaska and California. And for delivery.
- [11:43]
A
And I think that with Tesla, very in the first five years of its existence that I looked at it between 2013 and 2018, I said I like the company, I like the product, I like the way it's approaching the market, but I don't like the price. Then in 2019, after that episode with $410 funding secured and the stock price collapsed, it got to a point where for the first time I said, it's at the right price. And that's my opening is I want to take interesting great companies, but I want to get them at the right price. Which means I've got to track two things. What's the value, what's the price? And to me, that's the essence of investing is keeping your eyes on both sides of that divide.
- [12:29]
C
Yeah, you know Buying stocks is fun. Selling stocks is not as fun. So how do you think about, you know, a good company you own? You mentioned Tesla, Nvidia, these sorts of names. You know, when do you sell? Like how do you think about your cell discipline and rebalancing? As, as you know, your thesis plays out.
- [12:46]
A
I mean, as a value investor, I've never quite understood buy and forget philosophy that many old time value investors brought in. You buy a great company, just hold it for the rest of its life. Because when you're truly a value investor, you buy when something is undervalued. And the flip side of that is you stand if it's overvalued. So we want a margin of safety and you use the Seth Klarman version of it and you buy something because it's 25% undervalued, should you also be thinking about selling if not at 25% overvalued at 30 or 40%? There needs to be a trigger on both sides. Now, one of the things I do when I value companies is I've increasingly turned to using Monte Carlo simulations because historically valuations have been based on point estimates. Why? Because we had no choice. In the 1980s, you couldn't run simulations unless you'd access to a mainframe. So you made point estimates for revenue growth and margins, knowing fully well on each of these estimates that you could be wrong. Today you no longer have to wring your hands and say, I could be wrong. You could tell me how wrong you are in the form of a distribution. And I tried to do that. And it's not, it's not incredibly difficult word. It just means that you have to use the data more fully. And I run simulation. The advantage of simulations is not only do I have a median value, which is going to drive my decision, I have a distribution of value. And my distribution of value gives me a margin of safety on both sides. I could say, look, I'm going to buy the 30th percentile if you want a 20% margin of safety, but I'm also going to sell at the 70th percentile. But here I think one of the advantages of having this process is it allows you to take decisions that you find difficult to make out of your hands. Let me explain. The stocks that you fall in love with are the stocks that have done really well for you. And once they do well for you, it's really, really, really difficult, even when they get overvalued, to let them go. Right? You say, it's been so good to me. I don't want to let go. So we tend to hold losers too long and we also sometimes tend to hold winners too long. So sometimes you got to automate the processing. Look, you know, I'm going to buy it, but I'm also going to put in a limit set. If it hits this price within the next six months, you don't want to make it time just because obviously values can shift that way. If you are right and you are right too soon, and that sounds like a strange thing to say, but you're right. You bought something that is undervalued. My worry is sometimes that the correction happens too fast in the next two weeks, the next four weeks, and then it keeps going through. It hits your median value and keeps going through. It gets really difficult to revisit your decision because you're so busy patting yourself on the back. So I think we know and I don't think you're. I mean, if you look at books and investing, almost all of them focus on what you should buy. Look for the chapter on when you should sell. Most of these books don't have one. Right. You don't at least explicitly talk. So I think this is partly because we focus so much on buying the right things that we don't talk enough about when do those right things need to get sold for your portfolio to be. And you know, there's always going to be the, the, the, the hindsight issue, which is you can say if I'd only held on. No. So you sold because you were disciplined, but the stock kept going up. Right. And we said if only you'd held on, you'd have made a lot more money. You're right. But you can't selectively pick just this hindsight. You got to look at the collection of stocks that you sold and say, did I on average do the right thing? And I think on average I've ended up ahead by being disciplined when I sell as well.
- [16:34]
B
Let's unpack. Especially with regard to looking at the portfolio and looking at the opportunities and highlighter on the idea. If you don't have a sell discipline, where are you going to make room to buy new stuff? Unless you're adding new money to the portfolio. It's.
- [16:47]
A
Yeah. And that's part of the problem. Right. If you're an active money manager and money keeps coming in, you don't need a sell discipline in a momentum driven market. Just need to keep buying and buying and buying. I mean, let's fence it. You know, if you look at Bitcoin and microstrategy the reputation that allowed MicroStrategy to trade at a premium on the bitcoin. It's a bitcoin spec. Let's face it. At this point, MicroStrategy is not a company. It's a bitcoin spac. It's a bitcoin spac that trades at a. That trades at a premium on bitcoin. Because, you know, you might say, viewed as this genius timer, the only thing was he always bought, he never sold. But as long as you're buying something that on average is going up as much as bitcoin, you look like a genius. The reason the market is turned on him is because once things start to go down and you don't sell, people say, well, maybe you're not such a genius. Maybe you just latched onto something that went up a lot quickly. So I think that if you have enough money coming in and people keep sending you more, without sell discipline, a market that's going up, you can look good because you keep buying things and they all keep going up. But I think you're right. Once you hit this point where markets start to saturate, fund flows start to level off, then you start to separate, in the words of Buffett, those who are wearing their bathing suits from those who are not. Because as the tide goes out, you start to recognize who the pretenders in this game are and who the people are really playing the game are. So you're right. Without a sell discipline, that's when you start to separate the two. But in buoyant markets, often you can get lazy investors who look good just because they got on the ground floor of an elevator that's going up to the 25th floor.
