"Look, the emperor has no clothes"! Hedge fund manager: Can the trillion-dollar investment in AI be recouped?

On October 14th, the podcast Rebel Capitalist released its latest interview, inviting hedge fund manager Kuppy to discuss the fatal flaws of the AI boom.

Harris Kupperman (nicknamed "Kuppy") is an experienced hedge fund manager and investment blogger, currently the founder and Chief Investment Officer of Praetorian Capital Management LLC. From 2019 to 2021, the fund achieved a net return of up to 711%. He also runs a popular investment blog, "Kuppy’s Korner," where he shares his investment philosophy and market insights.

In the program, Kuppy calculated that there is a huge gap in the return on investment for AI data center construction. Over the next 3-5 years, AI data center construction will require investments on the order of trillions of dollars (for comparison, the annual budget of the U.S. Department of Defense is only $1 trillion), and to achieve a 10% capital return, revenues of $1-2 trillion are needed; to obtain good returns, revenues of even $3-4 trillion are required.

Kuppy also pointed out the flaws in AI business models, stating that AI products like ChatGPT/Gemini/Claude are highly interchangeable, with users showing no loyalty (as the free versions are sufficient); price wars will drop to just above energy costs by 1%, leaving no profit margin. LLMs are advancing rapidly, and free versions will always be "usable," with no one willing to pay; even if AI doctors/tax advisors emerge, there will be 10 free options available.

Regarding the recent hot discussions on AI cyclical trading, Kuppy believes that it is similar to the telecom bubble of 2000 (where Lucent/Nortel financed through suppliers/leasing to create fake revenues). Currently, companies like Meta/Microsoft are using equity/leasing data centers, effectively serving as a cover for PE/VC, creating "safe" assets.

Summary of Key Points:

• If you want to achieve about a 10% capital return, this seems a bit high for some speculative, rapidly depreciating capital equipment. For example, I don't know, you need about $2 trillion in revenue. If you want to achieve good returns, you need $3 trillion or $4 trillion in revenue. I mean, these numbers are enormous. Keep in mind that the Department of Defense spends up to $1 trillion annually, which includes aircraft carriers and so on;
• The AI commodity component does not prevent other companies or individuals from entering the market. OpenAI charges you $90, we charge you $50. Then OpenAI says, "Okay, we charge you $40." Then Claude says, "Okay, we charge you $30." And so on, you compete all the way down until costs are minimized, ultimately possibly just 1% above energy costs. And that 1% is the profit for OpenAI or Anthropic
• Investing in AI infrastructure is like building a railway; those who invest in this railway experience countless failures at various stages of the capital cycle, including debt, equity, and so on. I believe artificial intelligence will be the same. They will continuously invest more and more money, and then go bankrupt time and again. Perhaps, you know, Microsoft or Meta, they won't go bankrupt, but their level of bankruptcy is equivalent to, you know, writing off a $500 billion investment and starting over.
• It is called "cyclical investment" because all the numbers are fake. The same billion dollars just keeps circulating. The purpose is to lure Wall Street into investing more money. Eventually, Wall Street sees through the scam, but before that, retirees lose hundreds of billions of dollars, just like this cycle, except this time the losses will reach trillions of dollars, and they all go bankrupt.
• I have spoken with two people in the private equity investment field, one responsible for credit and the other for equity. I mean, these people told me, "We have these things worth tens of billions of dollars, but we think they are worthless."
• To build infrastructure for data centers, some companies are buying land next to natural gas facilities, hoping to sell it to companies like Google, Microsoft, or OpenAI. This reminds me of 2006 and 2007 when everyone was buying single-family homes and flipping them for profit because they knew housing prices would always rise. But now we are no longer flipping houses; instead, we are reselling land for building data centers.

The following is the original interview:

George:

Hello, friends of Rebel Capitalist! I hope you are all doing well. I'm here with my good friend Kuppy, and we have a lot to discuss. First, Kuppy, please tell us why you play an important role in the video?

Kuppy:

Well, let me explain the long-winded arguments people are making. Every dollar in a 401k account is used to buy Nvidia. My dad's 8% of his retirement fund might have been spent on Nvidia, and he doesn't really understand what Nvidia is. But you see, they, I mean, this is a cheat code. You see, every dollar you earn, every dollar you save, will be used to buy Nvidia. By 2030, about 150% of global electricity demand will be used for AI chips, which will create cat meme templates, and then we will all post them on Instagram in holographic form. Yes, that's the whole story; is everyone doing AI?

George:

Alright, goodbye. Great roots. Okay. Now, jokes aside, brother, you've published a few blog posts recently that I think are fantastic. One thing that frustrates me is that many of my friends are not actually investors; they come from different industries, and they send me podcasts like Jensen's, you know, the so-called rebuttals to these arguments, saying this is a cyclical AI bubble, whatever you call it, but they never provide specific numbers. For example, they say, "Well, you're investing in this, and this is investing in that," but they never break down the actual math, right? Where does the income come from? How much income do you expect? What are your costs?

Jensen's argument is always, oh my god, I invested in this company, I should invest more. What does that even mean? What I like about your blog post is that you really take out a pencil and notepad and say, hey, let me do some math. So, let's take a look back at the origins of the first article. And I really want to hear how many people in the AI field have contacted you before you wrote the second article.

