원문 보기
Most people will take this deal completely wrong. And the worst thing is that they will leave, thinking they have everything figured out. Right now, Michael Berry is playing for Nvidia shares to go down. And this was no small position. At the same time, he holds call options on Nvidia. This is a type of contract that only makes a profit when the stock goes up.
Two bets, opposite directions, one stock, one score. The internet has already decided what that means. He lost confidence. He hedges because he is no longer sure of his choice. He quietly shifts to an optimistic forecast, and no one notices. Each of these interpretations is wrong. What he does has a name on professional trading platforms .
And once you see the math behind it, the real numbers, not guesswork, you'll never look at a short position the same way again . It also teaches you something far more useful than one person's opinion about one chip company. It teaches you the difference between being right and surviving long enough for that right to be proven.
Most people never learn this difference. This is the reason why bearish forecasts end in disaster, even when the forecast itself was correct. So, here's what we're going to talk about. What he actually has, in his own words. Why the numbers you saw in the headlines about his rates were wrong about a hundred times .
Which, in his opinion, is quietly breaking down under Nvidia's record results. And then the mathematics of hedging itself. The part that makes the bet both ways suddenly obvious. Let's get down to business. Let's start with the facts, as almost every version of this story online has been simplified to the point of nonsense.
On August 26, 2026, Berry published a post outlining his positions. Three things are important about this. First, he bought December call options on Nvidia with a strike price of around $200-plus. Nvidia was trading at around $210 at the time. So, these are contracts that only make a profit if the stock rises significantly above its current level.
Second, these options make up approximately 3.5%–4% of his portfolio. Small. Intentionally small. Third, and this is the number that everyone misses. His total short position exceeds 21% of the portfolio, and this figure does not include his put options. So, the bearish side of the portfolio is about five to six times larger than the bullish side.
And he said something that is usually kept quiet. "I'm not playing here for profit." He wrote. He added that he would not have entered into this deal at all if he did not already have a large short position and put options under it. Read this again. Call options exist because there is a short position.
This is not an opinion. This is a tool. Here's the image I want you to keep in your head throughout this video. The stuntman tightens his five-point seat belt before crashing the car into a wall. Nobody watches him buckle up and says, "Look, he doesn't really believe he's going to crash." A belt is not a doubt.
The belt is what allows him to intentionally drive into the wall. This is the whole point of trading. But to understand why this "belt" is needed , why a professional spends real money on protection, instead of just opening a short position and hoping for luck, you need to see what shorting actually does to your risk.
We'll get to that math. First, we need to clear up the biggest misunderstanding in this story. You've probably seen the headlines about Berry betting almost a billion dollars against one company. This number was never real. And how it became real in the public eye tells us something important about how financial news works.
When a fund reports its assets to the SEC, option positions are reported at face value. That is, the market price of all the shares that control these contracts, not the money that the fund actually spent. One option contract controls 100 shares. So, if you buy 50,000 contracts, you are reporting exposure to 5 million shares.
Multiply 5 million shares by the price—and you get a huge number for a headline, even if you spent a penny to enter the deal. That's exactly what happened. Berry himself publicly cited these calculations. "I bought 50,000 of these for $1.84 each," he wrote. Then, marking the publications that made mistakes: "I spent $9,200,000, not $912 million." 9.2 million.
Not 912 million. The figure given was overstated by about a hundred times . Think about what a difference this makes. The Billion Dollar Bet is a story about a man who bets everything. $9 million in options premiums—this is a story about a man who buys a very specific, very cheap insurance policy with strictly limited losses.
These are completely different agreements. The media reported on the first one. He was making a friend. And this concerns you directly, it's not just a small thing. When you see a headline about what a famous investor is betting on, ask one thing. Is it money he spent or an amount under his control?
Almost no one asks this. This is the easiest way in the world to be misled by a number that is technically correct. There is a second reason why this continues. Big numbers spread quickly. Refutation—no. The billion-dollar version of the story spread everywhere in a matter of hours. The correction he made himself, by writing out the arithmetic, reached only a fraction of the same audience.
This gap is not a conspiracy. This is how attention works, and it's worth remembering this the next time a number in a headline stirs up emotions in you . So keep this picture clear in your head. The risky commitments are huge. But the money at stake is not. This combination is not reckless. This is the main point of buying options instead of directly opening a position.
