원문 보기
So, Michael Bury is making a prediction again, believing that AI stocks will crash and cause the next global financial crisis. And for a few of them, like Micron, he's betting they'll fall more than 50% by 2027. And for those who don't know, Michael Bury is the same investor who became the prototype for the hero of the movie "The Downside Game." Before 2008, he bet against the US housing market , and it brought his investors over 700 million, and him personally about 100 million dollars.
So in this video, I'll break down with you what exactly his short positions are, what they indicate, up to what point, and what he can actually lose. Then I will analyze two specific stocks that he is playing against, which are Nvidia and Micron. I will analyze them using my value investing methodology to see if he is only playing against low-quality stocks or also against high-quality ones.
And finally, I will express my opinion on whether he is right, whether he was in a hurry, or whether he is wrong again this time. I will also talk about what I personally do with my own money. Now a quick disclaimer: I am not a financial advisor. Everything I share here is just my personal opinion for educational and entertainment purposes.
Always do your own research before investing. Let's start with the first point: what his short positions actually mean. Before I begin, I want to note that I am well aware that Michael Bury's predictions change quite frequently. I often hear jokes like he predicted 200 of the last two recessions. By the way, back in January 2023, he even wrote just one word on Twitter: " Sell." Just months later, he wrote: "I was wrong to say that." So, as you can see, although he likes to make predictions and always has strong beliefs, he changes and closes positions very quickly as soon as he receives new information and sees that his hypothesis may be too early or wrong.
And, honestly , I think that's how the best investors in the world always operate. They always make bets with the highest probability of success based on the available information and change their decision-making process depending on new data. Recently, on September 23, he started playing against the AI market by buying put options.
This is exactly what he bought. And here are my calculations of how far each stock needs to fall before its bets start to make a profit, and the payoff deadline for them. So, first—Nasdaq 100, about 20% by September 2027. Next up is a chip ETF called SOXX, about 29%, and Nvidia—about 36%. Palantir—45%, all until September 2027.
Micron— over 50%. Nubia—about 60%, both until June 2027. And Oracle— about 60% or more by December 2027 . And among them, Nvidia is the asset I have. In this video, I will conduct a thorough analysis of the two largest companies on the list, using my value investing method to understand Michael Bury's hypothesis and assess its probability based on fundamental indicators.
Part Two: Two Big Short Positions in AI Stocks That Michael Bury Holds and How I Will Analyze Them. So, the first stock is Nvidia, which I have in my portfolio. Before you start thinking that I sold all my stocks, I want to note : I hold stocks, but in a very small proportion, about 35-40% of my portfolio, and the rest is in safe assets.
Regarding Nvidia: Michael Bury opened a short position by buying put options with a strike price in the mid- $100s with an expiration date of September 2027 . Nvidia is currently worth around $230. So, with this rate, Michael Bury is effectively saying that he expects Nvidia to fall by more than 30% by September 2027.
And this surprises me somewhat, because of all the companies I've looked at, Nvidia has extremely high profitability and growth rates, perhaps the strongest " economic moat", a cash-rich balance sheet, low debt, and moderate valuation multiples. I mean, take a look for yourself . Business continues to grow wildly.
When I check the SmartX terminal, revenue for the last quarter was $96.2 billion , up 106% from a year ago. Over the past 12 months, the company generated $127 billion in free cash flow. It has low debt and many short-term assets compared to liabilities, with a current ratio of 4.6. At a P/E ratio of around 30, it's not cheap, but it's not crazy either.
So why is Michael Bury betting against Nvidia? Aside from the general macroeconomics, which I always talk about in my videos about market cycles, why Nvidia? I think it has to do with who actually pays for Nvidia chips, and the risk that customers won't be able to pay and no longer be able to buy chips —like all these AI companies like OpenAI.
Let me give you a few examples. In August, OpenAI agreed to lease a data center in Ohio, and Nvidia promised the owners that if OpenAI stopped paying and they couldn't find other tenants, Nvidia would cover the difference in rent of up to $105 billion . And guess whose chips are in that data center?
Nvidia. This is effectively providing OpenAI with a loan guarantee so that OpenAI can buy Nvidia chips for use in this data center. It's almost like a car dealer co- signing your car loan so you can buy their car. In addition, the amount owed to Nvidia by customers has increased from $38.5 billion to $63.1 billion in 6 months.
And we've seen this before. Around 2000, telecommunications giant Lucent provided about $8 billion in loans to its own customers so they could buy their equipment. And when those customers went bankrupt, Lucent shares fell from $84 to 55 cents. Of course, in Nvidia's case it's a guarantee, not a loan, so the situation is a bit different, but the idea, in my opinion, is the same.