Kuppy:

So, if anyone hasn't read my blog posts, you can check out these two posts at precap.com. Although it's free, everyone keeps telling me how great AI is, and I'm not smart enough to argue with them.

Look, I'm using AI. I think it's great. You know, it makes me more efficient. However, it makes mistakes most of the time. It's a bit like a first-year analyst, but you know, it can do your first-year analyst work faster and for free. So I'm not against AI, but I started to think, if this is a great investment, then what should the numbers be? We will invest about $400 billion in data center construction this year, which is a huge number.

If you assume the depreciation curve for these devices is 3 or 4 years, I mean, data centers themselves are not durable; they just keep changing technology, and even the buildings themselves are almost worthless.

It's like I've heard about some data centers that are now shut down. They operated for three years, the technology changed, and the cooling systems became obsolete.

But if you think these things can last three to five years and assume a 25% gross margin, that's almost a fictional number. Who knows? You need about a billion dollars. I mean, for these data centers, you need a trillion dollars. For these, about $500 billion is needed. To break even, if you want a 10% return on capital, you need about a trillion dollars. This is not referring to the future of all AI, but to the data centers that will be built this year. Next year, they will build more data centers. Next year, you need $75 billion in revenue. I mean, you're looking at a trillion dollars in revenue. And you know, this is before paying a billion dollars in financing costs or any other expenses; this is the revenue you need to build these things, right? Like I said, if you want about a 10% return on capital, that seems a bit high for some speculative, fast-depreciating capital equipment, like, I don't know, you need about $2 trillion in revenue. If you want a good return, you need $3 trillion or $4 trillion in revenue.

I mean, these numbers are enormous. Keep in mind that the Department of Defense spends up to a trillion dollars a year, including aircraft carriers and so on. We can't shift everything to AI. I don't know where this money comes from. For example, the entire AI industry is expected to generate about $15 billion to $20 billion in revenue this year. Remember, you need $500 billion to break even Yes, so you need 30 times the growth. I don't know how you do that. It doesn't make any sense at all. The things these people are building have no chance of being profitable. I think those venture capitalists see this model. I mean, they did this model with Uber, and then they subscribed to it. They spent about ten years scaling it. I mean, that's great, right? You can go to any city you want for just $4. You know? It's like they got used to subsidizing this thing, and it's scaled so much that they'll go bankrupt before they reach profitability because in the end, no one is paying for artificial intelligence.

George:

Yes, one argument I've heard, and I think it makes a lot of sense, is that they can only make $12 billion a year right now because they're just doing B2C. But the real goal is to integrate AI into everyone's life so that people get used to using it. They will feel like they must have it. Just like with smartphones. Then they will start selling it in the B2B market. When they sell it in the B2B market, it makes sense for all these companies. Let's take Microsoft as an example; suppose they spend $50 billion a year on hiring employees, or say, on manpower. They could save that expense and just use AI. Then that $50 billion in revenue could flow to OpenAI.

I heard Jensen give an example; he said, you know, the global economy is $120 trillion. He said 50% of that is the wages of people working with computers, that is, non-manual laborers. And AI will replace all those jobs. Roughly speaking, that $50 trillion in revenue could flow to OpenAI. This revenue would be used to pay for all the data centers. So, what do you think of this counterargument? I believe you've heard it before.

Kuppy:

So I don't think it's $50 billion, nor is it $50 trillion. I think it's about $5 trillion. If you take away those people who work with computers, the highest earners, where does the income come from? I don't know how that works, right? Then there would be no income because all consumers, remember, 10% of the population accounts for half of the consumption. The people you're talking about have no jobs. So where does consumption come from? Where does income come from? It's a cycle. Even if these people win, they can't win.

You see, I completely understand. In ten years, I will have an AI doctor. I will use AI to handle tax issues. I will use AI to submit my ADV form to the SEC. But if I use AI today and the ADV form is filled out incorrectly, I will get into big trouble. We all know that AI has its good sides and bad sides. I don't want to bet my career on whether they comply with the requirements. I want to hire a very professional, experienced person to do this, and they need to buy a lot of liability insurance so that if they make a mistake, we can sue each other You know, I'm so happy running a hedge fund that I really don't want to testify at the SEC. I mean, we're not at that point yet, but it will come sooner or later. But if we spend $500 billion this year, it could be $6.7 trillion next year, or even $10 trillion. I've heard that by the end of this century, they will spend $5 trillion, which is just unbelievable, right? Because if you consider inflation, that's equivalent to spending seven times what we spend on interest highways. Highways are cool, useful, and AI can make cat memes. Well, I think we will achieve this, but who will spend $5 trillion to make it happen? What is the return after achieving it? I'm an old-school baby boomer, the kind that has money coming in and out. I got venture capital, you cash out, and it doesn't come back for years or even a decade, but at some point, the money has to come back. The longer the delay, the better the returns for the entire cycle.

As we said before, there are no returns. Even if you invest $5 trillion, maybe this year's funds grow from $15 billion to $50 billion, you still have no returns.

I mean, that's not profit; $50 billion is revenue, right? Because just like LLMs (large language models), they are excellent. The free version will always be good enough, especially if the free version can improve by 25% or 50% every year from now on. Who would pay for that? No one would pay for these. Yes, I think we even have things like doctor AI. There will be 10 different AIs. Yes, they will all be free.

