Today's show is sponsored by YCharts. Future Proof is on the horizon, and YCharts is taking the stage during the event to talk about one of the trickiest parts of Advisor's Day: communicating investment strategy on September 16th, 12:30 p.m. That's West Coast time. On the show and tell stage, YCharts will present the psychology of persuasion in client meetings, taking a deep dive into how the right visuals build client confidence and keep conversations moving forward. After the session, stop by their booth to see up close how YCharts simplifies every stage of the client relationship, from risk alignment and proposals to tough questions that come up mid-meeting. Click the link in the show notes to learn more about their session and get 20% off your initial YCharts Professional subscription. New customers only.
Welcome to Animal Spirits with Michael and Ben. It is Tuesday, September 1st. That means back to school.
You know you want to say it. Summer went so fast. Summer went—I can't believe it's fall. Summer went so fast.
No, no, no, no, no. I do that spiel in the first week of August.
But this is a good season. Great season.
Football is back. Football's coming. What? What? What? Say what you're going to say.
That's okay. It's a sign of middle age. I like—every year I like it, it gets better and better.
Football's coming back. This could be the Giants' year. We might win six games, maybe seven. Who knows?
It's a time of the year where you start to get into fantasy football emails, which, Ben, you're not a fantasy football guy. I know this, right?
And I want to give a shout out to all the ThanksKamish people. I know, again, you don't know this, but I think maybe there's two people in the audience laughing.
I've literally never signed up for a fantasy football team in my life. I'm pretty proud of that fact.
That's not a badge of honor. I don't think that's pretty lame. It's very lame. It's ludicrous behavior. No offense.
Have you ever talked to one person who's happy about playing fantasy football? Everyone who plays fantasy football is miserable.
I mean, we're past that stage in our lives, but you never did it.
No, just like I never signed up for Facebook.
Badge of honor. Enough of this. Before we get into the show, we have a listener survey. We want to hear from you, or we want to hear about you. I don't think we've done this in—I don't know, it's been years. We know nothing about our audience, the demographics of our audience. Who are you people? We hear from you. We love the emails, but we want to know. So, if you don't mind, take a second. Link in show notes. If you can't find us, email us. Thank you. Thank you. Thank you. All right. No, just for all the great recommendations we've given over the years, it's totally worth it.
All right. This was the cover story for Barron's, Ben. Is the AI capex bubble about to burst? What 250 years of market history tells us. This is the story, Ben. And this is the story that is going to be with us for the duration. I think sometimes we talk about the market as if there's like passing themes and passing fads, and of course there's lots of that, but this is one of them that I think will ebb and flow right over the years. It's not just—we're going to get sick of talking about this. I mean, I think I'm already getting there, but this is the big one.
It's not going away. So in the New York Times they also had a story: This is how the AI debt binge sinks the economy.
Okay. Yep. So this is going to be the thing, and someone wants to be the person who calls it. Like people saw what happened from The Big Short. I want to be the next Michael Burry. I want to be the next Steve Eisman. You have to plant your—someone has to plant their flag to be the person who says, 'Hey, I called how it was going to happen.' I give people credit for trying, but I think it's just going to be really hard to do.
Did you listen to the RTF episode with—yeah. He wants to be the next Michael Burry, right?
Yeah. He's planting his flag hard. That guy Ed Zitron, new to me, but he came in like a tornado. Very bearish.
He's been bearish since the start, but he's got all the hallmarks of wanting to be that. It's like, I'll move the goalpost across the field if I have to. Right? So that's how it happens. And you're right. In five years from now, are we still talking about this?
Well, here's a quote from Ben Wrighty from that article in Barron's: I think we're going to just keep going. This will be bigger than anybody thinks for longer than anybody thinks. I could, to use a Ben Carlson-ism, I could see that.
You called me out on that when I was in New York. It's true. Here's the thing. We talked about this with the economy probably five years ago. We had this conversation. We said, 'What if the economy just keeps growing for the rest of this decade?' And honestly, unless AI does it, we're not going to have a recession the rest of this decade. Like, if AI doesn't sink us, it's a higher probability event that we don't have a recession the rest of this decade than we do. Is that fair?
