On October 14, the latest episode of the podcast Rebel Capitalist featured a talk with hedge fund manager Kuppy, discussing the fatal flaws of the AI boom.
Harris Kupperman, known as “Kuppy”, is a seasoned hedge fund manager and investment blogger, currently the founder and Chief Investment Officer of Praetorian Capital Management LLC. During the period from 2019 to 2021, the fund achieved a net return of up to 711%. He also runs the popular investment blog “Kuppy’s Korner” where he shares his investment philosophies and market insights.
In the show, Kuppy analyzed the investment returns on the construction of AI data centers, revealing a massive gap. Over the next 3-5 years, construction of AI data centers will require an investment on the order of trillions of dollars (for context, the annual budget of the US Department of Defense is only $1 trillion). Achieving a 10% capital return would necessitate $1-2 trillion in revenue, while robust returns could demand $3-4 trillion in revenue.
Kuppy pointed out flaws in the AI business models, explaining that offerings like ChatGPT/Gemini/Claude are highly interchangeable, leading to a lack of customer loyalty (the free versions are often sufficient). Price wars might drive prices down to just above the energy cost at 1%, leaving no room for profit. With rapid advancements in large language models (LLMs), free versions will remain 'good enough', discouraging users from paying; even with the emergence of AI doctors or tax advisors, there will likely be ten free alternatives.
Referring to the recent hype surrounding AI cyclical trading, Kuppy likened it to the telecom bubble of 2000 (where companies like Lucent/Nortel inflated revenue through vendor financing/lease arrangements). Current firms like Meta/Microsoft are leveraging equity/lease agreements on data centers, effectively acting as cover for private equity and venture capital, to create “safe” assets.
Key points summarized as follows:
If you’re seeking returns of about 10%, this seems excessively high for speculative, quickly depreciating capital equipment; akin to a necessity of approximately $2 trillion in revenue. For healthy returns, one would need $3-4 trillion in revenue. Clearly, these numbers are monumental. For comparison, the annual expenditure of the Department of Defense is $1 trillion, covering aircraft carriers and more.
The composition of AI commodities does not prevent other companies or individuals from entering the field: OpenAI may charge you $90, while we might charge $50. OpenAI then counters with $40, and subsequently Claude comes in at $30. This competition could drive costs down to a level only marginally above energy costs, perhaps just 1%, which would represent OpenAI or Anthropic’s profit margin.
Investing in AI infrastructure mirrors railroad construction — participants in this capital cycle experience numerous failures, including at all stages from debt to equity. I believe AI will undergo similar trials, with continued investments leading to repeated bankruptcies. Microsoft or Meta might not go bankrupt, but they could write off $500 billion in investments, wiping the slate clean and starting anew.
The reason for describing this as 'cyclical investment' is that all the figures are fictitious. The same billion-dollar figures circulate endlessly, designed to lure Wall Street into investing more. Eventually, Wall Street will see through the ruse; but before that, retirees will lose hundreds of billions, potentially amounting to trillions this time around—all going bankrupt.
I conversed with two individuals in private equity—one focusing on credit and the other on equity. They remarked, “We possess these assets worth tens of billions, but we see them as worthless.”
Regarding the infrastructure for data centers, some companies are acquiring land next to gas facilities, hoping to sell it to titans like Google, Microsoft, or OpenAI. This puts one in mind of 2006 and 2007 when everyone was buying single-family homes to flip for profit, believing property values would only climb. Now, it seems we’re not flipping homes anymore but rather selling land to build data centers.
Here’s a transcript excerpt from the interview: George: Hello to all our Rebel Capitalist friends! I hope you are all doing well. I’m here with my good friend Cupy, and we have plenty to cover. First off, Cupy, what’s your crucial role in the video?
Kuppy: Let me explain the various long-winded arguments people bring up. Every dollar in a 401k account goes to buy Nvidia. I’m pretty sure 8% of my dad’s retirement funds are locked in Nvidia, and he doesn’t even fully grasp what Nvidia is. But you see, this is a cheat code. Every dollar you make, every dollar you save, goes towards Nvidia. By 2030, around 150% of the world’s power demand will go towards AI chips that make cat memes to post on Instagram as holograms. Yes, that’s the whole story. Is everyone doing AI now?
George: Well, goodbye to the great root. Sure. Now, putting jokes aside, I feel discouraged by something – a lot of my friends aren’t actual investors; they are from different fields. They send me podcasts like Jensen’s, you know, the ones countering claims of an AI cyclical bubble, whatever you call it, but they never provide solid numbers. They might say, well, you invest in this while also investing in that, but they never dissect the actual math, right? Where does the revenue come from? How much do you expect? What are your costs?
Jensen’s argument is always, oh my gosh, I invested in this company, and I should invest more. What does that even mean? What I love about your writings is that you genuinely throw down a pencil and notepad and say, hey, let me run some calculations. So, bring us back to the origin of your first piece, and I’d really like to know how many people in the AI sector reached out to you before you wrote your second article.
Kuppy: For those who haven’t read my blog posts, you can find them on precap.com. Although they are free, people continually tell me how great AI is, and I’m not bright enough to argue against it.
Look, I use AI, and I think it’s fantastic. You know, it makes me more efficient. However, it often makes mistakes, somewhat akin to a first-year analyst. It can do your first-year analyst work faster, and it’s free. Thus, I don’t oppose AI, but I’ve considered that if this is such a brilliant investment, what should the figures be? We’re projected to inject about $400 billion into data center construction this year, which is quite substantial.