George:

Yes, I think that's my biggest concern, and it's my biggest objection to the AI supergiant Hyper Bowls. What you're dealing with is something like a commodity; no one will have an affinity for OpenAI or ChatGPT, just like no one will have an affinity for Gemini. If Gemini releases a better version, you'll use Gemini. Or if Claude releases a better version, you'll switch to Claude. There are no network effects. It's like all your friends are on Facebook, so you have to be on Facebook. Or all the hedge fund managers you know are on Twitter, so you have to be on Twitter. AI doesn't have that.

Microsoft does have Azure products. I don't know if I'm pronouncing it right; you might know what I'm talking about. Azure is their backend, and they can leverage cloud computing and so on to create various upsells. So once a company starts using Azure, they become integrated into the ecosystem, making it hard to leave. I get it. So their approach is to introduce AI as a tool to ensure customer retention and prevent customer churn. But unfortunately, introducing this tool comes with costs, which will lower profit margins Kuppy:

I believe the argument here is that Microsoft could fire all its employees, and then they would... assuming they give each employee $100,000. So now they can suddenly offer OpenAI $99,000 or $90,000, which seems reasonable. But my point is that this commodity aspect does not prevent other companies or individuals from coming in and saying, well, OpenAI charges you $90, we’ll charge you $50. Then OpenAI says, okay, we’ll charge you $40. Then Claude says, okay, we’ll charge you $30. And so on, you compete all the way down until the cost is at least 1% higher than energy costs. And that 1% is the profit for OpenAI or Anthropic. So the issue for suppliers is that it is a commodity. What are your insights on this? Or do you think I misunderstood?

This situation will continue until someone refuses to fund it. That’s the answer. You have many participants, and it’s almost a matter of life and death. You have some super-large companies with real cash flow and profits, and strong borrowing capacity. You have some private equity firms that once you put in a penny, you know you’re making a lot of money, especially since you’ve marked this model for 10 years. So they only earn commissions, and no one even knows the value of these things. You also have some well-funded venture capital firms. I think the best analogy is the transcontinental railroad. We built some of these railroads, and it doesn’t really matter which route you take.

You know, there are better routes and less good routes. But the trains have to run, and the cost of running those trains and maintaining the tracks is fixed. So they operate these projects at a loss because a small loss is better than a big loss, and they are all competing for cargo. But there really isn’t enough cargo, and they have all gone bankrupt many times. And you know, those tracks, those lines are still in operation today. I mean, 150 years later, they have been refurbished. We are still using them.

It’s like we built the transcontinental railroad, which is great, but all the investors in that railroad, at various stages of the capital cycle, including debt, equity, etc., have experienced countless failures. I think AI will be the same. They will keep pouring in more and more money and go bankrupt time and time again. Maybe, you know, Microsoft or Meta, they won’t go bankrupt, but their level of bankruptcy is equivalent to, you know, they will write off $500 billion in investments and start over.

I mean, these people are a bit cunning, right? Just like those circular trades you mentioned. I don’t think anyone really plans to invest $100 trillion in these kinds of trades. I mean, they are just doing it to release press releases. This is actually a cover for private equity firms to get funding and build data centers, and you have to use the data centers. So this is a $10 billion data center. Maybe, you know, Meta or Microsoft, they participate with $500 million or $1 billion in equity. So it looks like, oh, they are taking on the risky part, which is very safe Then they signed the contract, and we will rent this data center for 15 years. Oh, it really looks safe, you know, because Microsoft is unlikely to default like we've never seen before; they charge or pay the contract, you know, sometimes it's cumulative. I mean, this situation happens every five years.

But, uh, in this case, you will encounter some private equity brothers who will buy 9% of the subordinated debt. You should know that if it's 9% debt, there must be a problem, right? Then you will see companies like Pacific Investment Management Company (Pimco), which means my dad's retirement account is buying super safe senior debt at a 7% interest rate, which should indicate that there is a problem, right? Because I mean, safe debt is U.S. Treasury bonds with a 4% interest rate. I think its performance will be very similar to CDOs; I mean, people take these CDOs, and basically, it's the square of the CEO. They take the worst parts of the CEO and put them into a new structure they call CDO square. All its advantages are AAA. We learned that if you mix raisins and... you basically get... and these things, the results are terrible. I mean, their loss rates are as high as 90%.

So, what happens when you have a data center, technology changes, and five years later, the cooling system needs to be installed underground instead of above ground? Oops, you didn't build the basement; this situation is common in data centers because technology is constantly evolving.

Look, I believe the people operating Meta and Microsoft are all good people, but the lease term is 12 years; do you think they will continue to pay? Or will they find an excuse to say in some rogue court, we want to default, and then you fight each other in court for five years, and you might get back two-thirds or half of the money. After deducting legal fees, the amount you receive will be much less.

But that's always the case. I mean, we've seen this in the "pay as agreed" contracts for pipelines. It's as if no one wants a pipeline to be filled for 20 years. If there are problems with the oil field, you can always find a way out. It's the same here.

So you are basically financializing these things. And companies like Microsoft, I mean, look, they are not getting rich through stupidity. They are finding ways to make up for these shortcomings. This includes companies like Core Weave, which is like a huge stock market.

I mean, look, hey, today's press release says the stock price went up that day because insiders only stole $2 billion today. You know, normal team operations, they usually sell $3 billion to $4 billion a day. Today, I don't know, maybe they went to buy yachts or something in the afternoon, you know, forgot to place an order for the new seller.