We don't have recessions anymore. Unless something comes—unless a meteor strikes or the AI capex story completely falls out of bed, we're not going to have a recession this decade. We're just not. That sounds really like a dumb thing to say, but I think with those caveats it's true.
My brain just short-circuited for a second. I remember what I was going to say. When did Galloway write that book, The Big Four or The Four Horsemen or whatever it was called?
Okay. What was the original MAG7 called? What was it?
FANG stocks. Remember it was FANG stocks. Ah, there we go.
Facebook, Amazon, Netflix, Google.
Then they added another A to add Apple. So it was FANG.
And then Microsoft, Tesla got in there.
All right. So anyway, but the point is FANG came along. I think Cramer coined it actually in 2016, 17, maybe 18. I don't remember. It's been seven or eight years of non-stop talking about the same group of stocks.
No, 10 years almost. Yes. So, I do think this is the thing though that could—So, Robin Wigglesworth wrote this one for the New York Times. He's got a great bond book that just came out that I got an early copy of. It's—
Wait, did he move—he moved to the Times?
So, he says, AI companies tapped debt for $27 billion last year. Okay. By mid-August, it's already close to $450 billion. It's going to be about $600 billion for 2026. That's more than the combined budgets for the Department of Justice, Transportation, and Education. Like, trying to put these things into context. Okay. Future capital expenditures planned to account for roughly 3% of GDP from 2027 to 2029. Back-of-the-envelope calculations suggest that it's almost a trillion dollars a year. So we've never seen anything like this. And it's kind of funny because people think with numbers that big that it has to end in a crash. But don't—this is my thing: if they keep spending this much money, there's no way the economy is going to stop and slow down.
Well, the bears would say, 'Jackass, of course they're going to slow down. This is unsustainable. They can't keep spending this much money.' And right, well, we shall see.
Down the hall to the right, right before the doors here.
That did not sound like a New York accent.
I work in a shared office complex and people are always poking their head in. 'Hey, do you know where this office is?'
Why are they asking you? Do you look like the guy? You don't look like that guy.
He asked me where the bathroom is. I get asked—people call me to my office all the time. I should put a sign up.
You just got one of those faces. You look friendly.
New background. I like it. Shelf space. Did you do that yourself?
I got new bookshelves. I put them—You know what? I'm a man of the people. I still will use an Allen wrench to put together furniture.
I feel like you're saying that as if I am not Mr. Allen Wrench.
Don't come to me with that.
You literally got one of our colleagues to hang up your—
That's different. That requires their skill in hanging photos.
You will put together furniture—
All right. Person locked. So here's the thing. I do feel like almost everyone thinks that the AI ending has to be black or white. It has to be something on the extreme, like we're getting 10% GDP growth with all this productivity, or we're getting the biggest crash in history.
Wait, you know what? Let's plant our flag. Ben, we're team middle. We're team middle, right?
We're team gray. So, it's interesting because Torsten Slok had this new slide out that he released, and he says if AI succeeds, rates will be lower. If AI fails, rates will be lower. So he's one of the few people who's saying like he's kind of bullish on bonds. So he says, 'Listen, if it generates all this money, it'll be massively deflationary,' which I still don't quite believe because of all the money they have to spend for these data centers. 'And if it doesn't work out, the bubble bursts, the Nasdaq is down 50%,' and then people go into bonds. So he's saying regardless, bonds are in a great position. I just don't know that it has to be one of the extremes. And I think that's just what people kind of want to happen secretly. Like we want to have this big bang, this ending crescendo, whatever, and you just wonder, what if it just isn't that, and it's not going to be an extreme? Is this an always feature of the stock market, is it just like waiting for the next chapter?
I just don't know if we've been trained because of the dot-com bubble blowing up and because of the Great Financial Crisis. We had twin peaks and that scarred a generation of investors, understandably so. I just wonder if even if we get the AI bear market, there will be an AI bear market at some point. Is it going to be a V-shaped bottom?
That's the thing that would leave no one—you wouldn't be very satisfied with that outcome if you're an extreme person.
I say of course, just kidding. Who knows? All right. So Kevin, they did the Jackson Hole thing this week, and I read his prepared remarks, and I liked it. I thought it was good.
I don't seem like the macro tourists were all happy that he actually gave some forecasts and talked about the economy. I don't know. I still don't think it matters.