- [18:46]
B
What looks like a bathing suit may not actually be 1 on the current portfolio, but also on the ideas. How do you track the life cycle and maybe explain what the life cycle or that staging framework is for the ages of companies? This metaphor, I think this is wonderfully useful. How do you track that on top of the price and the valuation?
- [19:09]
A
I mean, I think there are, There are financial metrics you can use. Revenue growth, operating margins. When companies are young and growing, revenue growth is booming, margins are increasing, but companies mature. One of the dimensions of maturity is revenue growth starts to level off, not because they're not trying, but because they've scaled up, they're much bigger, and margins start to kind of hit. Hit a plateau. You're not seeing the jump in margins you saw. This is, in a sense, the best place to be for investors often is because that's when cash can be returned. The company's looking good, they're making a lot of money. But just as human beings age, companies start to hit aging points where businesses start to lose their. And you start to see this when companies have to try much harder to get the same revenue growth. One of my, one of my signals that I need to take a closer look at a growth company is when they start to do a lot of acquisitions, right? I mean, let's face it, the history of acquisitions is not good for value creation. You go to acquisitions because internal growth is no longer delivering the oomph that it used to. So the very fact that you're looking outward makes me a little more wary about your growth because you're trying harder. The second thing you'll start to notice is your margins start to level off and decrease. Initially, the decrease is not going to be large because you have scale working in your favor. You take a company like Meta or Alphabet, they have a cash cow in advertising that delivers so much profit that they can screw up at the margin and the projects are not. But you don't even notice until years into the process. So I look for revenue growth. That's where you're starting to struggle to deliver. Double digit growth that you used to be able to do effortlessly and margins that are starting to stagnate or decrease. That to me is a sign that you're aging. And if you're aging, here's what I want to see in the management. Stop trying so hard. You don't have to be a growth company to be a great company. I'll tell you, one of the reasons I think Tim Cook should be put in the list of top managers of all time is you look at what he's been in charge of at Apple for the last decade plus. I mean, this is the greatest cash machine in history, driven by the iPhone, hundreds of billions of dollars of free cash flow coming in. And the fact that he hasn't done a large acquisition or gone crazy on AI investment, to me is a good sign. It's a sign that he says, look, we're a mature company. We deliver huge cash flow. You can be a great company and not be a growth company going out after new business and throwing tens of billions. That's a sign of recognizing your age and acting your age. I want to invest in companies that act their age. So if you're a middle aged company that keeps trying to squeeze yourself into tight jeans because that's what you used to wear when you were a teenager, I don't want to have anything to do with you because you're going to take my money and try to be young again. You're trying to do your facelifts and big acquisitions and nothing's going to work. Because like human beings, the aging process kicks in. I mean, you take, I mean, last year I wrote about Starbucks and Intel. Companies that at their peak were amazing companies, great companies. They both aged and they're both struggling with aging in different ways. And I think that when we talk about management quality, those are the companies that need the most disciplined managers. But unfortunately we don't put those managers on pedestals. I blame business schools for this. You know, we glorify empire builders, we glorify growth. I would love to have a movie made about somebody who made their company smaller because they'd reached a peak. Because that to me is a sign that you're really managing a company the way it should be managed. So to me, aging is part of the corporate process. And the way I separate great companies from good companies is great. Companies can extend their life cycle not artificially, but by doing things that keep them at that 8% growth for a longer period. They don't try too hard to get to 20% growth anymore because they know that that's going to cost them too much. So to me, the challenge with investing in mature companies is taking a look at management. Say, is this a management that's going to live within what it can deliver, or is it a management that's going to try too hard to go back to being a growth company? Because the history of trying to do that is not that good for investors.