$2 billion is indeed bullish, just like insiders not selling fast enough, which should indicate their view on this asset. Yes, you can announce transactions worth hundreds of billions of dollars among ten companies, and these companies are all doing the same thing. I don't know if I want to, I don't know if I want to say no, but I really have my doubts, right? Just like when I was a kid in 2000, I saw what was going on. Companies like Lucent and Nortel, you know, they are now NVIDIA. They would engage in vendor financing with customers who were telecom companies that could no longer raise funds. The capital markets had, to some extent, shut them out. So they would announce these huge deals, you know, we are going to lend this company billions of dollars, trying to make Wall Street feel, oh, this must be safe, you know, if Nortel is willing to lend money.

Then Nortel would say, we want to lease the spectrum and fiber capacity of your telecom system. What’s the use of that? Just to create revenue? I mean, why does Nortel need the capacity of the telecom system? What they were doing was creating fake revenue and then directly returning it to Nortel to pay the interest on the debt Nortel had lent, right? Then, when you get equity in return, you can inflate the equity value and make a profit like Nortel did with those tricks. So, the incentive mechanism is like, okay, let’s add a zero to all these numbers. Well, I’m going to lease this capacity at a price far above the market. Do we know how to handle market prices? Does anyone know what the market price should be? We won’t lease at a price far above the market. You would use this money to acquire us, to buy more from us. We’ll give you a billion dollars, and you’ll give us a billion dollars. We’ll do this over and over again.

It’s called “circular investment” because all the numbers are fake. The same billion dollars just keeps circulating. The goal is to lure Wall Street into investing more money. Eventually, Wall Street saw through the scam. But before that, retirees lost hundreds of billions of dollars, just like this cycle, except this time the losses will reach trillions of dollars. You see, they all went bankrupt. You all went to prison. It’s all a scam. I mean, people like Adelphia might still be in prison.

George:

Yes, yes. And, you know, it might not repeat, but it’s definitely Ryan’s at Ryan. It seems like, okay, we used free cash flow to finance all capital expenditures before, and now they seem to be financing again with debt and equity. At some point, this runway runs out. It’s like, can you get the expected profits, trillions of dollars in profits, from our current $1.2 billion in revenue? Can you timely complete the transition from financing with free cash flow to debt financing and then back to free cash flow financing?

I think that’s the real issue. And I think the likelihood of that is extremely low. Some giants like Microsoft and Google seem to be pouring billions into this or the metaverse space because they have no choice. It’s not because they think it’s profitable. It’s because I have to do this, or I will lose market share. Can you feel that atmosphere behind the scenes?

Kuppy:

I think it really is a matter of life and death. Actually, I think you’re right. I think a lot of people, after seeing it, will say, we don’t want to fall behind this new technology; we don’t really understand what it is If we fall behind, what do we have left? I mean, you can look at those big companies that spend the most money. Well, Google, I haven't used Google Search in six months, right? Do they have their own AI search? Well, no ads. Google Search, they do too much. You ask it a question, like what's the weather like tomorrow? It just sends you 100 things to click on with banner ads. It's so annoying, they have nothing. Search is their business. Waymo is cool, but it doesn't make money. So they have no choice. In 10 years or 10 months, their business could disappear. Meta, like who would still use Facebook? Instagram lost to TikTok. What will Zuckerberg have in five years? It's a matter of life and death. Yes. Today's cash glue. He might as well be rolling dice.

George:

It's like they have no choice but to keep investing. Just like retail investors over the past 15 years, they've been pushed out of the risk curve.

Kuppy:

But look at who isn't spending money. You know, Amazon isn't spending money. It doesn't matter to them. You know, Apple isn't spending money. It really doesn't matter to them. Yes, I think so.

George:

Bro, there's something interesting. A few days ago, I was researching pets.com because some feedback I received was that this is completely different from the internet bubble. Look at pets.com, it didn't even have revenue, any sane person wouldn't invest in that company. It's completely unreasonable. So I asked ChatGPT, hey, can you provide some bullish arguments and sources for pets.com? If you read the bullish arguments from 1998, you'd find them very reasonable, very logical, very sensible. I thought, hey, yes, that sounds good. Where should I buy? But interestingly, do you know who the biggest initial investor in pets.com was? Awesome. Bezos. It was Bezos from Amazon. So the point is, I think you say Bezos was a bit hands-off on this because he was the biggest investor in pets.com, which is quite interesting. So you say, maybe.

Kuppy:

Incredible. I mean, like you said before, these things often rhyme. I think it's very important for investors. Yes, I remember how severe the bubble was in 2000. Some friends of our family, you know, not super wealthy, but around 50 years old. They could have retired. But instead of managing their portfolios as usual, they put all their money into tech stocks, and some people in their 70s are still working, as if they went bankrupt because of margin debt or something.

These cycles are truly incredible. I never thought I would experience such a cycle again in my lifetime. I thought they would only happen once every 100 years. But 25 years later, we went through it again because they printed too much money, and some people, especially in the venture capital world, did too many stupid things and made too much money. They made so much money just by sheer willpower, putting funds into projects they thought would eventually have a business plan as long as they invested enough money I don't know. I don't know. I still don't know how they plan to monetize this money. It's really unbelievable.