If you assume the depreciation curve for these devices is three to four years, I mean, data centers themselves aren’t durable. They evolve with technology, and even the building itself becomes virtually worthless. I’ve heard about data centers that are now shuttered. They operated for three years before technology outpaced them, rendering their cooling systems obsolete.
However, if you project these things lasting three to five years and assume a 25% gross margin, which is almost a fictitious number… who knows? You need about $1 billion. I mean, it requires $1 trillion for these data centers. For a strong balance, you would need this entire sector to see about $1 trillion in revenue. And you know, that assumes a billion-dollar payment for financing costs or other expenses, which is the revenue needed to get these things built, right?
As I’ve said, if you aim for approximately 10% capital returns in speculative, rapidly devalued capital equipment, that seems a bit lofty; you’d need about $2 trillion in revenues. For good returns, expect around $3-4 trillion in revenue.
These numbers are immensely large. Keep in mind, the Department of Defense spends $1 trillion annually, including aircraft carriers and so on. We can’t shift everything to AI. I have no idea where this money is coming from. For instance, the total revenue of the entire AI industry is around $15-20 billion this year. Remember, you need $500 billion to breakeven. Yes, so that’s 30 times growth you’re looking for. I don’t know how you achieve that. It simply doesn’t add up. The things being constructed have no chance of profitability. I suspect those venture capitalists have noticed that pattern. I mean, they’ve utilized similar models with Uber, subscribing to the model. They invested around ten years to scale it up. I mean, that’s great, right? You just spend $4 to get anywhere you want in town. You know? They had gotten used to subsidizing those operations, and it got so large that they’ll likely go bankrupt before reaching profitability, as no one is prepared to pay for AI.
George: Yes, a point I find reasonable is that their current annual revenue is merely $12 billion because they are only operating in the B2C sector. But the primary aim is to embed AI into everyone’s lives, making people feel as though they must have it, just like smartphones. They’ll pivot to the B2B market thereafter. When they start selling in the B2B sphere, it will make perfect sense for all these companies. Let us say, for instance, Microsoft annually spends $50 billion on staffing, whatever the case may be. They could save that expenditure using only AI. Hence, that $50 billion could flow to OpenAI.
I heard Jensen state that the global economy stands at $120 trillion, suggesting that half of it revolves around computer-based roles—salaries of non-manual workers. And AI will take over all these jobs. Roughly speaking, this $50 trillion could funnel into OpenAI, covering all data center costs. What are your thoughts about this critique? I’m sure you’ve caught wind of it previously.
Kuppy: I don’t think it’s $500 billion, nor $50 trillion. It feels more like $5 trillion. If we exclude those computer-based positions—the highest earners—where does the income even derive from? I don’t understand how it works, right? Hence no income. Because remember, 10% of the populace generates half of the consumption. You mention those individuals having no jobs. So, where does consumption arise from? Where does the income stem from? It’s a cycle. Even if these folks won, they still can’t prevail.
Look, I entirely see it. In ten years, I will have an AI doctor. I’ll be using AI for my taxes. I’ll utilize AI to submit my ADV forms to the SEC. But if today I fill out my ADV form via AI wrong, I might land in serious trouble. We know that AI has both its upsides and downsides. I don’t want to wager my career on whether they comply. I prefer hiring a seasoned, highly skilled professional for this task, who would need ample liability insurance so we can litigate each other if something goes wrong.
You know, running a hedge fund is great fun, and I genuinely have no desire to testify in front of the SEC. I mean, we’re not at that point yet, but it’s bound to arrive. However, if we’re spending $500 billion this year, next year it may become $6.7 trillion, perhaps even $10 trillion. I heard by the century’s end, spending could reach $50 trillion; this is downright absurd, right? When considering inflation, that’s equivalent to seven times the outlay on interest highways. The highways are cool and handy, plus AI can concoct cat memes. Well, I believe we might reach that point, but who’s going to spend $50 trillion to make it happen? What’s the return once it’s achieved? I’m part of the old-school Baby Boom generation, the kind that sees money in and out. I received venture capital, you cashed out, and you wouldn’t see returns for years—maybe even a decade—yet eventually, that money needs to come back. The longer the delay, the better the overall cycle returns.
As we said before, there are no returns. Even if you invest $50 trillion and your capital increases from $15 billion this year to $50 billion, you still have zero returns. The numbers sound great, but I mean, showing $50 billion in revenue is merely revenue, right? Because, like those LLMs (Large Language Models), they’re exceptionally good. Free versions will always suffice, especially if they enhance their offerings by 25% or 50% yearly from now on. Who’s going to pay for this? No one will. Yes, I estimate we will even get things like AI doctors. There will be ten different AIs. Yes, and all will be free.
George: Yes, that’s essentially my core concern, and the largest counterargument I make against AI giants’ hyperbole. You’re dealing with something that is pretty much a commodity; there won’t be any emotional attachment to OpenAI or ChatGPT, just like with Gemini. If Gemini rolls a better version out, you’ll switch to Gemini. Or if Claude offers a superior version, you’ll transition to Claude. There are no network effects. It’s not like all your friends must stay on Facebook because they’re all there. Or your hedge fund friends have to be on Twitter because all of them are there. It simply doesn’t exist with AI.
Microsoft indeed has its Azure product. I’m uncertain if I’m pronouncing that correctly, but you know what I mean. Azure serves as their backend, allowing them to monetize various add-ons by leveraging cloud computing, etc. So, once a company relies on Azure, they integrate into the ecosystem and find it hard to leave. I understand this. Their technique is introducing AI as a tool to ensure customer retention; keep clients from churning. Unfortunately, the introduction of this tool incurs costs, thus suppressing profit margins.