- [24:03]
C
So yeah, you mentioned as well two companies that have been in the news recently. So Apple, right, they have been abstaining from the AI race and instead made a deal with Gemini and then Meta. Right. In the complete opposite fashion, it appears that Zuckerberg's firing his workforce in order to put all his resources behind this top level initiative, Meta Compute. Hundreds of gigawatts of data center capacity. So which approach is correct, I guess ultimately will depend on the way the AI Capex build out plays out. Right, so let's transition now to that, since it's kind of the driving force in markets. You know, what are your thoughts on the folks out there who are, you know, saying the B word that may be a bubble? Do you think it is? And, you know, how does that kind of inform your view not just on Meta and Apple, but just on the market more broadly?
- [24:53]
A
I think any major change in the way we live and work is going to create a bubble. Why? Because that's what human beings do. They overreach, they're over optimistic. That's what allows us to kind of extend what we can do. I mean, one of my favorite sayings in class is, do you want to live in a world run by actuaries? We'd still be in caves, right? Because if you assess the expected value and the probabilities, right. There's very little risk that's really worth taking. So I would expect AI to create a bubble because it is changing the way we live and work. I mean, look around you. My wife teaches 5th grade. She is affected by ChatGPT. This is not like the Metaverse, right. Which was this. I don't even know what that is and know how it's going to play out. But this is. So I think AI is going to create a bubble because people are going to overreach. And there are two ways of approaching this. One is the Apple way of outsourcing this AI investment phase. They're still going to use AI. You, you spend the money to develop it, we'll buy the product from you. The other is to make a bet on. I want to be one of the big winners in this AI space and invest a lot. And I'll make a prediction there will be one or two players who make those investments. So the investment is going to pay off big time. They're going to. It's going to be a winner take all business. They're going to take hundreds of billions of dollars. Collectively though, if you take all of these companies that are spending tens of billions of dollars, they're going to step back and that collective investment is going to have a negative net present value. The nature of bubbles, we overreach. I call this the big market delusion. When you have a big market and everybody's trying to get into big market, people overestimate their chances of succeeding, overvalue the businesses they're in and collectively, almost by definition, you're going to get. We could call it a bubble, but there's overvaluation that's going to get corrected. I mean, I know people like to draw analogies to the dot com bubble and there are similarities. Thing that changed the way we lived it created a bubble of its own in terms of companies. But if you look at the companies from this, there were some companies that came out of that bubble with incredible valuations. Amazon is a classic example. You look at the PC bubble from the 1980s, right? Lots of companies that basically crashed and burned. But Microsoft was the big winner for the PC bubble, not IBM, not Dell, not Compaq. So two things from that. One is it's difficult to predict who the big winner from a bubble is going to be early on. But if you are an investor who can do it, you're going to make a huge amount of money. And that's what draws investors into this space, is they think they can pick the winners in this space by investing in one of these companies. I don't think it's a high odds bet for an investor to make to know you know, because you don't even know who the winner is going to be. The winner might be a company you don't even see on the list right now. But I think it's too early to anoint Nvidia and Microsoft and Meta as the winners, as some people are. They know Nvidia benefits from the other companies wanting to spend, but Microsoft and Meta is still working on trying to get something that'll make the money. So if you feel you can call the winner in this AI race, go bet on Meta or Alphabet or Microsoft or whoever you think is investing them. If you don't feel comfortable doing that, you can still invest in AI, but invest on the periphery. Invest away from the companies that are spending tens of billions of dollars.
- [28:33]
B
One of the interesting developments here, especially inside of this AI buildout and we might need a better bathing suit metaphor for this, this might be going from speedo to onesie. What I'm talking about is the asset light to capital intensity of these businesses. Kai actually wrote a really cool piece on this. We'll share this with you. But that's the core thesis. He was taking companies like Meta and the other mag7 today and comparing them in capital intensity to utility businesses. And he even threw at and T in there at the height of the dot com bubble. That shift from asset light to more capex intensive. What do you make of that?