George:

Yes, that's the key here. So now I want to discuss something that's slightly off-topic because I know the market was crashing before our live broadcast. I don't know if that's still the case.

Kuppy:

It's still crashing. Is it really? I would continue to avoid it like crazy.

George:

Yes, but this is also very timely because we have the AI bubble, and people need to read your research on this issue, read the math. Then they also need to hear the stories from those insiders, and we actually contacted them by phone.

Kuppy:

I find this very interesting. Actually, let's pause for a moment and talk about this topic. How cool. I wrote this blog post, and I have quite a few fans, although not as many as yours, but still quite a few. Then someone suddenly came to me and said, "Kuppy, I agree with your point of view." I was glad to hear you say that. I thought it was crazy at the time. Every day when I get home from work, I ask myself, how is my company going to make money?

George:

These are all insiders. I want to make it clear that these are all insiders. They are not ordinary people like me who casually say on Twitter, "Oh, man, I'm glad you said that because I think so too." They are actually people currently working in the field of artificial intelligence.

Kuppy:

They are like friends who are pre-selected. I wouldn't talk to some junior employees because I only want to talk to those who know what they are talking about. I've talked to someone who has a lot of data centers. In fact, he is the CEO of a data center company, which is interesting. I've spoken with two people in the private equity investment field, one responsible for credit and the other for equity. I mean, these people told me that we have these things worth billions of dollars, but we think they are worthless. I've talked to someone who works in cooling technology. You know, I've talked to people in chip design, and I've also spoken with several venture capitalists, one of whom said, yes, we just completed another round of financing, you know, at a valuation that is tens of billions of dollars higher than our last round. Our revenue is about $1 million. I've been telling my board, hundreds of billions of dollars. Yes, he said, I've been telling my board, let's get this right and then finish this. It feels not worth it. Then the board said, the valuation doubles every six months. Why should we sell now? He said, you don't understand. I don't understand what I'm doing.

All these high-level people say they are surprised by those who truly believe this is the future. They are only focused on making money but feel that the numbers simply don't add up. We are facing negative growth, and the more money we invest, the more we lose. Typically, these types of companies eventually achieve some economies of scale. I've talked to about twenty people; I don't want to, you know, doctors, anyone, so I have to be careful with my words. But they almost unanimously think this is crazy, and they completely don't understand the reasoning behind it. I think the real meaning of my blog post is that it's a very interesting expression, like everyone knows, but no one knows that everyone knows George:

It's just like the emperor has no clothes. It really is.

Kuppy:

It's like, I ask these people, why, why don't you talk like this, like, I can't tell the board, they would fire me. They think I lack vision. I can't tell the employees that they all need to resign and go to a company that is passionate about this. I can't talk to people like these. It's almost like a crying session. They just call me because they have no one to confide in. They are really scared that if they say anything, it will get back to the people funding these things, and then they will be unemployed. Right? It's so strange. But it also makes me believe that my numbers are not wrong. You know, there was once a person with a lot of AI investments in a hedge fund who said the same thing as you, that many high-paying jobs, like doctors, lawyers, accountants, will be taken away. I think he is right, and I think so too.

George:

By the way, I just think this will be offset by energy costs and the costs brought about by the global increase in the use of AI?

Kuppy:

And I think he was also 10 years early. When you discount that money back to current cash flows, it’s worth nothing. But I understand what he means. This is the bull market rhetoric. No one can tell me you don’t understand. Next year it’s $100 billion, the year after that it’s $500 billion. I mean, I ask these people, and they say, yes, revenues might double, but even with such revenues, it might grow from $1.52 trillion to $3.04 trillion next year, a lot of which is actually back-and-forth revenue, right? You know, Microsoft buys things from Core Weave, and Core Weave also buys things from this company. How much of that is actual net revenue? After all, these are all private companies. I won’t report too much to you. If the numbers were really good, they would definitely release a pressure statement on it.

George:

Yeah, you know, one thing I’ve noticed is that you always have to pay more attention to people’s behavior rather than their actual words. Altman publicly states, of course, you know, he’s been touting everything, saying how great it is. But then you look at his actions, he’s making a lot of investments in other companies, so I think, well, if he’s so optimistic about his own company, why doesn’t he reinvest that money into OpenAI? Why is he taking that money to investigate companies that might not go bankrupt, might not commercialize, or might not have commercialized products? So, that’s something to be wary of. Another thing, I want to tell you, some anecdotes I heard last weekend on the golf course, I’m a member of a club called Madeleine. So you always see the green lane, tourists come here to vacation and play golf on weekends. I had previously chatted with a guy. You can tell he’s, let’s say, a retail investor, a bit scared of missing out (FOMO), likes to buy cryptocurrencies and the like. But he has a lot of money. He said what he’s doing now. As his career is booming, he and his partners or others have created a limited liability company, and they are buying land next to gas facilities and building infrastructure for data centers, hoping to sell it to companies like Google, Microsoft, or OpenAI I just think this reminds me of 2006 and 2007, when everyone was buying single-family homes and flipping them for profit because they knew housing prices would always go up. This reminds me of the same thing, but instead of flipping houses, we are suddenly flipping land for building data centers.

Kuppy:

Everything is so crazy. About six weeks ago, I attended a wedding and saw uncles and aunts I hadn't seen in five years. I always feel like a financier, and they always want to give me stocks.