Kuppy: I think the argument here is that Microsoft can fire all employees and hypothetically give every worker $100,000. Now they could suddenly offer OpenAI $99,000 or $90,000 for that matter. But my point is this commodity element does not prevent other companies or individuals from swooping in and saying, fine, we’ll charge you $90, you $50. To which OpenAI replies, alright, we’ll charge $40. Then Claude offers $30. Before long, you’re competing your way down to minimal costs, potentially nothing more than energy costs plus 1%. This 1% is where OpenAI or Anthropic's profit falls.
Consequently, the issue lies with the supplier; it’s treated like a commodity. What’s your take on this? Or do you think I’m misunderstanding it?
This situation will persist until someone declines to fund it. That’s the answer. You have a lot of participants, and it’s a matter of survival. You have some massive corporations with genuine cash flow and profit margins, alongside ample borrowing power. You’ve also got private equity; once they sink in a single dollar, they already see a tremendous profit, especially given their ten-year scaling models.
That’s how exactly they make their profits. You have funding-rich venture capitalists too. I believe the best analogy is the construction of transcontinental railroads. We built some of these railroads, yet the routes taken are less significant. You know, there are improved routes and inferior ones. But trains must run, and the costs to operate these trains and maintain the tracks are fixed. Therefore, they run these projects at a loss since a small loss outweighs a large one, battling for cargo. However, that cargo simply isn’t sufficient; they’ve gone bankrupt multiple times. And those tracks – they are still operational to this day. I mean, after 150 years, they've undergone renovations, and we still utilize them.
So, just as we constructed transcontinental railroads, it’s brilliant, yet all those who invested in these railroads underwent countless failures throughout the capital cycle, working through debt, equity, etc. I believe AI will be identical. They will continue pouring vast sums into it, only to declare bankruptcy time and again. Maybe Microsoft or Meta might not go bankrupt; their bankruptcy might entail something along the lines of writing off a $500 billion investment and starting back at square one.
I mean, people can be a bit sly, right? Just like those circular trades you mentioned. I can’t foresee anyone genuinely planning to invest $1 trillion in these types of trades. What I mean is, they’re merely orchestrating the releases to put into a press release. This serves as a cover, enabling private equity to raise funds and establish data centers, with you compelled to utilize these data centers. Thus, they merely create a $10 billion data center.
Perhaps, you know, when Meta or Microsoft stakes a $5 billion or $10 billion equity interest in it, it may appear, oh, they’re shouldering a risk portion; it’s all safe. Then they ink a deal to lease that data center for 15 years. Oh, that certainly looks secure since Microsoft isn’t the kind of titan that we’ve never seen default.
But in this instance, you encounter a few private equity dudes who will procure 9% subordinated debt. You should be aware if it’s 9% debt; there must definitely be issues. Then you encounter companies such as Pacific Investment Management Company (Pimco), which implies my father’s retirement account is purchasing super-safe priority debts at 7%. This ought to signal there’s something unstable since, fundamentally, a safe debt is akin to a 4% U.S. Treasury bond.
I suspect these will perform similarly to CDOs. I mean, they take those CDOs, and essentially it’s the squared version of a CEO's inefficacy, mixing its worst aspects into a new structure they label as “squared CDO”. Their advantages claim to be AAA rated. We’ve realized if you mix raisins with... you fundamentally receive... and those products yield adverse outcomes. I mean, their loss rates could approach 90%.
So, when you possess a data center where technologies evolve in five years, and the cooling system must be installed underground instead of above, what transpires? Oops, you never constructed a basement; such scenarios transpire regularly in data centers as technology constantly evolves.
You know, I trust the individuals driving Microsoft and Meta are all decent people, but you think they’ll keep making payments throughout a 12-year lease? Or will they craft an excuse to default in some rogue court and spend five years battling each other, perhaps returning two-thirds or half of the money? After legal fees, the amount you receive drops even further.
This pattern repeats itself. My team and I have witnessed similar cases in pipe contracts with unconditional contracts. No one desires a pipeline to remain filled for two decades. If an oilfield encounters problems, one will always find an escape route. It's the same dynamic here.
So fundamentally, you’re financializing these entities. And firms like Microsoft, I mean, look, they’re not getting wealthy through ignorance. They’re finding methods to mask those deficiencies. That includes companies like Core Weave, which functions like a massive stock market.
I mean, look, the news release states the stock price increased today due to insiders only swindling $2 billion today. You know, in standard teamwork, they often sell around $3 billion to $4 billion daily. Today, I don’t know, perhaps they wandered off to buy yachts or some such, you know, forgetting to place an order for new sellers.
$2 billion absolutely appears bullish if insiders aren’t selling quickly enough, reflecting their outlook on the asset. Yes, you can claim that ten companies are navigating transactions worth thousands of billions, and all are operating similarly. I’m not sure if I want to indicate that not everyone is doing this; I genuinely maintain skepticism, right? As a young person in 2000, I witnessed its unfolding. Look at Lucent and Nortel; you know, they are now Nvidia. They were financing with suppliers—clients unable to raise funds from capital markets. They got shut out from capital markets. Thus, they publicized these enormous transactions. We’ll lend this company billions, trying to convince Wall Street, oh this must be secure, you know, if Nortel is willing to lend money.