You know your maximum loss on the day of entry, and it doesn't increase. Which brings us to the crux of the argument . Why is he even playing for relegation? That's what makes it interesting. Berry is not saying that Nvidia is a bad company. He said almost the opposite. After the company reported second-quarter revenue of $96.22 billion, up 106% year-over-year , with guidance for more than $108 billion next quarter, he described the stock as grossly undervalued on paper, pointing to the low P/E ratio for a fast- growing company that is currently generating monopoly profits.
So, a person with over a fifth of their portfolio in short positions looked at these results and decided that the stock looked cheap. This is also not a contradiction. This is the tension on which all trade is built . His " bearish" scenario has three pillars, and the third one is what actually drives the position.
Pillar one : demand can be cyclical. What is worrying is that a significant portion of AI spending is funded, directly or indirectly, by the same companies that sell the hardware. Chipmakers are investing in AI startups and cloud providers. Then these companies buy the chips. Income is increasing. The demand seems bottomless.
Imagine a bartender who lends money to a customer so that he can continue to buy drinks. Today's earnings look fantastic . The problem only arises when lending stops. Now this is a criticism of these companies, not something they agree with. Nvidia's Jensen Huang dismissed the accusations of cyclical financing, calling the idea absurd and arguing that the amounts in question are only a small fraction of what these customers should raise anyway.
He has a strong argument. Keep both versions in mind. Second pillar: accounting. This is a more technical point, and it has created the most buzz. When a company buys servers, it doesn't write them all off as expenses at once. It spreads the costs over the expected life of the equipment. Stretch that period from 4 years to 6, and reported profits will increase without a single additional dollar of sales.
Berry argues that the big cloud companies have done just that with AI hardware, which is updated every two to three years. He called it one of the most common types of fraud today and estimated that they will understate depreciation by $176 billion between 2026 and 2028 . According to his calculations, by 2028 Oracle will overstate profits by 26.9%, and Meta will not lag behind it too much.
Be careful with these numbers. These are his predictions, not conclusions, not adjustments to reporting, and not what the regulator confirmed . Many analysts disagreed, arguing that older chips are transitioning to lighter tasks and indeed continue to generate profits for years. Treat this as a one-person model, clearly labeled as such— this is the third part, and it's the most important.
He believes that today's deficit will become tomorrow's surplus. He expressed himself as aptly as is possible. Real demand now, real shortage, real pricing power, and then—just a hint of future ghost towns. That's the whole thesis in five words. Golden cities are real as long as the gold rush lasts.
Buildings do not disappear when gold runs out. They just empty. There is a second bearish scenario that goes hand in hand with the first, and it is not about chips at all. It's about the " water supply" of the market. On August 4, 2026, the day the S&P 500 closed at a record high, Bury wrote that we may be close to a serious peak and possibly a 1987-like crash.
The mechanism he described is worth understanding, even if you've never played a downswing in your life. There is a huge pool of money managed by funds that target a fixed level of volatility. When markets calm down, regulations force them to increase leverage. Quiet market means more borrowing. More borrowing pushes prices up.
Rising prices further reduce volatility, which forces leverage to increase even more. As he said: “Lower volatility encourages volatility-focused funds and momentum strategies to increase leverage. This happens automatically. Nobody sits down and decides to do it. At that time, the volatility index closed at 15.86, right in the complacency zone.
Imagine a rubber band being stretched by a machine, pulling harder the quieter it gets around. It feels completely safe right up until the moment it stops. In October 1987, the Dow fell 22.6% in a single day—still the worst session in history, caused by just such mechanical forced selling in a market without buyers .
Pay attention to your word choice, however. Perhaps. Not " imminent", not " definitely", not " planned". This caveat plays a big role, and any honest review should take it into account. Okay, this is the part where everyone gets it wrong. So we'll do it slowly, with real numbers. First, one remark. The following is a simplified example built to demonstrate the mechanisms using rounded numbers.
This is an explanation of how the structure works, not a suggestion to build it. Options carry a real risk of losing everything you invest in them. Let's start with what a "short" actually is. You borrow shares, sell them today, and hope to buy them back cheaper later. Your profit is the difference.
Here's the problem that no one explains to beginners. The stock can only fall to zero. So the best scenario for a short position is 100 % profit. But there are no restrictions from above. Sell for $200, watch the price skyrocket to $600, and you will lose twice as much as you invested, and the losses will grow.
