The fraud cycle follows the financial cycle with a lag, and the longer the financial cycle goes on, the more fraud is ultimately uncovered on the down cycle. So I've already dubbed this the golden age of fraud, and I suspect that when we're on the down part of this cycle, the bodies will float to the surface as they always do. But remember, the corollary to that is the harshest prosecutor and the staunchest defense attorney of a company is its stock price. Nobody goes after frauds at all-time highs.
Today's number: 25. That was the percentage decrease in Canadian travel to the US in 2025. Ed, what did the beaver say to the maple tree?
It's been nice knowing you. How are you, Ed?
The weather's nice. I've been going out to Long Island, getting tan. Had a wedding this weekend, which was nice.
Yeah, the weddings, they're happening. They're coming. Every other weekend it feels like. At first it's fun, and then it gets a little that you're like, 'Wow, it's a lot of trekking out.' But it was a beautiful wedding and a really good time. So, you know, I'm very, very happy for my friends.
In two or three years, you'll experience the do-overs where people realize like we only got married because we were 27 and it didn't work. So there'll be a few of those. In 10 years, that's when the really ugly divorces set in because they have kids.
Then in about 20 years, you'll go to the second weddings when you decide to tie the knot. I just can't wait. I wish I could be a fly on the wall when you have a very rational conversation around you suggesting that you take the money for the wedding for a down payment for a house. You're the kind of guy that would suggest that, and I just wish I could see it.
Actually, I think it's the opposite with this. I think I want the awesome wedding and to splurge on the party, and I don't think she cares that much about it. But I think having a sick wedding is — I think you got to have a sick wedding. I'm a little probably don't have my head screwed on quite straight on that one. What I would suggest is just throw a huge party and don't tell anyone it's a wedding, because when people hear the term wedding, they mark everything up by 50%.
That's a really interesting point. Yeah. Maybe just don't get the wedding planner, get an events planner, pretend that you're sort of celebrating your birthday, and then you kind of Trojan horse a wedding. It's an interesting, interesting strategy.
Or if you do what I did, just go down to city hall with a woman who looks like she's about to start dilating and get married in front of a judge. That was the romance I brought, and we had to call four different people to get a witness that morning. That's how I expressed my love and my undying commitment.
I love it. I love it. So you're not a fan. Are you not a fan of a big wedding, or where do you stand on the issue?
So first off, what you have to acknowledge is in all of these decisions, you're an influencer, not a decision maker. And that is — I'm a big believer in dividing and conquering in a partnership. And in my partnership, I'm in charge of money and movies and everything else. She decides because she has much better instincts and judgment than I do across everything. So she pretends to listen to me and she nods, right? You know, I say, 'Well, you know, we should stick in the US for high school.' She's like, 'Oh, no. We're moving to Europe.' I'm like, 'Okay, just tell me where to be when. Just tell me where to be when. Send me the address.' So, but yeah, get used to that. Decide, pick one or two things you're really good at, and just acknowledge every other decision is going to be made by your partner.
Okay. Money and movies is a good — that's a good combo. That's fun to be in charge of those two things.
And you're good at both of those.
I'm outstanding at both of those things. Yeah, that's my value add.
Speaking of money, let's get to our guest. Let's do it. Our guest today is one of the most famous short sellers on Wall Street. Over the years, he's been given nicknames such as the Darth Vader of Wall Street and also the LeBron James of short selling. He first made a name for himself in 1982 when, as a junior analyst, he urged clients to bet against a piano manufacturer that had expanded into insurance. Just months later, the company filed for bankruptcy. But he is perhaps best known for calling the collapse of Enron before it imploded. That bet cemented his reputation as one of Wall Street's most respected skeptics. Now he is sounding the alarm once again, this time about AI. He has argued that today's AI boom may be an even bigger bubble than the dot-com era. So we wanted to understand why one of Wall Street's most successful contrarian investors thinks the market has become so euphoric and what he believes investors are missing. Here is our conversation with the legendary short seller Jim Chanos. Jim, great to have you on the show. Thank you for joining us. I'm going to start with a simple two-part question for you. Is this market in a bubble? And if so, is it going to pop?