You don't think what matters?
He could have spoken in Latin and it wouldn't have mattered to the market. I just don't think the Fed giving a forecast of the future is meaningful in any way.
He's not doing that. He's doing the opposite. All right. So, Ben, just let me read Kevin's speech because obviously you're naive to the subject. You don't care and you're not listening. So let me read it. Capital and labor have combined to create the large language models at the heart of AI. Users buy tokens to gain access to the models. Reports put annualized token sales for the two leading labs alone at more than $100 billion, an increase of 500-plus percent from a year ago. The Fed watches all of this attentively. We recognize that AI is a new variable, potentially a new factor of production that will have consequences for both the economy and the conduct of monetary policy. It opens some major lines of inquiry. Will the application of AI cause a significant sustained rise in productivity across the economy? And if so, when?
I'm falling asleep. It sounds like he used AI to write this. He's not saying anything new here that no one knows.
Yeah, listen, I'm not Mr. Fed Watcher, but I feel like you think you're like dropping a mic or something right now.
I'm not dropping a mic. I'm just saying, what did he say here that no one is already contemplating? It's interesting to hear the head of the central bank talking about the state of the economy as AI as it pertains to—no. Okay. He knows just as much as we do though.
Okay. What else do you want to hear? What else did Powell say that's interesting to you?
Sorry to poo-poo this, but I think it's worth poo-pooing.
Well, I have some of Powell's stuff sprinkled throughout the doc, which I feel like I should just delete now if this is how you're going to behave, Ben. Here's what else I thought he said that was interesting. Bear with me.
Wait, he used the word heterogeneity. That's pretty—
All right. Tell me if you like this, Ben. I wish our understanding of the economy were so precise as to provide a mechanical, tried-and-true answer that some simple function like a Taylor rule could be rigorously relied upon. But our knowledge just doesn't extend that far, at least not yet, and the factors most relevant to the proper conduct of monetary policy change over time. Providing forecasts to illustrate the Fed's reaction function works better in theory than in practice, better in the lab than in the field. I'm not alone in noticing that forward guidance in 2021, to cite one example, might well have slowed the policy response to high inflation. In my term as chairman, my colleagues and I will endeavor to construct more reliable models and more robust rules to guide policy decisions. We'll do this knowing that accuracy in economic forecasting is still just an aspiration.
Okay, that's actually pretty good. Not bad.
I mean, that sounds like Ben Carlson wrote it.
Yeah, that's fair. I like it. Duncan says you have an Allen wrench broker, by the way. That's pretty good.
Wait, so a couple weeks ago, I can't remember what we were talking about, but you said, 'Listen, we talk about something that gets disrupted or something that crashes or a risk that's bad. We talk about it and we look at every single angle and then we move on and never talk about it again.' And I feel like this has happened with software stocks. Software stocks got destroyed. They fell what, 35% as a group, IGV, the ETF.
Yeah. Individual names fell a lot more than that.
Yeah. 50, 60%. I remember you were dipping your toe for a while at maybe Salesforce or some of these stocks.
Yeah, I think I probably broke even.
Okay. So these stocks have almost round-tripped. IGV is almost back to all-time highs. I think it's 40% off the lows. And if you look at this, this is a V-shaped bottom for software stocks.
This doesn't look like a V to you?
No, maybe a W. No, this is an interesting chart, but no, it crashed, rebounded, fell hard again, and then—but I guess the look at ServiceNow and Workday and Salesforce and Adobe and Duolingo and Figma, those were not V-shaped bottoms at all.
But this was almost 40% off the highs. IGV, the group. I'm talking about software as a group. It's now less than 7% from the highs.
We don't need to argue over the letters, but the recovery was astounding. Yeah. I just—it's these stocks. Listen, 'software is dead forever,' this is what people were saying, and now they came roaring back already.
This is what happened. This is what's going to happen with AI. There's so much back and forth. And I think there's going to be—this is the overreaction function that we're going to be seeing. Companies are going to lay off a bunch of employees and go, 'Oh crap, why did we do that?' Like this is how AI is going to be for the next 10 years: that, 'Oh no, we way overhired people. We have to downsize.' Like there's going to be a lot of that back and forth.