- [29:13]
A
These are infrastructure businesses, right? And if you look at the history of infrastructure investing, you go back 120 years at&t building the first phone lines. The reason they were able to invest so much in infrastructure and borrow money to do it was they were building into a monopoly. So the history of infrastructure is you invest a lot in infrastructure, you have years and years of negative cash flows. But then once things happen, because your pricing power, either as a monopoly or as a duopoly, you can charge the prices to cover your infrastructure. That's your turn to business. The reason it's a regulated monopoly and the way governments Kind of restrict that pricing power is to make sure you earn your cost of capital. So it's not a great value adding business by definition because of the way it's structured. You get to the end of the 20th century, you saw the limitations of that when people started to build with the cable business, same infrastructure investment. The problem was they thought they were building into a monopoly. But technology operators, wildcard, because by the time they built the cable business, the business had shifted on them, they'd lost the monopoly power. So if you look at cable companies that built the infrastructure, they were not able to get their money back, at least collectively, because the business shifted. With AI, you're putting this on steroids. You're building infrastructure into what's definitely not going to be a monopoly, right? I mean, there might be a winner take all business at the end, but you don't know what structure it'll take. You don't even know what products you're going to sell. It's a remarkable bet on an amorphous business, right? That's what makes it different from historical infrastructure. Luckily, at least in the collective sense, the debt used to build the AI infrastructure is relatively small. I know there are some companies in this space that are using large amounts of debt, especially the smaller AI architecture companies. But look at the big tech companies. The reason they're able to spend this without borrowing money is because existing businesses are cash cops. I mean, let's face it, Meta's, you know, the Facebook advertising is what allows Meta to spend what they do. So from that perspective, if this crashes for a company, and it will crash for quite a few, the good news is the people who will bear the brunt of that cost will be the people within the company the shareholders employed, saying, why is that good news? Because the one thing you worry about with infrastructure crashes is this socialization of costs where you take costs and everybody else has to step in now to bear that cost. So the one concern I would have in this space is those AI infrastructure companies that are borrowing money, I think are making a bad choice from a corporate finance perspective. You don't borrow money to build infrastructure into a risky business. And those are companies that I worry about because if they go down and many of them are funded by private debt, which is another unregulated, unrestricted space, you have the potential of creating societal costs which then spread across other businesses. Now, the essence of risk taking is it's great if the risk takers are the ones who end up bearing the brunt of the downside. But risk Taking can be dangerous when you start to share those losses with everybody else who's not really involved in the upside. And I worry about at least a segment of the AI architecture business which is more debt funded and what the consequences will be when some of those businesses don't get the revenues and operating income to cover their infrastructure.
- [33:04]
C
Do you think that there's a sort of entanglement risk here too? Even for like the, let's call it, let's say Google or Microsoft. Right. They are cash flow machines, yet they are entangled with OpenAI, Oracle, Core, Weave, a lot of these companies with shakier finances. So yes, there only is a subset of companies who are explicitly debt funded. But Also with the SPV's, meta and you might mention private credit, is there a risk that one link in this chain goes down and infects and contaminates the balance sheets of the others?
- [33:39]
A
Yeah, and I think that's one of the problems with something that we haven't historically seen in the us this cross ownership you're seeing across the AI entities is something that is unusual for the us. The US historically has not had much cross ownership across companies for a variety of reasons. I mean, this is more of a Asian phenomenon where you have family group companies that invest across. By creating cross holdings, you're in a sense linking yourself with the successes of that cross owned entity, but also the potential failures. So that is a potentially worrisome development that you're creating this circle of companies all connected to each other and that one of them makes a big enough mistake. And you saw some of this play out with the Oracle right off of Hunt. I mean, you can make a big enough mistake as a single entity to start dragging other players into the game. So I think those companies that are in these cross ownership structures should take a closer look at them because here again, Apple's been more wary about doing this. And I think that's the right approach to take here, is don't be so quick to jump and buy 20% of another company, no matter how critical you think it is in the AI space, I think part of it is driven by this fear at least, that if you don't create this closed space, an outsider is going to come in from China, from somebody else, enter the space. And that fear was put in their minds by Deep Seek. And what it wrought in those few weeks that it rolled markets is that can't happen, that all of this money that AI is spending building an infrastructure could be blown up by an outside technology that delivers much of What AI is promising with a fraction of the investment, that you don't need these extraordinarily expensive chips and these huge data centers to deliver 90% of the AI products and services. And that's, I think, a potential worry for that entire space is that disruption could come from outside.
- [35:44]
B
When we think about AI in valuations, and there's so many different ways to approach it, but maybe at the top level first, like at the index level, do you look at it as a way of short term increasing profit margins, long term hurting it? Do you look at it as pushing valuations?
- [36:01]
A
I'll give you my prediction. Overall, I think AI is going to lower profit margins collectively across companies. And here's my reasoning for that. They're going to be companies, obviously, that make money from selling AI products and services. The key word is they're selling it. It becomes a cost item to companies. So let's suppose you develop this great AI product for grocery stores. You go to Kroger's and you say, if you get this product product you're going to get customers going to the right aisles get more revenue. So Kroger is going to invest in it because they think they're going to get an upside from investing it. The only problem is, unless you get some exclusivity, that AI product is now going to go to Safeway and to Walmart and to Target in their grocery sections. Everybody's spending more on AI now because they've been told that. But if everybody's getting it, nobody's getting it. I mean, it's like, you know, collectively think of how much money we spend on SAT prep now, and then ask yourself, are we collectively better off? The answer is no, because by spending this much, we will essentially weaponize the process where you have to spend to keep up with everybody else. But you're not getting the benefit because if everybody's spending on it, they're all getting the same crap. So I think in many ways I think AI is going to become a cost item that reduces. The other aspect of AI is just like online retailing gave consumers more weapons to play off retailers against each other. I think that AI is actually going to give you and me more weapons to then essentially play off companies saying, okay, I can get this from you, but I can get this from somebody else. So I think it's going to put downward pressure on overall margin. So there'll be some companies that benefit. But collectively, AI can't push up the profits of all companies and the value of all companies because revenues for one Company become costs for another.