Kuppy:

Yes, yes, I want to say that the current 60%/40% portfolio is 60% Nvidia and 40% Bitcoin. That's right, that's right. Some people like the 3x leveraged version. These people, like my cousin, have seen their stock prices rise 150% this year, and I feel sad about it because it's for a professional life. My performance this year has been good, but not at the 105% level. Yes, they are more like replicas. You can buy call options, and with these 3x leveraged things, your profits will multiply by 6. You know, yes, I've seen this before. It's like I don't know the ending of the story; I would feel it would last longer. They would make more money.

I really like many of the "emy" you see today, which are those who start spending after making profits in the stock market. Yes, I completely agree. The vast majority of people have been crushed by inflation. But if you hold 6x leveraged Nvidia, you will go out and splurge, go on vacation, do whatever you want. This is basically the last factor driving emy because as long as the video shows a two-week decline, all these people will be margin called. It's like, I've witnessed this, and it's also the reason for the market crash. I mean, we've never experienced a situation where everyone holds 3x leveraged stocks and they drop 10%. I mean, these stocks require margin calls, which is scary, right? You know.

George:

Another ironic thing is that I think Nvidia's ability to become a $4 trillion company is largely due to passive fund inflows. Inflows into SMP. This is obviously a weighted index, so for every dollar that flows in, they get 8 cents. But if AI is so great, if AI takes away all the jobs, then passive fund inflows will turn into passive fund outflows. Now suddenly, for every dollar that flows out, 8 cents come from Nvidia. So, if you do this, if you...

Kuppy:

Don't do that, I think this situation is very common. And look, my view of the world is very simple. I don't know, you know, no, look, I think this AI will actually take away a lot of jobs. I think it starts slowly but will continue. This won't be, you know, an immediate thing. It's more likely that people will say maybe we don't have to hire, let's see how AI performs a year from now. You know, our business will grow. Let's have everyone work overtime. Because if you just train them and then fire them, why hire people? I think you just have a very inflationary cluster George:

In that environment, you face the pressure of deflation, especially when it erodes the job market, unemployment rises, passive capital inflows turn into passive capital outflows, the market declines, and a bull market arrives, all caused by the spending of a K-shaped economy.

Kuppy:

Inflation is inevitable. But I think, especially with someone like Trump, looking for someone who will make plans and pretend to be a free market person, he will definitely like his central planning. I think they will take controlling measures and strange fiscal policies, and you will see all sorts of strange adjustments everywhere. Trade and investment in this world will become very difficult. But I also think you might see a world like the Great Depression, where in the 1930s, the living standards of ordinary people fell by 15% to 25%. Why wouldn't we repeat that? Unless it happens during a period of inflation. This is not the deflation that everyone fears. It looks more like Argentina. I think this is the worst-case scenario because there won't be many good investments in that world.

George:

I think the situation in the U.S. now is similar to what it was in 2021 and 2022, or similar to the situation in the 1940s. I always compare it because I think if the government stands by and does nothing, the economic fundamentals will lead to deflation, or even complete deflation. But if the government intervenes, that's a good example. Like I said, if the government had been hands-off in 2020 and 2021, we might have seen a 15% deflation. I mean, the entire economic system would completely collapse in the context of deflation.

But the government intervened in your view, distorting the economy with "stemmies" (referring to economic stimulus plans). You experienced all these supply shocks, and the end result was that CPI rose by 9%. So, I think this is very similar to the situation in the 1940s, you know, when price controls were implemented and so on. So I agree. I think the economy wants deflation, but the government will intervene, taking these crazy measures that lead to a temporary spike in inflation, very similar to the situation in the 1940s. I reminded everyone that the CPI back then was 19%, and two years later it became -2%. It's nothing. This kind of extreme rollercoaster fluctuation looks like we are about to experience something extremely similar. But you talked about Trump, you can like him, hate him, and so on. He is definitely an unconventional person. So you posted again. I don't know if it was a blog post or something you posted on Twitter, where you talked about your daily work as a hedge fund manager, what you were trying to do during the Trump era, and how it differs from the Biden or Trump 1.0 era. Can you give us the Reader's Digest version?

Kuppy:

Yes, of course. I mean, you see, I kind of miss the Biden days. He would wake up at noon, they would dress him in a suit, he would shake hands with a cloud, and then walk off the stage as if nothing had happened, right? Yes, and then Trump, he is just incredible, right? He is full of energy, like throwing bombs at everyone But he is not like some female politicians who leak some information to The Washington Post, and then they see the results of public opinion polls elsewhere. Sometimes Trump just presses a button on his tweet and then says, "I'm going to play golf." The whole world is asking what this tweet means? It's like, look, I'm talking to you right now, and the stock market is down 2% because of this tweet.

George:

Oh, that happens. Yes, I know you guys have been watching, by the way, because just before we went live, as we mentioned earlier, something happened, and the market crashed. I don't know, Kuppy said it was caused by something Trump said. So, think about it.

Kuppy:

Yes, you see, all of this can change in 10 minutes, but we are doing pretty well today. We are shorting the global market, going long on volatility, going long on yen, and going long on gold.

Well, I think we tweeted again. We've been talking for 20 minutes, but just like in Trump's world, you have to think differently. You can't use maximum leverage. We can't be fully exposed to risk. You see, we don't have any extra cash, so buy some crash put options, and then you want to hedge it. Trump always does the same thing. Shocking, and then he steps back a little. Right. So you better have a market maker's mindset rather than a trend-following mindset. Makes sense. You know, I have to adjust the way I deal with the market. We all should. But it also creates a manic frustration; you never know what he will do next.