The market is very, very expensive. I learned a long, long time ago, about 40 years ago when I started my firm, that the market was inherently unpredictable, but that there were companies and sectors within the market that are often much more predictable. And so I don't know if the market, broadly speaking, is in a bubble. It's certainly quite expensive, as expensive as pretty much it's ever been, right up there with 1999, 2000, the dot-com bubble you referenced. But look, I mean, it's been expensive for a number of years. However, what we are seeing right now is an unprecedented capex boom, and capex booms tend to end badly. They tend to leave behind very productive assets, as they did during the dot-com and telecom buildout, or going back further, the railroad buildout of the 19th century. But investors often get burned along the way in financing that buildout, and that's my concern right now, particularly as it relates to physical assets related to AI.
What are some of the biggest concerns that you're seeing in the market as it relates to AI?
There are several things going on. There's the circular financing. There's the explosion in AI debt. There's the fact that the debt appears to be increasingly going off of the balance sheets of the hyperscalers and stuffed into these SPVs. I mean, there's a lot going on to be sort of skeptical of.
Which things do you see as the largest concerns in AI right now?
Well, if we go back to the dot-com analog, most of the spending that was done, and Scott knows this, back in the late 90s was by enterprise and by relatively profitable telecom companies. The spending by the dot-coms and the CLECs and the fiber guys was a relatively small amount of the total capital spending. The companies that had, in effect, unprofitable business models that remained unprofitable, and basically a lot of them went bankrupt. Most of the spending was done by AT&T, by Bank of America, by Coca-Cola, who were networking their systems together to take advantage of the internet. And then on top of that, you remember we had the Y2K, and I remember my firm, we replaced all our PCs in 1999. So a lot of the capital spending, budgeting and spending was by fairly profitable companies. What happened in the dot-com era was that they just got too optimistic in terms of how much they needed, and order books collapsed beginning in late 2000 into 2001. And so S&P earnings collapsed 40% from the middle of 2000 to the middle of 2001. They dropped as much as they did during the global financial crisis, which a lot of people find hard to believe. And so nowadays we are seeing much, much more concentrated capital spending in the form of buildouts by the hyperscalers, by the so-called neoclouds, and by the AI companies themselves. So we're seeing a lot more risk in a lot smaller subset of the market. The broader question, of course, is will there be a return on this investment, and I think that question remains to be seen. I pointed out to people, interestingly, the US GDP growth in the 10 years prior to Netscape was exactly the same as it was in the 10 years post-Netscape, from basically 1996 to 2007. And S&P profits, which have a long-term sort of trend growth rate of about 6% a year, S&P profits grew 6% in the decade before Netscape, 6% per annum, and they grew 6% per annum in the decade after Netscape. So for all of the wonderment of the internet, and it certainly changed their lives and it brought forth all kinds of new businesses and killed a lot of old businesses, you wouldn't have really kind of noticed it in the aggregate economic or even arguably financial statistics. And so the question will be, what will AI bring in the form of productivity? What will it bring in the way of enhanced profitability? And how much of that is being front-loaded today? And that's a big question because, like the dot-com boom, we have an accounting identity problem that follows these capex booms. Namely, that the companies that are spending the money do not expense immediately most of that money that's being spent. It's capitalized and depreciated over five to ten years. The companies receiving a lot of that money, the Nvidias of the world, the Caterpillar tractors of the world, the utilities, they're receiving it in terms of revenues and profits immediately. So the same dollar is basically contributing to profits in a far greater extent than it does in a more normalized economy where it would be recognized as revenue by one company and expense by another. And that's what we're seeing, and that's why S&P profits have taken off in the last two years. It's because of this mismatch.
Could you describe more or speak more to this expensing problem? Because I think a lot of the argument as to why we shouldn't be worried about any of this is: one, GDP growth is growing as a result of AI; two, S&P earnings are also growing because of AI, and significantly; and then three, a lot of the financing, a lot of the debt that was kind of the undoing in previous cycles, I mean, it's coming from companies that actually do have significant profits, that actually do have significant cash flows, and that you could make the argument that actually they have the money and the credit to build the amount of data centers that they are building. So what would be your response to people who make that argument, and how does that relate to these expensing problems and perhaps these accounting problems that you bring up?
