원문 보기
In 2008, one man looked at the US housing market, something everyone thought would never crash, and bet his entire fund on it collapsing. His own investors thought he was crazy. They were trying to withdraw their money. The media laughed at him. And then it turned out that he was completely, devastatingly right.
His name is Michael Burr. And on June 30, 2026, he did it again. He made a new series of bets, real money, real positions against the most beloved , most massive trade on the entire planet–the artificial intelligence boom. The one that everyone believes can only grow. Today I'm going to show you exactly what he did, exactly why he did it, and then the part that almost no one else will tell you about–honest arguments for why he might be wrong.
Because this is not a video about selling everything. This is a video that the smartest bear in the room sees. Decide for yourself. If you only know Michael Bure from the movie, here's the two-minute version. In the mid-2000s, Boer ran a hedge fund called Scion Capital. While the rest of Wall Street treated American home prices like a law of physics, he read the actual mortgage documents, the fine print that no one else cared about, and realized that millions of these loans were designed to explode the moment interest rates reset.
So he did something that almost no one had done before. He approached the largest banks on Wall Street and asked them to create credit default swaps. Essentially, insurance policies that would pay out huge payouts if these mortgage bonds defaulted. The banks thought they were taking free money from an eccentric doctor.
They were happy to sell. But the Boer came early, and being early is expensive. Month after month, he had to pay insurance premiums to keep those bets alive. As housing prices continued to rise, his clients turned away from him. He faced a revolt from investors who demanded the return of their capital.
Convinced that he was burning through their savings, he refused. He restricted withdrawals. He held on. And when the housing market finally broke, his analysis proved to be absolutely correct . He walked away with $100 million for himself and over $700 million for his investors. The entire saga became a book and an Oscar-winning film, The Great Depression, starring Christian Bale as the barefoot death metal genius who saw it before anyone else.
However, here is the detail that matters the most . Warren Buffett once nicknamed him Cassandra after a character from Greek mythology who was given the ability to truly see the future but was cursed so that no one would ever believe her, right? But he was ignored. Remember this. This is the whole story of Michael Burr, and it will be relevant until the end of this video.
That's what makes the news of June 30 truly extraordinary. Boer no longer works as a regular Wall Street fund manager . In November 2025, he did something dramatic. He returned to outside investments and announced that he was closing Scion Asset Management . He officially withdrew his firm's registration with the SEC.
And this one bureaucratic step has huge consequences. This means it no longer has to file quarterly 13F reports , which force large investors to publicly disclose what they own. In other words, Michael Boer intentionally excluded himself from the network. He no longer owes anyone any disclosure . He manages his own money privately, not accountable to anyone .
So now he only speaks where he wants, through a paid newsletter called Cassandra Unchained, which costs about $379 a year. And that changes how you have to perceive it all. When a regulated fund manager discloses a position, they may be compelled to do so. When a person who has removed all obligation to speak still decides to publish their agreements to the world, they do so because they want you to see them.
This is a message sent intentionally. So, here's the deal. In a post titled " trading post" dated June 30, 2026, Boer laid out five brand new short positions, all opened on the same day. When you short something, you are betting that its price will fall. He sold shares of Nvidia, the company at the center of the whole AI story, along with Applied Materials for $72,940 and the XXX Semiconductor ETF for $642.80, which is $1,989.
He sold shares of Nvidia itself for just over $198 and added Tesla to the pile. But the one that made everyone stop was Caterpillar. Yes, the Yellow Bulldozer Company, your grandfather's industrial stock. Boer noted that this was the first time he had sold Caterpillar stock, saying that previously it had only made him money on long positions.
So why now? Because Caterpillar has quietly become an AI stock. As data center construction exploded, Wall Street began treating Caterpillar like a game of picks and shovels. The company, which sells generators, earthmoving machines, the physical innards of every new server farm, has nearly doubled in value, up 86% in the first half of 2026 alone.
Burris read: “When the market places a premium on AI for a bulldozer company, the deal has gone too far.” And there's a detail about Tesla that tells you all about his way of thinking. That day, Tesla shares jumped about 10%, closing the previous session at $379.71 and soaring higher. Most short sellers wait for weakness.
Boer did the opposite. He sold Tesla shares short at $41,622, writing that he was glad they had bounced back to give him a better entry. He did n't chase falling stocks. He stepped in front of the growing ones. This is a person with conviction or a person who is about to be knocked down. We will come back to what the trap is and how it is reported.