Yeah. It's really hard to invest actively through mega disruption.
It is. I know it is, right? Picking the winners and picking the losers, because look at how quickly this changed.
Yeah. And this in particular, I said I wasn't going to nitpick over the letters, but I'm just going to dive back in for a second. If you were trading this, and I was, this is brutal. So, a slow fall, then a crash, a little rebound, back to new lows, a monster run from—let's use the ETF for example, IGV, from 75 to 105, so a 30% rally or more in a couple of weeks, and then another quick crash before rebounding to new highs or new recent highs. It kind of looks like, remember when Adam Sandler tries to draw a cursive Z in Billy Madison? It looks like that, because he doesn't know how to do the cursive Z, which I—
I mean, BB gun to my head, could I draw a cursive lowercase Z right now if I had to?
I'm thinking—yeah. No. No.
Maybe one of the most useless things I ever learned how to do is write in cursive.
But you could also make the case this is why this is so hard. This could be a dead cat bounce easily.
Well, I think you saw a lot of software stocks. Like you could say, okay, in five years, that was a dead cat bounce. What were people thinking? But the truth is, and this is very unsatisfying—the stuff that we're saying today is very unsatisfying to people, like, 'Oh, you don't know how it's going to play out.' But that's the thing, no one does. Well, this is why especially in the world of attention-seeking, which is the world that everybody is in these days, somebody like Ed Zitron, who I never heard of until I saw him on the calendar, can go from 0 to 100 because people seek certainty. And when you speak with confidence and you speak fast and you speak with numbers and you cite sources, you build an audience.
So, TB had this quote that he stole, I guess the Roman army had this thing where they say, 'Action removes fear.' Like that was their—
And so that's the thing. If you grab hold of the steering wheel and you think, 'I'm going to do this because I think I know it's going to happen,' or that person knows what's going to happen, it makes you feel safer and more comfortable.
Well, you telling people you have no idea how any of this works or any of this plays out, it's like, all right, so I guess I'll just listen to somebody else. Thanks.
Right. I don't know either, but the thing is, I think you still have to have a baseline. Right? This is the baseline I think will happen. I think this is the most probable outcome, and I have these two tails over here that could also happen. I think—
Credit to me, my baseline changes every week.
That's true. Yes. But I think you have to have a baseline.
All right. Last week, my knee-jerk reaction to the drug using AI was, 'Who cares?' Now I understand why some people view writing as sacred and holy. I really understand that point of view. Druckenmiller immediately was like, 'Yeah, of course I used AI. I'm not a writer. I'm a trader. I'm 73 years old. Why wouldn't I be using AI? These are the tools available.'
But that was the only response he could have. But that was one of those things where, like when I catch my kids stealing candy out of the pantry, like he was caught red-handed. He had to go, 'Oh, of course I use AI. Why wouldn't I?' Like that was—he probably didn't think he was going to get caught. Let's be honest.
I don't think he gives a damn at all.
Okay, that was one of those—
Why would he care? It's honestly—I don't even think it's embarrassing from his point of view.
It's a little embarrassing. I think it's embarrassing in the Wall Street Journal to use a piece written entirely by AI. It's very embarrassing.
I don't—there's no way that he wasn't embarrassed, and he totally spun—that was the biggest spin zone of all time.
I hear that point of view. I just disagree. I don't think he cares.
He completely spun it. That was one of those, 'I can't believe people—wait, people can know these things?' If you're that age, you don't know that someone could look up if you used AI? Come on.
All right. Great chart from Rich Bernstein. This is interesting. So they looked at the consensus growth EPS estimates by company in the All Country World Index. So any of these stocks have a projected earnings growth rate of more than 25%. So this is looking out pretty long. And he said this is interesting. So it's like 200-some stocks of the world. There's only one MAG7 company in here. And he color-coded them by foreign developed, emerging markets, US ex-MAG7, and then MAG7. And there's a lot of foreign companies on here.
It's interesting. So this is a really good cutout as far as you would say: Investors scoffed at our suggestion that the stock market could rotate away from the momentum-driven AI and technology theme without a bear market. However, that has indeed been 2026's story so far. This is this thing where we're seeing potential earnings growth come from all these different stocks. This is not something people had on their parlay for this year.