I remember before I could go to meetings, you know, I used to have lunch with some friends and talk about the market. Now I can't. I have to stay here and watch the tweets, you know, you never know when he will tweet at you. I check my phone every three minutes. Just like you said, I probably wake up every 30 minutes at night because these are the targets we are going to impose tariffs on today. You know, then you stop and you're like, okay, I mentioned it, you know, we have no contact with this person, you know, right. For someone like me who is very risk-averse, this is really difficult and frustrating. I try to stay disciplined. But it's harder because you really don't know who is immune and who is safe.

A lot of what he does doesn't really have much logic. I think it's a mosaic format, just like I understand the goals he wants to achieve. But a lot of times, this strange collateral damage is, you know, something happens, and then, you know, it's not something my company intended to do, it's just, you know, I like to joke, you know, having lived in Mongolia and running a business there for over a decade. I like to joke that, you know, my biggest fear has always been the government passing a law to confiscate all the assets of someone named Kit. Any last name starting with K, our assets would be confiscated. Then, as the dream continues, one of my Mongolian friends comes to me from parliament and says, Kuppy, we are sorry. We are going after Mr. Kim, that Korean. You are just a version of K. Oh, if you don't say anything to the government and just act a little friendly, we will return your assets to you in six months This is how Mongolia operates, and it's also how Trump operates.

The unique dictatorial rule and mentality of the Mongolians make investing difficult. Yes, I believe that due to the bizarre volatility based on macroeconomic outcomes, stock price-to-earnings ratios should be lower. You say our valuations have reached historical highs. I think we should be more like a banana republic because we increasingly feel like a banana republic. However, that's how I think.

George:

So, as a fund manager, what would you do? I mean, you don't have to detail all your strategies and so on, but what are some simple things you're doing now that are different compared to the Biden administration?

Kuppy:

First of all, I have a negative view of the world economy. I want to invest in things that do not rely on government revenue, government support, or say, things that the government does right in terms of permits and such. You know, I want to stay away from government-targeted areas. I want to invest in industries that do not affect GDP. I mean, that's really important to me. I want to invest in industries that won't be affected even if the economy performs poorly (which I also expect). You know, economic growth usually benefits businesses like mine more, but I don't want to be affected by that. My portfolio composition is very different. Lastly, my risk exposure is smaller. I used to run 1 minute 15 seconds, 1 minute 25 seconds, with almost no short-term runs. Now, I mean, I run 80 to 90 seconds, and I'm flexible. If a crash happens, I will take crash put options and some other measures, just like it has changed the way I handle things.

George:

Yes, that makes sense. You know, I'm completely an amateur in this regard. But one thing I've done that has worked quite well, and I would love to hear your opinion on it, is that I haven't built positions directly but have tried to think about the spread. For example, at that time, the spread between the 2-year and 10-year was about 40 to 45 basis points. Now I look at a chart of the 2-year and 10-year spread, going back to other points of curve inversion and inversion easing, you know, usually, things explode at that time. But typically, you get a larger bowl-shaped curve because the Federal Reserve will quickly cut rates in response. Then, when they lower rates to 0 or other levels, you get a bear market curve. So, from inversion to a spread of 150 basis points, then to 250 or 300 basis points. I thought, this makes sense. The best part is, if we really fall into a massive recession, you might get a bowl-shaped curve, in which case you win. If you don't fall into a recession, if you're completely wrong, and the economy takes off completely, with inflation expectations and growth expectations rising, you might get a steeper bear market curve. So, the only thing you lose is a flat curve, and the probability might be slightly lower than the other two outcomes. So I've been trying to do that instead of just following the 10-year interest rate trend directly or shorting the 10-year interest rate directly Kuppy:

Yes, what I mean is, you see, I've never really played those spreads. I feel embarrassed to admit that. I’ve been long on 10-year and 30-year bonds. I really hate bonds. I know they are worthless. I just feel like they might go through another Pavlovian "Pavlovian" response. Oh, we are experiencing an economic recession. For example, let’s buy some bonds. You see, I go on Twitter, and not a single person is long on bonds. The 60/40 portfolio is no longer viable. I mean, the 40-year bond is really like Bitcoin, or if you are a baby boomer, it’s gold. No one holds bonds anymore. Everyone says nothing can stop this train unless it falls into a ditch. Just like emy is really bad. I don’t know. I feel like a bondholder. It makes me sick, but I don’t know much about bonds, and I know where to stop loss if it doesn’t work. So it’s a bit crazy...

George:

Let me give you a few examples that might make you feel better about the bond position. I’ve researched this as much as I can to see if I really have an edge because I’ve done whiteboard videos on this. You know, I’ve done so many whiteboard videos, I can’t even remember how many, 1,000, 2,000. So, to do a whiteboard, you have to do some research.

I think bonds are curve-related because they are usually tied to growth and inflation expectations, not supply. The reason holding bonds makes you feel sick is that you see supply increasing, and that is definitely going to happen. In my view, the deficit will exceed the debt in three to five years, reaching $50 trillion. Absolutely.