- [38:07]
C
That's really interesting, and that's an interesting contrarian take, I think. So, one thing just going back to a point you made that I want to just zoom in on here. You were talking about infrastructure booms. You mentioned the telecoms and kind of utility like build outs. One thing that you find if you study some of these historical episodes is this kind of, kind of ironic situation where the folks, the infrastructure builders, who were actually building the future, right, laying the railroads in the 1860s in the US or building the fiber optic cables for the Internet in the late 90s actually were the ones who made no money. They created a ton of societal value for the economy, for consumers, for other enterprises, yet they themselves failed to capture a decent share of the profits, right? Most of the railroads went bust. Many telcos did too. And those that survived, like AT&T, weren't exactly great investments. And instead it was the Netflix and Metas of the world that came out later to enjoy the subsidized bandwidth costs that were a function, of course, of the excess capacity we saw, right? This goes back to the capital cycle idea, this big market delusion you mentioned that this arms race dynamic that pushes firms to over invest in the upswing of a cycle. And when demand doesn't materialize, you mentioned deep seq being a risk, there are others as well, then you end up with stranded assets as excess capacity and falling prices, which in this case, in this thesis here would be at the detriment, of course to the infrastructure builders, but would be a benefit to the adopters of the technology, the users. You're also saying here that the collective margins of companies may actually be under pressure from AI. So let me ask you this, which is if we do believe, and maybe this is not your premise, that AI will create value through the economy, where in the chain, where in the value chain, where will the value accrue? Which companies will be winners and which losers? Not specific names, right? I'm not looking at Google versus Amazon. I'm thinking about like sectors of the economy, the chip makers, the cloud providers, the, you know, financial companies that are, that are, you know, using AI to make their operations more efficient.
- [40:25]
A
I think in the first phase, the architecture builders benefit, right? The chip makers, the data. So in a sense, we are in that golden age for the architecture company because you're building up the architecture. Historically though, you're right, once the architecture gets better, the glory fades and you move on to others. I mean, I'd like again, going back Cisco in the early phase of the dot com bubble was a big winner, right? There was a brief point in time in 1999 with the largest market cap company in the world, and then it became an afterthought, right? It basically not only did, did it, did it fall off, but its market cap actually dropped by almost 70% over the next decade. Because people said, okay, the architecture has been built. That will happen with AI infrastructure. With a caveat, which is to the extent that this infrastructure will need constant regeneration, and that's basically what the Nvidias of the world want to try to make it is these chips are not permanent. In five years you will need even more powerful chips and even more expensive chips. They're hoping to make this an infrastructure investment that is not one and done or one in 20 years, but one in five years. Whether they will succeed or not is really the bet you're making when you invest in these architecture companies. Is this AI architecture for the long term or is this AI architecture that's got to get renewed and often renewed using the same company? One of the first things, one of the first questions I asked when I valued Nvidia, especially after the AI boom, was let's say I start to build an AI architecture around Nvidia chips and let's say I want to expand or replenish this center five years from now, do I have to keep using Nvidia chips or can I switch to AMD chips? I don't know enough about AMD are. The answer I got was surprisingly messy, which is it's easier to stay with the same company in terms of chips because bringing in new chips. And that's when I heard about Cuda and the software that goes with the chips. And, you know, so clearly the companies are aware of this as well, is their glory will fade and they're trying to figure out ways to keep themselves a key part of this process. But I think there will be a phase here where you're going to see a shift away from the architecture companies towards the companies. And I would include the LLMs also as architecture companies, right? Because ChatGPT and Groq are not the devices themselves that make money, it's others who license them and make money that they hoped. So right now that's where the bulk of the market cap is, in the architecture companies. But there will be a point where you're going to see product and service companies which deliver value. My view is they're going to be more likely to be B2B companies rather than B2C companies, because I look at the AI products and services that consumers get, and many of them you don't need the heavy investment that you're seeing in. But there will be a point where, and that's what as investors you're going to keep your eye on, is if you can get one or more of those companies, perhaps a portfolio of those companies that you think are doing something that is truly value creating in the AI space that goes beyond, you know, the ChatGPT version of AI. I think those are the companies that are going to be the next phase of the winners. Right. And I think that that's going to happen with AI. Justice is I think that when I talk about lower margins, I'm talking about the fact that these companies which produce the product and services will have to sell them and the companies buy them. So there will be a value shift, just as there's been over the last 30 years, away from the consumers of AI to the producers of AI products and services. And that'll be interesting to see where the value shift comes from. But I think that that's what you should expect to see over the face of AI, the long term.