Then Nortel would reply, we’ll lease your telecom system's spectrum and fiber capacity. What’s the point? Merely to create revenue? I mean, why does Nortel need the telecom system’s capacity? All they’re accomplishing is fabricating false income and recycle it back to Nortel, using that to pay the interest on the debt they borrowed from it, right? Then, when you receive equity in return, you push up equity value, gaining profits, just like those tricks they’ve pulled. So the incentive mechanism is—it’s like, alright, let’s append a zero to all those numbers. Okay, I’ll lease these capacities at a price far above market price. Do we understand how to handle market prices? Do we know what market prices should be? We wouldn’t lease at prices significantly above market rates. You’d use that money to acquire us, buy more from us. We’ll give you $1 billion, and you’ll pay us $1 billion. This will repeat indefinitely.
This is termed 'circular investment' because all the numbers are fictitious. Those same billion dollars perpetually circle each other. It seeks to lure Wall Street into investing further. Ultimately, Wall Street will uncover the ruse, yet prior to that, retirees will incur losses in the hundreds of billions, just like this cycle; only this time, losses could reach trillions—everyone is bankrupt.
You see, they’ve all gone bankrupt. You all ended up in jails. This is all a ruse. I mean, people like Adelphia might still reside in prisons.
George: Yes, indeed. And you know, it may not repeat itself, but it’s clearly a Ryan-style scenario. It appears, alright, we previously financed all capital expenditures through free cash flow; now they appear to be leveraging debt and equity again. At a certain point, the runway simply ends. It’s like can you derive anticipated profits from our present $12 billion revenue stream that equates to tens of trillions in profits? Can you successfully transition from free cash flow financing to debt financing and then back to free cash flow financing?
I believe that’s the crux of the issue. And I suspect the odds are very slim. Giant firms like Microsoft and Google seem to be pouring billions into this or the Metaverse if not out of whim, simply because they have no alternatives. It’s not that they see this as profitable, rather, it’s a must to avoid losing market share. Can you feel that behind-the-scenes atmosphere?
Kuppy: I believe this genuinely pertains to survival. Indeed, I think you’re correct. Many folks perceive it, saying we don’t want to be left behind regarding this new technology, without truly comprehending what it entails.
If we lag behind, what will we have left? I mean, just take a look at those corporations that are splurging the most. Alright, Google, I haven’t used Google search in six months, right? Do they possess their AI search yet? Well, not without ads. Google's search has been oversaturated. You ask it a question, say regarding tomorrow's weather, it bombards you with banner ads for a hundred things to click on. It’s maddening; they’re falling behind. Search is their business. Waymo seems cool but isn’t generating income. Thus, they face no choice. Whether in ten years or ten months, their operations could cease to exist. Meta, for instance; who still uses Facebook? Instagram fell to TikTok. What’s Zuckerberg going to have in five years? It’s a matter of survival. Indeed, it’s today’s cash adhesive. Throwing caution to the wind amounts to little but gambling.
George: It’s like they lack options to continue their investments. Just as retail investors have been pushed off the risk curve in the last fifteen years.
Kuppy: But consider who isn’t spending. You know, Amazon isn’t spending; it doesn’t matter to them. You know, Apple isn’t spending; it’s really irrelevant for them. Yes, that’s my belief.
George: Brother, it is quite interesting, just the other day I was researching pets.com, as I received feedback indicating this scenario is entirely different from the internet bubble. Looking at pets.com, it had zero revenue. Any rational individual wouldn't have invested in this company. It’s utterly illogical. I interacted with ChatGPT, requesting it provide bullish arguments and sources for pets.com. Upon reading the bullish arguments from 1998, they appeared coherent and rational, striking home as compelling. I thought, hey, this sounds great; where should I invest?
But intriguingly, do you know who the largest initial investor was in pets.com? None other than Bezos himself. Yes, the Bezos of Amazon. So the point emerges; I think you indicate that he was somewhat detaching himself from this investment scenario which is quite compelling in its own light.
Kuppy: Unbelievable. I mean, as you previously stated, these occurrences often rhyme. I think it bears tremendous importance for investors to witness. Yes, I recall vividly how severe the 2000 bubble was. Some friends of mine, you know, not exactly ultra-wealthy but nearing the mark of around 50 years, could’ve retired.
Yet instead of managing their portfolios as usual, they funneled all funds into tech stocks; some people past 70 are still employed, as they collapsed amidst margin debt nonsense.
These cycles are astounding—I never envisioned encountering a cycle like this again in my lifetime. I presumed they’d emerge every century or so. Instead, a mere 25 years later, we are facing another as they printed excess money while some specific sectors, especially VC, committed extensive blunders, scaling up and earning massive profits. They merely profited through willful intent, disbursing funds for projects they believed would yield business plans with enough future investment. I don’t have clarity; I still struggle to determine how they will liquidate these funds. This situation is utterly astonishing.
George: Indeed, that’s the central issue at hand. Now, the conversation has shifted slightly; I know prior to our live session, the markets were crashing. Is that still the case?
Kuppy: It’s still in free fall. Is that really the case? I’ll avoid risks like madness.
George: Yes, but it has a sense of immediacy, given the AI bubble, people need to read your analysis on this matter, delve into the math behind it. And they ought to listen to those insider stories, as we genuinely contacted them over the phone.
Kuppy: I find it interesting. Let’s pause and discuss this topic for a moment. Earlier, I authored this blog post, gaining a decent audience, though not as extensive as yours. Then suddenly, someone approached me declaring, Kuppy, I agree with your point. I was thrilled you said that. I thought it was utterly insane. Every day when I returned home from work, I would ask, how will my company ever become profitable?