Now, before we go any further, I have to protect you from the biggest misconception you will see in the headlines about this. And that's important because that's exactly what keeps people from taking the Storm seriously or taking them in the wrong direction. You will see articles screaming that Boer has bet billions against Nvidia and Tesla.
Be very careful with this. Boer did not disclose the size or structure of most of these deals at all. Neither the dollar amount nor the number of shares. And when he exercises options , the number that the media likes to quote is the par value, which is the stock price multiplied by the number of shares the option controls.
This is not money he actually risked. Here is a specific example. At the end of 2025, he held Palunteer put options with a face value of approximately $912 million. Sounds like a billion dollar bet. But the actual premium he paid was a tiny fraction of that amount . Same story with his famous 2021 Tesla short position.
The documents showed put options tied to 800,100 Tesla shares, and the press called it a half-billion-dollar bet. But, as journalists later admitted, this figure was not the capital he actually risked. So when you hear that Boer paid a billion against AI, mentally translate that to Boer paying a much smaller premium for the right to profit if AI fails.
The conviction is real. The dollar figures in headlines usually aren't. So what made him do it on June 30? One statement in particular . The day before, on June 29, South Korea unveiled a stunning national plan aimed at becoming the center of the world of artificial intelligence. The package is worth more than $576 billion up front, with Samsung and SK Hynix pledging investments reaching $1.3 trillion over the next decade.
New factories, new data centers, the largest industrial bet in the country's history. The market saw this figure and applauded. There's a limitless demand for artificial intelligence, right? The Boer saw the exact same number and saw the opposite. He wrote that spending outside Korea, he said, was the beginning of the end, and that it was only a matter of time.
Now think about the psychology that arises here. For Burr, trillions in government and corporate spending are not evidence of healthy demand. This is the sound of the summit. They all pile up at the peak at the same time, afraid of missing out, which is exactly the sign of a late bubble build-up. He bluntly said that this moment is similar to the final stages of the 1999-2000.com bubble, and that the entire market is holding what he calls a two-letter AI thesis that everyone thinks they understand, but almost no one actually does.
To make the parallel concrete, Bohr keeps pointing to one company from the last bubble – Cisco. If you're too young to remember, Cisco created the networking equipment that physically connected the early Internet, the ultimate game of swag and spades of the DotA era. When the story was going around that the Internet would eat the world, Cisco became the most valuable company on Earth in March 2000, worth over half a trillion dollars .
The story was flawless. The demand seemed endless. And then the growth stalled, and Cisco shares lost more than 80% of their value. She never returned to that peak. Not in 26 years. The Fury line is straight. At the center of it all is Cisco again, with his buzz and shovels and grand vision . And its name is Nvidia.
And he's not just waving his hand . He points to a specific figure in Nvidia's own books. Its procurement obligations. Money locked up for future deliveries rose to $95.2 billion from $16.1 billion a year earlier. This is exactly the move Cisco made, committing to huge shipments on the assumption that growth would never slow down.
It worked until it stopped. This is where Boer transforms from narrator to forensic accountant. And this is his most serious accusation. When a tech giant spends billions on artificial intelligence servers, it doesn't factor in the entire cost at once. He spreads this cost over the entire life of the equipment.
Several years of smaller expenses instead of one gigantic one. Burr's argument is that AI chips actually wear out quickly. From an economic perspective, a cutting-edge chip becomes obsolete in about 2-3 years because the next generation is much faster. But large consumers, he says, write off these chips for five or six years in their financial statements .
Stretch the timelines and each year's expenses will look smaller, making today's profits seem larger than they actually are. Boer did the math. He estimates that this maneuver will understate depreciation by about $176 billion across the industry between 2026 and 2028. He singled out two names, predicting that their earnings were overstated by about 27% and 21%.
Oracle and Meta. He called the practice one of the most common scams of the modern era. It's a word he does n't use lightly, and, fair warning, it's an aggressive, contested claim, not a proven one. We will move on to the answer. He adds two more concerns. He claims that the same companies have huge lease obligations on data centers that haven't even come online yet.
A debt that is real, but easy to overlook. And he scoffs at what he calls token maximization. Managers are forcing their own staff to spend resources on using artificial intelligence to make demand appear greater than it actually is. What all three have in common is that the demand you are shown may be artificial and the profits may be an illusion.