- [44:32]
B
I always think of this as the differential of moving between Moore's law and the law of. But wait, there's more. You start in that exponential growth curve, you end up in trying to bolt on stuff and keep people around in some infomercial.
- [44:47]
A
And I think that'll happen. I mean, right now it seems like it's a long ways off. But I wouldn't be surprised if it comes sooner than you think because I mean, I think one of the things we've Learned in the 21st century, the cycles that used to take 70 or 80 years in the 20th century are working in 25 to 30 year cycles. So I think as investors we need to factor that in. This cycle is going to be much quicker than the cycles you saw with railroads or phone lines. Because the world is just globalization. Technology has kind of speeded up the process. We're aging in dog years.
- [45:22]
B
You mentioned before about maybe creating a Damodron bot. AI in your process. Are you using any of these tools and how.
- [45:30]
A
No, first, I mean I wouldn't create it because it requires a tether. So there are two demoderant bots already in development that I know of that I've kind of given. One is a bot development by my good friend Vasanthar who teaches at nyu. He's actually written a book about the bot and there's a paper out there and essentially it's just an AI entity That's read everything I've ever written, watched every class I've ever taught, looked at every evaluation I've ever done, and remembers everything. So in many ways, it has an advantage over me. I don't remember what I wrote 10 years ago. It does. And he actually has used the bot to value companies like, you know, he used to value. I think he took byd, very complex company. And. And he sent me the report that the bot did, and if it were a student in my class, I would give it a B plus, which means it gets the mechanics well, it even gets my language right, the story right. What it doesn't do yet is make that leap of intuition that as human beings we make sometimes badly, sometimes well, but it's remarkably good. It's kind of scary because so much of investing in valuation is mechanical. That part is going to be very easily replicable. There's another version of the bot that's a purely appraisal bot. It's designed for people who want to value their companies. It just focuses more on the mechanics. So those are two bots that I know of. There are at least a half a dozen unofficial bots that are out there, some with dark rationales that trying to do scams. There was a bot early last year on Instagram that was a video of me inviting people to invest in a fund that I was supposedly running. And the video was actually completely AI generated, but it sounded like me, sounded like me, not just in terms of my voice, but in terms of the words that I use. It in fact, picked Palantir and Nvidia as the two companies. Knowing that those were companies, I talked about my valuation. And it took off, you know, a couple hundred thousand people, you know, kind of watching it, a few actually sent money. And so I wrote a post on it, basically, and the post I took was, look, I can sit here and complain about the fact that there's a scam out there with my name on it, but this is the world we live in. So I said, I'm going to look at it as a teacher and say, what does this scam do well, what does it do badly and what kind of creates a separation? And I tried to take as clinical a look as I could at that scam and say, okay, what does it get right? It gets a language right, it gets the words right. It sounds like me. What does it get wrong? I mean, anybody who's read my work for a long time knows that almost nothing that I do that I charge money for the very fact that it was asking people to send money was a false note. And people who knew me enough knew that right away. So I said, the first thing it tells you about the scams that are coming, and there are several coming, is when you listen or hear somebody promise you something that's too good to be true, you need to do your background research on that person. Say, does this sound like what this person would do? And I think, especially with people that you only have very mild connection to, that requires work. So it is going to create. And I think it also resolves, I think, one of the great contradictions of our time, which is we have more resources as investors than ever before in history, right? More data, more podcasts, more books. But collectively, as investors, we're doing worse than we used to. So there's something happening that's creating this disconnect between what should be and what is. And I think one of the reasons for it is we're being drowned by disinformation as much as information. It's getting more and more difficult to figure out what's real from what's fake. So there are bots out there, there'll be more out there, some with dark intentions, and there's little I can do to stop that process. All I can do is keep telling people who I am as a person. Hey, if you find something that's not me, don't buy into it.
- [50:01]
C
You talked about trying to diversify across the company life cycle, right? Growth companies, mature companies. You know, there's this notion out there that companies are staying private longer, right. And whether or not this is true, the narrative is that in order to get the access to the best companies doing the most innovative things. Let's talk about AI. You need to kind of go to private markets. So what would be your response to this? That, you know, maybe the fact that companies are ipoing a little bit later or not at all and staying private longer, maybe limiting your opportunity set as someone who wants to be diversified across life cycle.