George: All these are insiders. Let me clarify, these are indeed insiders. They’re not random individuals like me making comments on Twitter, saying, “Oh dude, I’m happy you said that because I feel it too” – they are active in the AI sector today.
Kuppy: They are akin to friends preselected to share insights. I intentionally do not speak with junior employees because I aim to engage with those who genuinely understand their perspectives. I engaged a CEO of a data center company, which proved enlightening. Furthermore, I’ve had conversations with two individuals involved in private equity, one focused on credit and the other on equity. Indeed, they mentioned having these assets valued at tens of billions yet deem them to be worthless. I’ve talked with someone involved in cooling technology. You know, I’ve engaged chip design professionals and a few venture capitalists, one of whom mentioned, yes, we’ve recently completed another round of financing, significantly outpacing last time’s valuation by several billion dollars. Our revenue sits at roughly $1 million. I’ve consistently told my board, this isn’t worth the effort. Yes, they exclaimed, I keep nudging my board, let's execute and finalize this task.
They don’t perceive its worth anymore; they say, every six months, valuations double. Why sell now? They don’t grasp it—I fail to comprehend my involvement in it.
All these senior personnel express astonishment at individuals fervently advocating this as the future. They’re solely fixated on making profits while realizing those figures are fundamentally unworkable. We’re facing negative growth, with investments multiplying losses. Typically, such enterprises ultimately achieve economies of scale. Although I’ve spoken to scores of individuals, I would only extend caution to doctors or any professional, so I must articulate judiciously. However, they nearly universally acknowledge it’s preposterous, and they grasp little of its intended functionality. I think the real meaning behind my blog post boils down to an intriguing expression, mirroring that everyone inherently knows yet no one explicitly states.
George: It’s like the emperor has no clothes. Indeed.
Kuppy: Look, I inquire of these professionals, why don’t you express these concerns? I can’t inform the board; they’d terminate my employment. They’d consider my foresight lacking. I cannot advise employees to resign to pursue their passions at companies filled with zeal for this opportunity. I can’t convey these narratives to such individuals; it almost feels like a support group. They merely contact me because they lack a platform for such release. They genuinely fear that should they voice anything, it might reach the ears of those funding their endeavors, consequently resulting in termination, right? It’s odd. But this reinforces my belief that my figures are accurate.
You know, there once was someone with considerable AI investments within a hedge fund, echoing your sentiments regarding many high-paying jobs, like those of physicians, lawyers, and accountants, being at risk of being replaced. I believe he’s right, and I concur with him.
George: By the way, do you view this being offset by energy costs and the expenses stemming from the global rise in AI usage?
Kuppy: I feel he’s a decade premature with his thoughts. Once you discount the money back to the present cash flow, it becomes next to worthless. However, I understand what he’s indicating. This reflects bullish market sentiment. No one can explain to me otherwise – next year will yield $100 billion, and the year thereafter $500 billion. I mean, when I consult these individuals, they concur income might double, but following from this income, next year may see growth from $15 to $30 billion—much of which ends up being pass-through revenue, correct?
You know, Microsoft buys from Core Weave, and Core Weave purchases from this firm. How much of that is genuine net revenue? After all, these are private entities, and I wouldn’t burden you with too much if those figures genuinely appeared bright, they would assuredly initiate pressure statements on them.
George: Yes, you know, one thing I noted is that you ought to focus more on people’s behaviors rather than their actual speech. Altman publicly claims, of course, he’s been extolling its magnificence. Yet when you consider his actions, he’s making substantial investments in other companies. I pondered, well, if he’s so bullish on his own company, why doesn’t he reinvest those funds into OpenAI?
Why would he allocate those funds toward investigating other firms that might not crash, might not commoditize, or perhaps offering no product at all? So, that seems rather foreboding. Another thing I want to share, some anecdotes I picked up over the weekend while on the golf course because I’m a member of a club where visitors come to vacation on weekends to play golf.
Previously, I had a conversation with a guy who could be seen as an individual investor. This person appeared terrified of missing opportunities (FOMO) and excited about purchasing cryptocurrencies. Yet he has substantial capital. He revealed what he’s engaged in now. Given his thriving career, he, alongside his partners or others, established an LLC, purchasing land next to gas facilities and aimed to sell it for data center construction to giants like Google, Microsoft, or OpenAI.
I reflected on 2006 and 2007 when everyone was buying single-family homes and flipping them for profits, knowing property values always trend upwards. It resembles that scenario; however, instead of flipping homes, we’re offloading land for data centers.
Kuppy: Everything seems utterly insane. About six weeks ago, I attended a wedding and encountered uncles and aunts I hadn’t seen in five years. I perpetually felt like a financial broker; they always wanted to gift me stock shares.
Kuppy: Indeed, I mean, I must say that today’s 60%/40% portfolio encompasses 60% Nvidia and 40% Bitcoin. No kidding. Some individuals relish a triple-leveraged version of that.
These individuals, like my cousin, saw share prices surge by 150% this year. I felt somewhat envious, as I performed decently this year but didn’t reach the 105% mark. Yes, they often purchase these leverage options; these triples leverage ultimately multiply their profits sixfold. You know, this isn’t new to me.
I find it amusing that many of the “emy” you see today are those who begin to indulge after capitalizing in the stock market. Yes, indeed, I fully agree. An overwhelming majority suffered under inflation pressures. But if you’re holding Nvidia at six times leverage, it’s not peculiar that you’re indulging, vacationing, or living lavishly. This essentially spurred emy as the moment videos document a two-week downturn, all these individuals would face margin calls. I personally witnessed such scenarios, which attribute to the market crash.