Unlimited risk on one hand, limited reward on the other. This is the deal you agree to when you " short". That's why professionals almost never leave it uninsured. Now about hedging. Let's say you sold 100 shares for $ 210. You received $21,000. If the stock drops to $ 100, you buy it back and pocket $ 11,000.
But you remain unprotected. So you spend a small amount on one call option with a strike price of $250—say, $15 per share, for a total of $1,500. This contract gives you the right to buy 100 shares for $250, regardless of what happens to the market. Two scenarios. Scenario one: you are right. The stock falls to $100.
Your call option depreciates in value and you lose the $1,500 you spent on it. But you make $ 11,000 on closing the short. Net you get about $ 9,500. Insurance cost you a portion of your profit. That's how insurance works. Scenario two: you made a mistake, seriously. Something happens and the stock price skyrockets to $400.
Without insurance, that's a loss of $19,000 on 100 shares, and if the price continues to rise, the losses will increase. But you have the right to buy for $250. So your short loss stops at $40 per share. Add the $15 paid for the contract, and your maximum loss on the entire deal is $55 per share. $5,500.
That's all. The stock could go up to $1,000, and that figure wouldn't change . That's the whole point. The call option did not make him optimistic about the market. He turned unlimited damage into known, predictable, and controllable. And now his words become clear. He is not playing for growth with these options.
He buys himself a "ceiling". That's why it's important for you, even if you've never played a downgrade in your life. The investor graveyard is full of people whose analysis was correct, but they rushed the time. Being premature and being wrong—looks the same on a broker's report. And if the position is large enough, it ends the same way.
Structure is what gives you time to discover who you really were. Now let's be fair to the other side, because there are a lot of them . In his own words, the short position in Nvidia is the only asset in his portfolio that is suffering losses. His bets against other companies worked. This one—no. Wall Street is overwhelmingly against him .
Before the latest report, 58 of 61 analysts tracking Nvidia recommended buying its shares. And the market reaction to these results was the opposite of what the "bear" would expect . On August 27, the stock jumped about 8%. Then there is the issue of capital. Berkshire Hathaway increased its stake in Google by about $17 billion last quarter.
One of the most valuation-disciplined institutions in the world is writing a huge check to a company that is in the midst of the same spending cycle it considers empty. There is a limitation that cannot be corrected by any justice. Time. Options have an expiration date . His Nvidia call options expire in December.
Its put options expire in 2027. Every day, a little bit of the premium evaporates, regardless of whether the market moves or not . If Nvidia simply moves sideways—without a crash, without a glut, without a 1987—the structure is exhausted, and the thesis doesn't even pass muster. This is what I find really interesting.
He didn't just make the statement, he set a deadline for it and paid for the privilege. And notice what this does to his arguments. Most market predictions cost nothing to those who make them. They may be vague, consistently premature, or quietly withhold their prediction if it doesn't come true. A dated option position eliminates all of this .
It has an expiration date. There is a specific point at which it either worked or it didn't, and no amount of storytelling will fix the numbers. That doesn't mean he's right. Many expensive and precisely dated bets turned out to be wrong. But it's a useful filter for you when deciding whose alerts deserve your attention.
Ask what this warning cost the person giving it. If the answer is nothing, weigh it accordingly. So, let's put it all together . Michael Bury is betting both ways not because he's confused. He plays the bearish game with a small volume and has spent a clearly defined amount of money to make sure that even if he is wrong, he still stays afloat.
A short position expresses an opinion. Call options control the consequences of this belief being proven wrong. These are two different tasks, and the confusion between them is the only reason why it once seemed like a contradiction. Three things to remember. The loud numbers about famous investors are often the amount of risk, not the money spent, and the gap can reach a hundred times.
Check what you are looking at before you react. Confidence and risk management are not opposites. Those who survive, despite premature predictions, are people who have decided in advance what a mistake will cost them . And the most boring advice, which is also the most useful. You don't have to take anyone's side in this argument.
You don't have to play for Nvidia's downgrade or defend it. You need a position size that you can hold even if you are wrong for longer than you expected, because that—much more than the quality of your analysis—determines how it will all end for you. The gold in the hills is real. Just like ghost towns.
The only important question is whether you have built something that will survive after you find out which one you are in .
Comments 0