Check this out. I had Chart can make this yesterday. I'm throwing it in the doc. I asked them to show me net income for the 493.
Right. Because if it's all just circular spending—
—and debt-fueled earnings.
Yeah. Take out the circular spenders.
All right, so look at the 493. The net margin has exploded to 13.6%, up from like less than 12 in 2024, which is a big move.
It's interesting because if you're an AI bull, you think the technology—if you think the technology for AI is going to do what the people say it's going to do, you have to be bullish on these other stocks, on the 493.
Yeah. And you have to think that this margin is going to keep inching its way higher and higher. So it's gone—so for people who aren't watching, in Q4 of 2023 it was below 11%. It's projected by Q4 of 2027 to be almost 15%. That's a meaningful boost to margins for these other companies.
Is it just a coincidence that these margins are expanding and exploding when AI hits the scene? Can't be, can it? I don't know how much of it is because of AI. I have no idea. But—
That's what's hard to say with these relationships because there's other stuff going on, but AI is the thing people latch on to.
Speaking of AI bulls, I listened to the entire podcast of Leopold Aschenbrenner from Situational Awareness and Dwarkesh Patel. It was over four hours long.
And how long ago did they publish it?
Okay. Before anyone really knew who he was, at least outside of tech circles, right?
Yes. So this kid is different, obviously. He went to Columbia at 15 years old.
Yeah, he's Doogie Howser.
Yeah. Graduated, was a valedictorian at 19 years old. And I found a lot of the conversation to be completely over my head. I didn't retain much of that four hours and 15 minutes. But some of the takeaways I had—so at the end of it, they spoke about, like in the last 15 or 20 minutes, they spoke about the fund, Situational Awareness, what he was planning to do. And of course he said he was citing a couple of hedge fund managers that got the thesis right but didn't manage the fund properly and blew up, and he said rule number one and number two, don't blow up, which obviously didn't age great. But I found him to—I was surprised by his humanity. I thought that, like, oh, this kid is just completely a complete alien and has—
He's not just a robot, you're saying?
Yeah. I thought that he was speaking about his exit from OpenAI and he got booted, and he acknowledged that he could have been politically more savvy about how he went about doing things there. Like I thought that he—I was surprised. I liked him and I'm rooting for him. I came away liking him.
All right, I'll give it a shot.
Well, I'm not saying you have to listen to it. I just wanted to hear this good.
Let's talk about bull market behavior. Go ahead.
James Safford from Bloomberg put this together, and I was kind of thinking of doing something similar, but he did it better than me because he used the headlines. All this stuff has happened in the last, I don't know, six months I guess. So in the past week, Vanguard purchased Altruist for four or five billion. T. Rowe Price acquired FM Investments. We've had Alex Morris on our show a number of times. Goldman Sachs acquired Neos. We've had them in Talk Your Book before. Goldman Sachs also acquired Innovator. We've had them on the show before. First Eagle got acquired for $7 billion by Victory Capital. This is all total bull market behavior. Asset managers getting acquired, custodians, right? Building scale. Like, these are the kind of headlines you see. This is rampant bull market behavior.
This stuff obviously doesn't happen in a bear market. Do you think that if 2027 and 2028 are really pretty shitty years because investors are underwhelmed or whatever, the mood changes, that we're going to look back and say like, 'Man, that was embarrassing. Overpay?'
No. No. Not these companies. I mean, you and I.
Are you going to be embarrassed for not seeing that this was obviously a bubble?
I'm going to say I don't think I will be.
No. We've been hashing this out for years now. Yeah. How every check, every thing is checked. Every box is checked for this being a bubble.
This could either go on for years.
This is great. It's hedged. Yes. But it's also true that these headlines have been going on for years as well and it hasn't meant anything.
So, we've had five years of this. The thing is, how could these deals—
Dude, hold on. It's been longer than five years. In 2018, people were saying this market is long in the tooth. Right, since the beginning of our careers, people were saying—
I mean, if we get another 20% year this year, we really are approaching late-'90s levels in terms of returns. And yes, it'll be obvious. Listen, all the debt they were taking on, the circular deals, the fact that AI wasn't having as much of a productivity boost as people thought, the jobs market wasn't as impacted. 'Of course, you idiots, of course.'