Now you could dismiss all of this as the opinion of one bear, except that the overall market data is flashing the same warning lights as he is. Take, for example, the Schiller PE Index, a valuation measure that smooths earnings over 10 years to remove noise. It is currently around 40. For comparison, its all-time high was around 44, peaking at 2000.com, the most expensive figure in the history of the US market.
We are actually knocking on this door. And there is also the Buffett indicator. The total value of the stock market divided by the size of the economy. Warren Buffett called it the best single valuation metric. Recently, it was around 222% of GDP. For comparison, at the peak of the dot craze in 2000, it peaked at around 136%.
Read this again . By this measure, today's market is not as stretched as a point bubble. He far surpasses her. And the costs themselves are historic. The four largest AI spenders expect to spend about $725 billion in capital expenditures in 2026, up about 77% from about $410 billion a year earlier. Amazon alone accounts for about $200 billion .
That's an ocean of money being invested in concrete and silicon, considering that demand will continue to grow forever. And the Bur chip index is selling short positions. For the first half of 2026, it traded more than 65% above its 200- day moving average. A level of overstrain that, he said, has only been seen once in the history of the market.
The year 2000. Okay, deep breath. If I stopped there, you would walk away, certain that the crash would be next week. So let me be an honest voice in your feed, because this is where most videos lie to you. Michael Burr has a documented habit of being right about what is happening but very wrong about when.
And in the markets, time is not a detail, it's everything. Let 's go back to 2021. Boer took a huge short position against Tesla, calling it wildly overvalued. Tesla continued to grow. He was squeezed out of position and closed the deal by paying premiums for a bet that didn't pay off. Then, in early 2023, he tweeted one word: “ sell.” At that moment, the S&P 500 index fell by 4.19.
After 2 months, he publicly changed his mind, admitting: "I was wrong when I said ' sell.'" And the market did not collapse. He experienced one of the greatest ups and downs in modern history. The same index closed above 7,300 by May 2026. Everyone who sold, keeping his word, left a fortune on the table.
That summer, he also placed $ 866 million in put options against the S&P 500, as well as a bet against the NASDAQ, which represented more than 90% of his portfolio , and quietly closed those positions while the market continued to rise . So, this is a record. Brilliant structural analysis, terrible sense of timing.
He sees the fault line years before it ruptures. If you acted on every crash warning, you would lose huge profits waiting for a crash that would take years or never come. And there is another side to the argument that is truly compelling. Many serious people believe that Boer is simply wrong this time .
It's their business, it's not 1999. Back then, companies with zero revenue were doing IPOs with a dream. Today, the giants driving this, Nvidia, Microsoft, Meta, are printing real, verifiable money. Nvidia just posted 62% revenue growth in one quarter. Firms like JP Morgan and Morgan Stanley are presenting this not as a bubble but as an infrastructure supercycle, closer to the introduction of electricity than to pets.com.
The CEO of Palanteer went on television and called the sale of these companies madness, pointing out that they are the ones who literally make all the money. And in an accounting dispute, cloud service providers deny that old chips continue to generate revenue for years on secondary markets. So their depreciation over five or six years is not a scam, it's just reality.
Wall Street analysts, if it's worth anything, remain mostly bullish, with Nvidia's average price target still hovering around $255 per share. Here is an honest synthesis, and that is the whole reason why this story is important. Michael Burr is not a fortune teller. He is a reader of patterns, and the pattern he denotes is real .
Estimates are at or above the most extreme levels in a century. Insane spending that assumes demand never slows down and a market that clings to one story that everyone has decided on is a foregone conclusion. These facts are not opinions. They are recorded on tape. But being right about the settings is not the same as being right about the timing.
Bubbles can inflate years after the smart money predicts a peak . He warned in 2021. He warned in 2023. Both times, fundamentals ultimately justified caution. And both times the market humiliated anyone who acted too early. This is the razor's edge on which this whole story hangs. So the conclusion is not that you need to sell everything because Boer says so.
It's the opposite. The thing is, the person who read the fine print when no one else did will reread it again. And it points to the very deal that you probably own a part of. Maybe he came earlier. Maybe he's wrong. Or maybe, like in 2008, he's just the only one in the room willing to say the quiet part out loud while everyone else cheers.
Cassandra was cursed to be right and ignored. The question is not whether Michael Bohr sees something real. The question is whether this is the time when the rest of us finally hear, or the time when he came early again. And that's honestly the one thing none of us will know until it happens. .
Comments 0