- [50:38]
A
And you could add to that the fact that supposedly these private investments are less correlated with public markets. So it gives you, hey, that's the alternative investing pitch, right? And the question I would have is how has it worked out for institutional investors in the last 20 years jumping into alternative investing? The answer is not very well for three reasons. One is that correlation is close to one, right? Much as VCs like to claim that their returns are not that correlated, they're close to one. The only reason they don't look close to one is because the way in which value gets estimated for VCs has a lagged effect. I'm not saying there's some devious reason, but there's a lagged effect. So the correlation is close to one. Second is if you gave me a chance to invest in these private businesses costlessly, like an index fund, I'd jump in. Right? But the problem is you charge me 2 and 20 upfront or some absurd cost structure, how the heck am I going to overcome that and deliver? So as an investor, not only do you have to ask yourself, is that a good place to be, but what does it cost you to enter the space? I don't think anything good can come from individual investors being invited into private equity and the VC space. In terms of net returns, it's going to be great for the players in the game selling you this stuff. But as individual investors, I think you're playing a game that's loaded against you, not because you're a trading against people who are more informed than you, but because of the cost structures that go with these investments. So I'm willing to give up on our wait until these companies go public. But it is true they're waiting longer. In fact, when you compare IPOs in the 1980s to IPOs in this, and especially in the last five years, the companies that go public are far bigger in terms of market cap. Far, far bigger. It's not just inflation effect, it's 10 times larger, 20 times larger in terms of market cap at the time they go public. You have $100 billion companies going public, which never used to be the case. But here's the catch. The companies are less formed as businesses when they go public, which is strange because you think they're bigger companies. If you look at Apple and Microsoft's prospectuses from Prospecti. I don't know what the plural is. From the 1980s, these were small companies that had actually figured out how to make money by the time they went public. You look at Airbnb in 2020 or Uber or any of the companies that are big companies have gone. These are companies with. They still haven't figured out how to make money on a consistent basis. So we're creating these bigger companies that are actually less formed before they go public. And I'm not sure that's a good thing. I think that the supply of capital in this gray market space has allowed them to scale up, but nobody seems to be asking them. You're scaling up, but are you building a good business? So maybe some discipline needs to enter that part of the process where you're building up companies in the private space and you're also asking the right questions about them building businesses while they're getting larger.
- [53:50]
B
You made a mention, I believe it was in the Scott Galloway interview, but I think you've been talking a little bit about it elsewhere, too. You're considering moving some portion of the portfolio to cash, collectibles, physical assets.
- [54:02]
A
I mean, earlier, Matt, we talked about selling things when they get overvalued. I've always done that. Historically, when I've sold things for overvalued, I put them back into things that were undervalued. And the last year or two, that's the part of the process that's changed a little. I feel less willing to load into something that's undervalued because I don't. First, I'm not finding enough that is undervalued to jump in. The second is I'm wary. I'm not a market timer, but I'm wary about where we are at the market in terms of pricing, in terms of what we're building in. It seems while there's a pathway we can justify where we are today, there are also multiple pathways where things can go bad. So from that perspective, I want to move my money into things that are not correlated with equities, and that is getting tougher and tougher to find. I can go into cash. That's always a choice. But 30 years ago, I could go into another geography and be okay. So if I were primarily in US equities, that's not working anymore, because in crises, equities across the world move together. You know, I could look for Real estate used to be uncorrelated with equities. That's no longer the case. Real estate is starting to behave more and more like an. In our zeal to securitize things, I think we've also created a nightmare, which is we've created asset classes that used to be separate asset classes that are now starting to behave like equities. You're saying, what about Bitcoin or cryptos? The problem is bitcoin behaves like very risky equity. If I think stocks are overpriced, the last place I want to be is Bitcoin. Because of the correction to tech stocks, it's going to be double that correction in Bitcoin. So there's no easy place for me to go. And I said, I've historically not looked at collectibles. You know, I include gold as a collectible because even though it's a commodity that's basically its role. But I think that this might be a time where if you cash out on a stock, some of that money should find its way into things that are. And that remains one of the few spaces which is relatively encore with equities. So I think that's what I was talking about. I wasn't talking about selling off half my stocks and buying gold with it because that's not something that I think will work for me in the long term. But holding something in an asset class that's not going to do as badly as stocks of stocks have a bad year is something that I think about more actively.
- [56:41]
C
So what you're saying is take some chips off the table, put them into assets that are uncorrelated with the stock market. We don't know when the correction will come, if at all. But that gives you some optionality. In the event that stocks do fall, you can get back in. So let me talk to you about that, which is how do you think about reentering the stock market? Are there trigger points? You mentioned limitations sell orders before. Is there a limited buy order? Because obviously there are, you know, plenty of horror stories of folks who, you know, got out and got out before GFC never made it back in. Right. So. And that's exactly how are you enforcing discipline on yourself?