Yes. It'll be obvious. Yeah.
Whatever outcome happens will be obvious, as always. But the fact that people have realized that asset management and wealth management are a good business, especially as markets are growing, you know, relatively high margins. You get a built-in growth kicker for these businesses because people look at these numbers and they go, 'Oh my gosh, how are they paying that much money for those amount of assets?' And in that sense, they don't make—but if you think about the built-in growth rate underlying these businesses and the margins they have, then all of these deals probably make sense.
That assumes that the markets keep going up, of course.
All right. Put this one in here this morning. Does anyone care about dividend yields anymore? Our colleague Tadas Viskant posted this chart. I guess he got a note from Bespoke about this. The dividend yield on the S&P 500 is now a touch below 1%. 98 basis points.
This is, I believe, the lowest level in history. And so if you compare bond yields to dividend yields, obviously it's drastically different. And for a while there, dividend yields were actually above bond yields, which is very rare.
This is the type of thing that is completely—and there's so many examples of this. It's just distorted by the tech companies.
Right. They don't pay dividends. And they're 40% of the index.
Think about what the dividend yield would be. I'm sure someone can do the math on this, or I could ask Claude.
You call him John Claude now. Is that it? I kind of like that.
What would the dividend yield on the S&P 500 be if they took all the money they're spending on capex and instead paid investors in dividends? And what would investors rather have? Companies reinvesting in this huge innovative technology or giving them cash back in the form of dividends? And what do you do with those dividends? You just call it back into the market, I guess. So, I guess my point is no one cares about dividends anymore. And I think especially with all of the other forms of income that are available now, no one really—like people are so much more enamored with options income than dividend strategies. I don't think this really is a big deal. Now, again, in five years, people go, 'Hey, idiots, the dividend yield is below 1%. How could you not see this?'
All right. We haven't spoken about South Korea in a few weeks, the epicenter of the mania.
My favorite stock market on earth right now.
Well, it's not anymore. It's over.
But they're going through this boom-bust cycle so fast, and all the retail investors are learning some valuable lessons. So there's a chart in Bloomberg showing the turnover of Korea single-stock levered ETFs tied to Samsung and SK Hynix, and it went basically to zero from a peak of almost $50 billion in—oh, that was just in a week. Holy cow. All right, so they said key to the collapse in demand has been a series of regulatory tightening moves. Most recently, a rule to complete five-day simulated trading. Investors must download a Windows-only program on PCs and spend at least an hour a day learning the ropes and the risks of leveraged trading with virtual cash. Yeah, nobody has time for that. Nobody's doing that. So, no more trading. Good.
Isn't it kind of funny that when bad behavior happens, all these other countries can kind of wave a magic wand and just put rules on it and it goes away? But we would never do this in a million years here.
No, there would be Robin Hood traders rioting in the streets if this happened.
Hey, it's a free country. I could blow myself up. What are you—
Yeah. What are you doing? South Korea, they're like, 'Listen, we don't want people blowing themselves up on leveraged trading, so we're going to do that.' That's hilarious. I didn't see this before that they have to download—Why do they have to download a Windows-only program on a PC, so you can't do it on your phone?
They're making it hard. All right, Ben. Somebody tweeted, 'Dividendology.' Speaking of dividends, if you could be guaranteed a 7% annual real return on your investments every year, would you take it? Speaking of bull market behavior, you should see the replies here. But—
All right, this is not a clever answer. The answer to this is an obvious yes. Of course. Yes.
This is—over the last 100 years, the annual real return on the US stock market is 7%.
10% nominal. Take away 3% inflation. You get—if you were guaranteed 7% real every year and you could just put your money in there and book it.
I'd put 95% of my money in there and keep 5% for gambling.
Yes. So you're saying the comments were people saying no, that's way too low? I mean, there were some intelligent people jumping in.
So this is above the rate of inflation, whatever inflation is.
Of course you would take this.
Yeah. Turning to the economy. This is an interesting juxtaposition of the United States versus Europe. The Wall Street Journal has a headline: Corporate America's profits are booming and signal more good times ahead. That's the Journal. The FT: US corporate profits surge to record as worker payouts wilt.