The reason this was irrelevant in the past is due to the mechanism of bank balance sheets, because banks ultimately, especially under the euro-dollar system, are still marginal buyers. So, if you hold a 10-year U.S. Treasury bond, assuming its trading rate reaches 8%, and the U.S. nominal growth rate or nominal GDP expectation (let’s say 4.4-5%), that’s a huge opportunity for them because their financing cost is 2%, and they can earn an 8% yield from the Treasury, thus hedging the maturity risk. So, it’s a bit like an equalizer. Therefore, if you look at the chart of nominal GDP and the 10-year U.S. Treasury yield, going back to the 1950s, you will find that U.S. nominal GDP shows a steady skeptical trend. I think this is because banks are marginal buyers. If the yield rises out of balance, then for marginal sellers, if the yield is too low relative to inflation expectations, because they can leverage their balance sheets and can borrow funds into the real economy, then the risk-reward makes more sense.

Kuppy:

I mean, that makes complete sense to me. I mean, I really never thought about it this way before. I’ve always just tracked it based on inflation expectations. But I mean, nominal inflation, you know, a large part of it is actually just inflation, not real growth. For example, I’m not pretending to be a bond expert, but I’m the kind of person who predicts that a major recession is coming, and I don’t know what else to do George:

Yes, maybe I created a monster. Well, those who have watched my whiteboard video know my thoughts on this. So I don't want to elaborate further; let's talk about some things regarding gold. I mean, gold prices reaching $4,000. I think this is another thing we can learn from professional hedge fund managers because from my video comments and from many retail investors I've heard on social media, most of them are long on gold and have been for a while. So they are doing very well, but they are all a bit panicked, a little frantic, saying, "Oh my gosh, is it time to sell? Is it time to sell?" Especially those holding small-cap stocks. So how do you view gold and small-cap stocks?

Kuppy:

Right now? So I have no small positions. I wish I did. Gold. I think it will go higher. I mean, when you hold gold and we don't know any situation, I mean, my personal view on gold, but when you hold gold, I think you need to look at it and say, "Am I planning to be a long-term owner of this?" As for my physical gold, I will never sell. Then am I planning to trade it? Look, it is really overbought. Things will become more and more overbought. Yes, feeling anxious, sell some. Historically, when gold exceeds twice the total maintenance cost, it is really overvalued. It is expensive. Gold is expensive. The gold-silver ratio. Gold, you know, I think silver has caught up a bit recently, but the gold-oil ratio, like the gold-health ratio, yes, it is becoming expensive.

Or you might just think the world is really changing, and they are resetting the monetary system, which is exactly my view. And this is a good opportunity, oh, this is a good opportunity, we have some good things, I mean, we express our view on gold in two ways. One is we have a company called Sprott. They are an asset management company. They manage about two-thirds of precious metals and one-third of uranium, optimistic about precious metals and uranium. When the prices of those underlying commodities rise, their assets under management (AUM) will also rise. Then, because Americans love to indulge, they will bring in capital inflows. So you can participate in the rise of gold prices in a light-asset way.

More importantly for me, I think the reallocation of investors towards gold. In terms of their exposure, the holdings of investors are simply absurd. I mean, most investors hold zero, and most investors should hold a few percentage points, or at least that's how it has been historically for decades. Therefore, as these funds are reinvested into gold, I think the AUM in the spot market will rise, and they will also charge some fees. The management is a bit greedy about their revenue sharing, but they will share a little with me; you see, this is not very suitable for us.

You know, especially this year, I also invested in a company called Major Drilling. They are the largest drilling company in Hard Rock. If I understand the gold cycle, eventually people will give money to those despicable junior companies, and those people will go turn the woods into Swiss cheese looking for gold You know, my Major Drilling company should get what they deserve, or maybe even better. Oh, all that money is spent looking for gold, right? It turns out that this is indeed a good way to invest. What's happening in the gold market? You know, I'm not against small companies under $4,000. They make a lot of money. I just know that once gold miners start making money, the government will start looking around and saying, "Hey, why are these people making money over there?" You know, then they start nationalizing and taxing. You know, the CEOs of gold companies don't understand what capital allocation is, and they suddenly start acquiring small junior companies and nurturing new talent, and you never see that money again. I believe in you get what you pay for. I want my dividends, I want my buybacks, I don't want you to invest in mines. I need ten years to make money. In this crazy world, ten years is like a lifetime.

So, no, I don't know any small oil and gas companies, but I think large drilling will perform quite well, waking up stock prices after being dormant for 15 years. But the last cycle's increase was about 20 times. I'm talking about the cycle from 2002 to 2008. So I think it's not as good as the cycle, but the cycle could be crazier. Who knows?

George:

Have you put some money into oil? What are your thoughts on oil? I know I've heard you say before that you are very bullish. I don't know if you still feel that way now.

Kuppy:

Well, the problem with oil is that if... so, I think we are going to see the biggest economic recession in history. Yes, this is from an internal activity perspective. It doesn't look like a recession because they will do a lot of stimulus measures, they will do a lot of inflation, and they will lie about inflation. So they will somehow reverse-engineer actual GDP. But when you ask the average person you meet at the bar how their standard of living is, they will tell you it's terrible. And this is usually detrimental to oil consumption. What we see is that there is a lot of oil in the world, but underinvestment. So if oil demand grows by 2 million barrels a day, that's not enough growth. If it's 1 million barrels, I don't know. It's not that exciting