I mean, we’ve never encountered a moment where everyone holds triple-leverage stocks, and suddenly a 10% drop occurs. This situation necessitates margin calls, and that’s simply terrifying, right? You know that.
George: Another point of irony occupies my thoughts: I firmly believe that Nvidia’s ascendance to a $4 trillion value is largely attributable to the influx of passive funds. These funds flowed into SMP. This is evidently a weighted index, so for every dollar that enters, they earn eight cents. Although if AI is truly exceptional, if AI displaces all professions, passive inflow reverses into passive outflow.
Now all of a sudden, for every dollar leaving, it’s eight cents that go toward Nvidia. So, if you…
Kuppy: Don’t do that; I sense that’s a frequent occurrence. Look, I view the world very simply. I don’t know; no, I mean, I think AI will indeed siphon off many jobs. It’ll initiate somewhat slowly but persist over time. This won’t be an immediate shift. More likely, individuals will conclude maybe we can manage without hiring, let’s assess AI after a year. You know, our business might blossom. Let’s extend hours for everyone. Why would you hire someone if you can just train and then lay them off?
I perceive a considerably inflated aggregation.
George: In such an environment, you confront deflationary pressures, especially as it erodes the job market, unemployment spikes, passive inflows transition into passive outflows, the market declines, bulls emerge, and all exacerbate the K-shaped economy’s expenditures.
Kuppy: Inflation feels inevitable. However, I particularly sense folks resembling Trump, searching for someone to forge plans, feigning as free-market enthusiasts while coveting central planning. I foresee they would undertake controlling measures and obscure fiscal policies, and you’ll observe unusual adjustments emerging. The trade and investment landscape will become increasingly arduous.
Yet I perceive a reality reminiscent of the Great Depression era where ordinary citizens see a 15%-25% decline in living standards. Why wouldn’t we repurpose history? Unless this unfolds within an inflationary context. This doesn’t echo the deflation dread gripping everyone. It resembles Argentina. I deem this the worst-case scenario because in that scenario, worthwhile investments are scarce.
George: As of now, I view persisting U.S. conditions akin to those of 2021 and 2022, or similar to situations in the 1940s. I persistently draw parallels because I believe that unless the governments act passively, the economic fundamentals will lead to deflation, or even a thorough deflation. Conversely, if the government intervenes, that materializes as an ideal example. As I articulated, had the governments remained uninvolved in 2020 and 2021, we might have encountered a 15% deflation. I mean, the entire economic structure would crumble under deflationary circumstances.
Yet the government intervened to the tune of "stimmies" (referring to economic stimulus packages), distorting the economy. You navigate all these supply shocks, leading to a CPI surge of 9%. Therefore, I deem the situation to reflect that of the 1940s when price controls, etc. were active. Thus, I concur. I believe the economy desires deflation, but the government will interject with erratic measures, fostering a temporary inflation spike akin to the 1940s scenario. I’ve noted that during that era, CPI was at 19%, only to drop to -2% two years later. It means enormous volatility types should be anticipated, it appears we may navigate exceedingly similar events soon.
But you deliberated on Trump. You may appreciate him or loathe him, etc. He is certainly nontraditional. You also dropped a post; I’m unsure if it was a blog or something on Twitter. On Twitter, you mentioned your experiences as a hedge fund manager, what actions you took during the Trump era, and how those differed from Biden or the Trump 1.0 experiences. Could you provide us with a digest version?
Kuppy: Yes, certainly. I mean, I somewhat with nostalgic fondness recall the Biden days. He’d awaken at noon, they’d garb him in a suit, and he’d shake hands with a cloud, strolling offstage as if nothing occurred, right? Yes, and then there’s Trump, he’s incredible, brimming with energy, hurling bombs at everyone.
But he’s not like those female politicians leaking stories to the Washington Post, only to note polling results elsewhere. Occasionally, Trump merely presses a button on his tweet and thinks, “I’ll go play golf.” The world will ponder the meaning behind that tweet. It’s like, look, I’m presently speaking with you, and the market collapses 2% due to that tweet.
George: Oh, that happens. Yes, as you’ve been observing—and just before we went online, as we stated—some incidents are occurring; the markets dropped. I’m not sure, Kuppy states this was due to Trump’s comments. Therefore, ponder on this.
Kuppy: Yes, you see, all of this could flip within ten minutes; however, we’ve had a decent day today. We’ve shorted the global market, taken long positions on volatility, buoyed the yen, and are long on gold.
Alright, I think we just tweeted again. We chatted for twenty minutes, but within Trump’s realm, one must change perspectives: avoiding maximum leverage. We can’t remain entirely exposed to risks. You know, we’re left with no excess cash; therefore, we attempt to acquire crash puts and then consult to neutralize. Trump behaves similarly. Shocking, then mellows slightly. Thus, it's advisable to utilize a market maker's mindset rather than following trend sentiments. It makes adequate sense. You know, I’ve had to adjust how I interact with the market. We must all adapt accordingly. Still, it generates a frenzied frustration as you remain in the dark regarding his next moves.
I recall I could attend meetings, you know, I lunched with friends, discussing markets back in the day. Now, however, I’m entirely prevented. I must remain connected to Twitter because you can never anticipate when he’ll tweet. I check my phone every three minutes. Just as you indicated, I find myself waking every 30 minutes at night because of our targeting today. You know, and then you halt and feel, alright, I mentioned this, you know, we’ve had no engagement with this fellow.