Well, all three of those were present in 1999 and 2000. As I said, the companies that were taking on the most debt and obligations to build out their networks were, by and large, the largest, most creditworthy companies in the United States. And so that happened then. And it's really an important point to make. However, there was also a mismatch. S&P earnings grew 30% from mid-98 to mid-2000. So they're growing even faster today. But then, as I said, there was a collapse. We had a mild corporate recession. I think GDP dropped 1% and corporate profits dropped 40%. And they dropped because order books collapsed. People realized they didn't need 10,000 routers, they needed 2,000. So they cancelled their order books. But the guys who were building the routers, you know, had expense levels that were predicated on the boom lasting longer and growing further than it actually did. So this mismatch in earnings is substantial right now. It's in the hundreds of billions of dollars a year and is the real reason that corporate profits are growing way above trend. I mean, I think the economy, we agree, is doing fine, which means corporate profits maybe should be growing eight or nine, but they're growing somewhere like 28 or 29. So you get an idea of really just how much of the profitability is going to the chip companies and to the infrastructure companies for this buildout. So then the question becomes, and Scott's a better judge of this, I think, than I am, is what is the ultimate ROI on all this spend? And can we take the tokenized economy and turn it into real productivity gains for Bank of America and Coca-Cola and 3M and my company and Riverside and what have you. And that, I think, is a little bit more nebulous right now in terms of what companies are seeing and relative to their spend.
Jim, it's nice to finally meet you. I've been following your work for what feels like three decades. And yes, I do remember the dot-com era, still nursing those wounds. The economists perfectly called the dot-com implosion. They said it would go from B2C, then to B2B, then the infrastructure guys would be hit. They just laid it out, the implosion perfectly. They laid it out though in 1997, and from that point, the NASDAQ tripled. And the question I would have for you is, do you think we're in '97 or '99? And I think you're going to say we're in '99, but how do you discern the difference between something that's overvalued and things are about to go insane versus we're in the beginning of the end, if you will. Is there a way to tell if we're in '97 or '99?
Well, first of all, I would argue, Scott, that tell me anything in the corporate world or the technology world that hasn't gotten faster since 1997. So in terms of the ability of investors to sort of react to things, in my experience, has been time-compressed dramatically in the last 30 years. So these things get sort of figured out much quicker than they did back when I was shorting iOmega and scouring the Yahoo message boards, you know, to figure out whether it was overvalued or not. So that's number one. Number two, we're not so close to the ignition point as we think. I mean, ChatGPT was fall of 2022. So we're now entering the fifth year of this. It's not the Netscape moment of 1996, I don't believe. And then think of all the VC infrastructure and other sort of ecosystem around technology that exists today to take advantage of these fabulous investment opportunities that really, you know, didn't exist as much in 1996 or 1997. So I would say that there's one other indicator, however, that I think has been foolproof in trying to figure out whether you're closer to a beginning or an end. And that is equity issuance. And equity issuance really didn't start picking up until, you know, late '98, '99 in the dot-com boom, and then crescendoed in the first quarter of 2000. We haven't had a lot of equity issuance. We had a blip in 2021 at one point post-GameStop. SPACs were raising about two and a half to three billion a night, which at the time was equivalent to the US savings rate. And of course we know how that ended. It didn't last long and gave retail investors some indigestion in late '21 and '22. 2026 is an entirely different animal, as SpaceX would indicate. We are now seeing massive equity issuance, and we're going to probably, you know, unless things really cool off a lot faster here, we're going to see probably record amounts of equity issuance in 2026. And so I've always joked that Wall Street has a printing press as well as the Fed. It just takes a while to get it going, and Wall Street's printing press is now going full bore in this year. So I think that there's a fair amount of indicators that are telling us we're closer to a '99-type moment than a '97 moment. But who knows? And by the way, so I do my investment conference every year. We used to have it in Miami in February. And in February 2000, we met and we were about 10 days away from the peak. We didn't know it, but one fellow gave a short and pointed out that it had doubled in 1999. It had doubled in January of 2000. It had doubled again in the first two weeks of February 2000 and did a final double the week of our conference, which was the third week of February. So to one of your points, I mean, when you're in a parabola, you know, you have price risk if not time risk, and as a short seller, I'm well aware of that.