The seller helps the buyer pay. Here's what I think about this . I think Michael Berry is right: any slowdown in one of the big AI companies will cause a domino effect that will hit all the other companies in the industry hard, like Nvidia. I also think Berry is right that the real risk for Nvidia lies in its customers' balance sheets, not its own.
However, I am also convinced that it is extremely difficult to pinpoint the exact month or time. So Michael Berry probably has more insight into why he's betting that a crash could happen within the year. And you know, on our channel we analyze macroeconomic factors, so this is quite consistent with the real estate cycle —we mentioned that 2026 could be close to peaking, and 2027 is likely to be a period of extreme risk .
So Michael Berry may well be paying attention to these macro factors, such as rising interest rates and funding for AI infrastructure. The second stock is Micron. Micron makes memory chips , which are essentially the short-term memory needed to power every phone, laptop, and AI data center. Now the shares have increased approximately 16 times in 3 years—from $68 to $1,074.
And Michael Berry's put options reportedly have a strike price of around $500 with an expiration date of June 2027. So Michael Berry needs everything to fall more than half in less than 9 months. And now, to be honest, their business is incredible. When I check Micron's terminal, revenue last quarter was $54 billion— almost five times more than a year ago.
The company kept about 87 cents out of every dollar of gross profit and made $37.7 billion in profit in 3 months. The balance sheet is very strong, with a liquidity ratio of 3.3, meaning three times more short-term assets than liabilities, and a debt- to-equity ratio of zero, indicating an almost complete absence of debt.
But the competitive advantage is assessed as weak, and that's the whole point of Micron. Memory is actually a tradable commodity. This means that a chip from Micron does exactly the same job as a chip from Samsung or SK Hynix. And now these three companies occupy about 90% of the entire market. So when there is too much supply, everyone's prices fall at the same time.
And it is often very easy to change supplier. As Warren Buffett says, in a business selling a standardized product, it is impossible to be significantly smarter than your dumbest competitor. And this creates a memory loop. Scarcity drives prices and profits skyrocket, like it is now, so every memory manufacturer uses those profits to build more factories.
And it takes years to build these factories. And then, when all the new supply appears at about the same time, prices fall and profits collapse. Michael Bury wrote in July that Micron defines cyclicality like no other, and history backs up his words. Over the past 25 years, Micron shares have lost about half their value or more five times: in 2000, 2008, 2018 and 2022, and most recently last April.
So, here is the point at which many investors lose their vigilance. Micron is trading at only about seven times its expected earnings for next year, which would look very cheap for most stocks. But for a cyclical business, a low P/E ratio near the peak is a warning of an upcoming downturn, not a bargain buy, because earnings themselves are at their peak.
In 2018, Micron was trading at just five annualized earnings just before the stock more than halved. To be fair , optimists say things are different this time , as Micron has signed 26 multi-year contracts under which customers pay whether they pick up the chips or not, covering 35% of revenue through 2030.
But that's only about a third of the income. And most of these contracts have both a price ceiling and a minimum threshold. So, they protect against falls but limit growth. And price growth is already slowing from about 60% per quarter in the spring to 10-15% now, as Samsung, SK Hynix, and Micron rapidly build new plants.
Point three: who is right, what I am doing and why. In my opinion, Michael Bury is right about the direction, but a little hasty with the timing. In 2008, he hit both . In 2023, he was in a hurry and made a mistake. And this year, his short position in Tesla worked, while his short position in Oracle worked against him for a month before changing in his favor.
According to the cyclical model I follow, we are in about year 14 of an 18.6-year real estate cycle, the final stage when borrowing peaks and everyone feels safe. And, in my opinion, the biggest factor that continues to attract capital right now is AI spending. So the year that Michael Bury just moved his bet to, 2027, is the same window that I'm also watching.
I know I mentioned 2026 as the peak period before, but I also said it could be stretched out, so there should be some margin for error. And with the emergence of new information, 2027 now seems more likely. So , like Bury, I too can hurry. As I mentioned in previous videos, at this stage of the cycle, I'm not making any noise because that's not my specialty, but I do have a small stake—about 35% right now.
As soon as the fundamentals of any of my stocks deteriorate, I will get out of them, but my goal is to wait until stocks with great performance and value appear at good prices, and then invest all the cash. If you want to dive deeper into stock analysis and use the same features I showed earlier by analyzing your stocks across various value pillars and financial statements, check out my stock analysis and research platform.
I also have my own microcycle analysis there along with the corresponding indicators. So check out the link in the description.
Comments 0