- [57:17]
A
Yeah, that's exactly why I would not sell 80% of my stocks or 60% of my stocks. Because even if you're right about the correct. Let's take your best case scenario. You get out of stocks just before the correction hits. You're patting yourself very vigorously on the back for great job done. Right. The problem you're going to face is you're now sitting on a mountain load or a hill load, depending how much money you have of all cash. And you know, you should get back in. But you're worried. You're saying, but what if this is not the right time? That's the scenario where people who got out in 2008 did not get back in 2009, 2010. There's still people who got out of the market after the 2008 crisis that have not gone back into stocks. So one of the reasons not to get into a 50% cash, 60% cash, 80% cash scenario is that decision becomes more monumental. But if you end up with this big cash balance, here's my suggestion. Put it on autopilot. What does that mean? Say, look, I'm not ready to put all of this into the portfolio, but I'm going to pre commit and you Got to find a way for that pre commitment to stake to take a quarter of this money and invest every six months for the next two years so end up back in stocks. You can either if you feel that you can't pick stocks, you can put it in index funds and let it sit in an index fund. But you got to get that money working for you again because you can't leave it in cash because it'll stay in cash for much longer. And I think you look back at people who timed the last correction and you look back at what their overall returns have been over the last 15 years. Timing the last correction got wiped out very, very quickly and they've ended up in cash too long. So if you find yourself with a significant amount of cash, you might want to take that decision out of your hands because clearly it scares you too much to make that judgment at one go. So split it up into pieces, go in pieces so that at least some of that money goes into play every three months or every six months.
- [59:19]
B
I've got one more question for you and it's about this logic of what the portfolio looks like as you look at what's. Because we're all alive here, a diminishing time horizon. How do you think about setting up those rules, those practices and the wealth that's going to be around beyond you.
- [59:37]
A
As you get to a certain age? You're not investing for yourself, you're investing for the next generation. In my case, the next two generations. And maybe, you know, so I think in a sense your time horizon is not your life, it's what your investment is designed to do. Right? So otherwise at 78, you're saying my time horizon is two years, I'm going to do whatever I want. Right. So I think that you're, you're the person, the people that you're, that you're managing. The portfolio will change over time. So your time horizon will change as you go through. Of course, if you have nobody to leave the money to, your time horizon then becomes with charity and leave the money too. And that's, you know, that too can affect your time horizon. I think actually I'm more willing to take longer term bets now than I was 30 or 40 years ago because I know I'm less wary of my, you know, of other things that can happen that are unanticipated. I mean, healthcare is the, you know, one of the things you worry about to the extent that you, that that could be a big chunk that might be a reason to hold some cash. But I think your time horizon will shift over time for lots of reasons with aging for me has not been one of the top five reasons. It's been other issues.
- [60:50]
B
As well. It's always a great pleasure to talk with you and go through all this work. People want to find more. The books you're writing, where would you like to send them?
- [61:00]
A
I mean you need to find. I mean one of the advantages of having an unusual name is you do a Google search, you're not going to get a bunch of aswathi mother and you're going to get just one. So you can go to my My launching pad is my webpage. That's you can find everything from there if you're if you don't. But it's it looks like a 1990s web page because it is a 1990s web page. So it's got that old structure to it. It looks old fashioned. But you know, I find that, you know, I need to keep the content there. So to me, if I play with the web page and create JavaScript, I completely lose control of it. People offer to rebuild that web page. My blog is where I go whenever my thoughts get out of control or when I start to worry about things like or question that I can't answer. I find that writing allows me to think through my confusion. And often when I think through my confusion, writing about that confusion actually helps as well. So most of my blog. My blog is the first place I go with my thoughts and from there everything else flows. My teaching, my writing, my data. So my webpage and my blog would be where I would send you.
- [62:13]
B
We'll put it all in the comments. Thanks for watching. This is Excess Returns. Kai. Thanks for joining me today, Professor Demidiron. It's always a pleasure. Like subscribe all the things below. We will talk to you real soon.
- [62:25]
A
Thank you. Take care. Thank you for tuning into this episode.
- [62:28]
B
If you found this discussion interesting and valuable, please subscribe on your favorite audio platform or on YouTube. You can also follow all the podcasts in the Excess Returns network@excessreturnspod.com if you.
- [62:40]
A
Have any feedback or questions, you can contact us@xsreturnspodmail.com no information on this podcast.
- [62:47]
C
Should be construed as investment advice. Securities discussed in the podcast may be holdings of the firms of the host.
- [62:53]
A
How many discounts does USAA Auto Insurance offer? Too many to say here. Multi vehicle Discount Safe Driver Discount New vehicle Discount Storage discount.
- [63:02]
C
How many discounts will you stack up?
- [63:04]
A
Tap the banner or@usaa.com autodiscounts restrictions apply or their clients.