So the market being at near historic highs in terms of valuation, rational argument, AI being the epicenter for driving that, what feels like irrational valuations, also very rational when you look at within the AI ecosystem, assuming that's sort of ground zero for leverage upon leverage or the tail of the whip of overvaluations, and presenting opportunities for a short seller such as yourself or someone who shorts the market. Do you find the ripest targets to be the hyperscalers, the Nvidias, the SpaceXes, or do you find that it's the adjacent and maybe don't have the same cash flows, the same brand equity? Like, what is the white meat or the soft tissue where you think these things really have the potential to go down, you know, 90-plus percent, if you will?
Since 1996, we run our portfolio or advise our clients to be hedged, right? So we're long the market and short a portfolio of 40 what we think are structurally overvalued situations. And in the AI space, we focused really over the course of most of this year in the adjacent companies. So we're in effect long the hyperscalers. We're in effect long Nvidia since we own them through the indices. But where we've been short are the Bitcoin miners who have suddenly become data center companies. The so-called neoclouds that are not estimated to make money until 2030 or beyond, despite a booming market for what they do. And those types of narratives that have sprung up and companies that have sprung up to take advantage of investor capital, where you can't really make the business model work. And it's getting tougher even for the hyperscalers. We could talk about that in a few minutes. They are seeing increasingly lower returns on their invested capital. But there's businesses out there that are doing deals for sort of low single-digit, mid single-digit pre-tax returns on capital who have weighted average cost of capital of 12, 15%. And they're doing it because it's growth. They can announce deals. And similar to the dot-com era, those kinds of companies usually are the ones that end up in the most trouble because they're capital intensive, they're low return, and when sentiment changes and the capital markets tighten up, they can't access capital anymore and they collapse. And so there's just a lot of those out there. But I want to add just one other interesting observation about this. And as much as the AI and the leverage and some of the accounting and corporate structures are questionable in this boom, the overall stock market itself is a lot more expensive than it was in 1999. And by that I mean in 1999 it was the so-called TMT section that you remember, technology, media, telecom, that just had stratospheric valuations, and then there were just lots and lots and lots of companies that traded at 10 and 11 and 12 times earnings. And in fact, value guys did pretty well coming out of the dot-com boom as those stocks were bid up as everything else collapsed. Now we're seeing all kinds of sort of what I would call relatively senior companies or mature companies trading at 40 and 50 times earnings. Take a look at Walmart, take a look at Caterpillar, take a look at a lot of companies that are not squarely in the AI height that are trading at extremely, extremely high valuations. One of our favorites, just as an example, is WD-40. WD-40 has grown its revenues and earnings, I think, something at about a 3% pace over the past couple of decades, right in line with GDP, and it trades at 40 times earnings. And so I mean, there's just lots of those kinds of companies that, because of the trend of passive index investing and the fact that retail and households have the largest share ever of their assets in stocks, means that the broad market is relatively expensive relative to just technology.
We'll be right back after the break. And if you're enjoying the show so far, send it to a friend and please follow us on YouTube, Spotify, or wherever you get your podcasts.
Support for the show comes from VCX, the public ticker for private tech. For generations, American companies have moved the world forward through their ingenuity and determination. And for generations, everyday Americans can be a part of that journey through perhaps the greatest innovation of all, the US stock market. It didn't matter whether you were a factory worker in Detroit or a farmer in Omaha. Anyone can own a piece of the great American companies. But now that's changed. Today, our most innovative companies are staying private rather than going public. The result is that everyday Americans are excluded from investing and getting left further behind while a select few reap all the benefits. Until now. Introducing VCX, the public ticker for private tech, now available wherever you buy stocks. VCX by Fundrise gives everyone the opportunity to invest in the next generation of innovation, including the companies leading the AI revolution, space exploration, defense tech, and more. Visit getvcx.com for more info. That's getvcx.com. Carefully consider the investment material before investing, including objectives, risk, charges, and expenses. This and other information can be found in the fund's prospectus at getvcx.com. This is a paid sponsorship.