For individuals like me, who harbor significant aversions to risks, it's incredibly challenging and infuriating. I strive for discipline but find it demanding as you genuinely lack understanding of who is immune or secure.
His actions frequently lack specific logic. I perceive it to resemble mosaic formatting; although I comprehend the goals, the odd byproduct seems to lead to instances where incidents transpire and, you know, it isn’t something my firm intended to execute, merely, you know, I enjoy jesting, spending over a decade in business in Mongolia.
It’s amusing to jest about my constant fear of the government suddenly enacting a law that would confiscate所有資產 titled to anyone with the surname that begins with "K." You know, any K surname individual in our assets would be seized. Then, upon reflection, a Mongolian acquaintance from parliament approached me saying, Kuppy, we’re sorry. We must focus on Mr. Jin, the Korean. You’re merely the K-alias.
Oh, if you stay quiet and cozy with the government, simply appear friendly enough, we’ll return your asset in six months. That’s how things function in Mongolia, and it’s somewhat analogous to how Trump operates.
The unique dictatorial rule and mentality of the Mongolians complicate investments. Yes, I believe the macroeconomic results’ eccentric volatility should prompt lower stock PE provisions as you expressed, arguing that our valuations have reached historical peaks. We should align ourselves with a banana republic, as we increasingly exhibit traits of one; that’s how I’m perceiving it now.
George: So, as a fund manager, what actions will you now undertake? I don’t need specifics regarding your strategies, etc., but what minor adjustments are you making compared to during the Biden administration?
Kuppy: Firstly, I maintain a pessimistic perspective regarding the global economy. I enuntiate, I seek investments devoid of dependency on government incomes or government backing, or more precisely, where the government executes things accurately vis-à-vis permits and such. You know, I prefer to steer clear of targeted sectors favored by the government.
I aim to invest in industries that remain unaffected—even should the economy underperform (which is my forecast). My portfolio composition has undergone significant transformation. Lastly, I’ve minimized my risk exposure. Previously, I’d run at 1:15, 1:25, and had nearly no shakiness in my shorts. Now, I’m running about 80 to 90 seconds while remaining resilient. If a crash occurs, I’d deploy crash puts and undertake other measures, fundamentally altering my approach.
George: Yes, that makes sound rationale. You know, I’m effectively a novice in this area. However, one element I’ve tried that yielded decent results—I'd like to hear your stance on this—is that I refrain from constructing positions directly and instead consider spreads. For instance, a while back, the two-year and ten-year spread resided around 40 to 45 basis points. Presently, I track the two-year versus ten-year spread chart, referencing periods of inversion and other rates calming down from inversions, recognizing that invariably, these moments cause volatility.
Ordinarily, you’ll witness a more pronounced bowl-shaped curve since the Fed typically reacts swiftly, lowering rates to address these. Later on, you achieve a bear market curve when they cut rates to zero or lower another level. So, from inversion to spreads reaching 150 basis points, then extending to 250 or 300 basis points, I ponder that this rationale appears compelling.
The most enticing aspect or target is, should we find ourselves embroiled in a significant recession, you might obtain a bowl-shaped curve, wherein you’ll have succeeded. Conversely, should you not experience a recession and miscalculate entirely, and the economy takes off entirely with heightened inflation and growth expectations, you might attain a steeper bear-market curve.
Thus, the only scenario wherein you've failed is on a flat curve. Yet, the odds are perhaps a tad less compared to the other two outcomes. So, I continuously endeavor to approach matters this way, rather than directly following the ten-year interest rate path or executing short plays on ten-year rates.
Kuppy: Yes, I mean, I’ve not truly engaged with spread metrics. I feel a bit ashamed admitting that. I genuinely hold long positions in ten and thirty-year bonds while harboring a burning dislike for them. I’m painfully aware of their lack of worth. It’s a feeling I can’t shake as I predict they might go through a Pavlovian trigger. Oh, we’re experiencing recession now, let’s purchase a few bonds; however, you know, I tweet right now—no one’s genuinely bullish on bonds.
The 60-40 portfolio once held resonates with notions of Bitcoin or the equivalent of gold for baby boomers. Nowadays, it seems no one continues to own bonds; everyone expresses nothing can halt this engine, except a derailment.
Just like emy is in shambles. I’m not certain. I perceive myself as a bondholder, which honestly makes我feel nauseous, although I hardly possess comprehensive knowledge on bonds, and I recognize where my stop-loss would trigger if things ever went awry. It all appears quite mad.
George: Allow me to offer some examples that might ease your bond perspective. I’ve done as much research as feasibly possible to determine whether I truly have leverage as I’ve produced numerous whiteboard video segments. You know, I’ve lost count of how many whiteboard segments I’ve done; at least 1000 or 2000.
Thus, to pour effort into whiteboards requires a bit of research. I theorize that bonds often signify curve traits aligning with growth and inflation expectations while remaining detached from supply correlation. The reason bonds convey a queasy vibe stems from witnessing supply spikes, which absolutely materializes. For my point of view, deficits are scheduled to eclipse debt within three to five years, reaching $50 trillion. Absolutely so.
The past bears minimal impact due to bank balance sheet mechanics, as banks remain the marginal buyers. If you’re holding a ten-year Treasury, and the trading rate hits 8%, some nominal growth that remains considered (say at 4-4.5%), that equals a massive opportunity for them as their financing expense rests at 2%, allowing for 8% Treasury returns, effectively hedging against maturity risk. Thus, it operates like an equalizer.