We're back with Profy Markets. You mentioned some of those kind of funky new names that are data centers, neoclouds like CoreWeave and NBS and all of these other names. And then there's of course some of the newer memory names like SanDisk, SK Hynix. I mean, a lot of these companies that suddenly just have exploded, and I think a lot of people would agree that there's probably a lot of speculation going on in there and therefore a lot of risk. Is it your view though that big tech, Google, Microsoft, Amazon, Meta, that they are relatively safe and protected right now? Or are they just as at risk as anyone else?
I don't know that they're as at risk given their balance sheets. We know Google just went into negative free cash flow in this last quarter. These companies are still funding most of their build from internally generated cash. So in that respect, they're in much better shape than some of the other companies. But again, that didn't prevent General Electric from going down, you know, 40% in 2001, 2002. And increasingly, the returns on that investment that they are increasingly making are dropping. They're still robust, but they're not what they were. And in fact, you know, we track that pretty closely. And most of the return on incremental invested capital, that is how much additional operating profit are you getting from an additional amount of investment, has been pretty much cut in half for the hyperscalers over the past year and a half. And that's a lot. And companies that were earning 100% incrementally on their capital are now earning 25%. And a couple of the companies are now earning in the teens on their incremental investment. And if that continues to deteriorate over the next 12 to 18 months, I think the C-suite in Silicon Valley for some of the giants are going to be having some interesting conversations about their continued spend. I know that they feel that they can't, quote-unquote, lose the race, but there could come a point, you know, 18 months from now where they're all losing the race. And that gets back to our original question, which is unanswerable, which is what is the ultimate ROI on all this spend? Not for Nvidia or OpenAI or Anthropic, but for you and for me and for, you know, the plumber down the street and corporate America, and what advantages and what increases to productivity and profitability will that bring? And that, I think, is still the 64-trillion-dollar question before the market, because we haven't seen a lot of that yet. We might, and hopefully we will, but that will remain to be seen.
It seems that the breaking point that you outline in your scenario of what could happen over the next 12 to 18 months, the turning point is when the people in the C-suite at big tech have that conversation and say we're not sure that this actually makes sense. Are you surprised that that hasn't happened yet, given the fact that we haven't really seen the ROI yet in AI, that we haven't seen it at the consumer level at least? You know, you point out the return on incremental invested capital, which as you correctly point out is declining for big tech. I mean, why hasn't that conversation happened yet? And do you think that it will happen soon?
Human nature is still human nature, and this is the shiny new object, and people are still enamored with it. And the boards of directors haven't yet started to ask difficult questions because companies are still making their earnings estimates and generally being rewarded or not completely trashed for it. But as I say, if the trends continue that we clearly see over the past two years, it's going to be hard to keep justifying growing your capital spending, you know, 40, 50% a year if your operating income is only growing 10% a year. That will get noticed, and it will get noticed by the market, and it might be already noticed by the market. The hyperscalers are underperforming. You know, the big buyers of compute have underperformed in the past sort of three, four months. So, who knows? We'll have to see.
That has been a fascinating question to me, is to what extent are the points that you're making already priced in? And in a lot of ways, it seems that there are certain corners of the market where it is priced in, certain corners of the market where it doesn't exist whatsoever. And I can't quite understand what to make of that, and more importantly, what to do about that from an investing perspective. What is your view on that?
Yeah, I mean, I would agree with you. You know, there's still these buckets of speculative capital that keep sloshing around to keep looking for the hot new area. So if we give up on hyperscalers, we buy neoclouds. And if we give up on neoclouds, we buy chip companies and memory companies. And if, you know, we tire of those, we buy something, we buy Apple, which is hitting a new high, who's not spending any money. So it's as old as stock markets, right? I mean, people just keep looking for the hottest area. But I would think that if overall the AI spend begins to get pulled back because of lower returns, I think the whole ecosystem, much like the whole ecosystem really prospered in 2024 and 2025, everything went up. Now we're starting to see some discernment, and I think that'll continue if the ROICs continue to drop.