Thus, should you analyze nominal GDP and ten-year U.S. Treasury yield graphs, tracing back to the 1950s, nominal GDP remains potentially consistent across skepticism trends. I suspect banks persist as marginal buyers—for yield imbalances to occur would pressure marginal sellers. Should yields remain balanced in response to low anticipated inflation and growth rates, banks can utilize balance sheets and extract capital from the real economy, thereby making a better risk-reward scenario.
Kuppy: This completely resonates with me. I’ve never ruminated on that concept before. I’ve always tracked based on inflation expectations. However, substantial components of nominal inflation exist more as inflation indicators than genuine growth metrics. I’m not pretending to be a bond specialist, yet I’m someone predicting an imminent recession and uncertain about alternative approaches.
George: Yes, perhaps I’ve birthed a monster. Well, viewers of my whiteboard videos recognize my perspective on this matter, so I don’t wish to reiterate; let’s pivot to discussing gold. I mean, the gold price surpassing $4,000 has become pertinent, leading me to believe this constitutes another insight we can learn from seasoned hedge fund managers.
Hearing from comments from numerous retail investors, most are long on gold and have maintained such positions for quite a while. They display precise sentiments; they often become quite anxious, crossing over into a state of panic regarding when it would be appropriate to sell. Are you still bullish on gold and small-cap stocks?
Kuppy: Currently? I harbor no small positions, but I wish I did. Gold holds my optimism; I believe its price will climb further. You see, while possessing gold amid lack of clarity within the situation, I posit, it requires individuals to consider: Am I an extended holder of it? Regarding my physical gold, I perpetually refuse to sell. Should I dictate when to trade? Observing it, gold tends to become excessively purchased. Things can increasingly appear overbought.
Yes, feeling nervous, leisurely sell portions. Historically, gold exceeds twice its total maintenance cost; that’s when it genuinely seems exorbitant. It’s pricy. Gold counts as a luxury. The gold-silver ratio indicates gold is nearing peaks, you know, I feel silver slightly aligned recently, but the gold-oil ratio, as well as the gold-health ratios—truly, they are adjusting to steep valuations.
Or perhaps people perceive meaningful shifts approaching, restructuring currency systems, which directly correlate with my projections. It’s an opportune moment; oh，此刻的情况真不错，某种意义下我们拥有资源。
In terms of our perspectives on gold, we own an entity known as Sprott. They function as an asset management company managing roughly two-thirds precious metals and a third uranium, maintaining a bullish outlook on both domains. Thereby, when the prices of those base commodities escalate, their assets under management similarly swell.
Then, due to Americans indulging in fine dining, an influx is ensured. Hence, you can engage in gold price surges through a light-asset structure. For me, it’s more pivotal to indicate I believe investors are reconfiguring their gold distributions. Considering their exposures, investors’ holdings feel extremely ludicrous. I mean, the majority hold zero positions; most should allocate a handful of percent at the least; historical contexts indicate a magnitude naturally preferred.
Thus, with funds reallocating into gold, I anticipate the asset under management in the physical market, portfolios will experience surges and likely charge their fees. Management’s greed regarding revenue sharing is somewhat elevated, but they will provide me with a bit of that, consequently making it less convenient for us.
You know particularly, this year, I’ve also invested in a company named Major Drilling at the forefront of hard rock drilling. Given my insights into the gold cycle, individuals will ultimately funnel cash toward those mischievous junior firms, converting woods into Swiss cheese whilst pursuing gold. My Major Drilling should indeed have their due share, or potentially yield better than expected.
Oh, those funds earmarked finding gold altogether, right? It has recently proven to be a transcendent investment strategy, yet you ponder is the gold market presently observing? You know, I don’t oppose those small-cap companies under $4,000 or subsidiaries. They’ve generated substantial revenue. I recognize once mining firms begin turning profits, governments will likely exhibit inquiries, wondering, “Hey, why are these entities making profits?”
Then they commence nationalization, taxation. You know, CEOs in gold companies typically lack understanding related to capital allocation; it turns out they suddenly commence acquiring small junior companies to cultivate new talent while these resources invariably disappear. I consider value for value; that’s the way it operates.
I want my dividends; I’m after my buybacks; I don’t wish for you to squander investments in mines. It takes ten years to see returns, and in such insane landscapes, a decade resembles a lifetime.
Thus, no—I'm unaware of any small oil and gas enterprises, but I suspect significant drilling will yield satisfactory returns, allowing stocks to awaken post-15-year hush. Compared to last cycle fluctuations of 20 times, I say the cycle between 2002 to 2008. Therefore, I belabor it could be less lucrative than that cycle, but cycles may become more frenzied. Who knows?
George: Have you funneled part of your funds into oil? How do you perceive oil presently? I recall you professing considerable optimism previously; I'd be curious to know if you still are.
Kuppy: Well, the oil issue revolves around looming prospects implies…I suspect we’re fated to witness an unparalleled recession in history. Yes, from a budding activity viewpoint, it won’t seem like a recession, given their extensive stimulus coercion. They will enact aggressive inflation measures while also deceiving the public about inflation rates, thus inversely engineering genuine GDP measures.
However, when you casually query those ordinary individuals you meet in a bar about their living standards, their responses echo: it’s dreadful. This scenario typically detracts from oil consumption. We observe ample oil exists globally; yet under-investment abounds. Therefore, should oil consumption escalate by 2 million barrels daily, this growth appears insufficient for sustainable optimism.
Should it be 1 million barrels—I’m uncertain. That doesn’t elicit considerable excitement.
Comments