원문 보기
Imagine you could look at the most powerful companies in the world and decide that the thing everyone thinks will make them richer might actually be the thing that breaks them. That is essentially what Michael Bur has been doing. The investor made famous by the big short became known for betting against the and now years later his attention has moved somewhere very different.
Artificial intelligence. In late 2025, Bur's Scion asset management disclosed put options tied to roughly 912 million of Palunteer and 187 million of Nvidia, more than 1.1 billion in combined notional exposure. But here is where the story gets interesting. That 1.1 billion number does not mean Bur simply wired 1.1 billion to a broker and bet it would disappear.
These were options. The notional value represented the underlying shares controlled by the contracts. the actual premium at risk was dramatically smaller. And that distinction matters because Bur wasn't simply betting that AI was worthless. He was betting that the price investors were paying for AI could become disconnected from the economic reality underneath it.
And in September 2026, Bur took the argument even further. He began shifting several bearish positions into longerdated put options, including positions tied to Nvidia, Palanteer, Micron, Nebius, Oracle, Caterpillar, the semiconductor ETF SOXX, and the NASDAQ 100. He also said his timeline for a potential AI bubble break had moved forward.
So, the real question isn't is Michael Bur betting against AI. The better question is, what does Bur think the market is getting wrong about AI? Because if his thesis is correct, the biggest risk isn't that artificial intelligence fails. It's that AI succeeds while investors still lose money. And that distinction takes us to the first piece of the puzzle.
Point one, Bur isn't betting against AI. He's betting against the price of AI. Here's the mistake investors make when they hear the words, "Michael Bur is shorting AI." They imagine a man sitting in a dark room yelling at a computer screen, "AI is fake." That's not really the argument. Bur's concern is much more subtle.
A great technology can produce terrible investments if the price becomes too high. Think about the internet. The internet changed the world. But that didn't mean every internet stock bought during the com boom was a good investment. Some companies had revolutionary technology and terrible economics.
Some had incredible stories and no sustainable business model. And some simply had valuations that required the future to go perfectly. That's the problem. Bur is focusing on AI can be transformative. Nvidia can be an extraordinary company. Palunteer can continue growing and investors can still lose money if expectations become impossible to satisfy.
That's the uncomfortable part because markets don't reward companies simply for becoming successful. Markets reward investors based on the difference between what is already priced in and what actually happens. If investors expect perfection, excellence isn't enough. You need something better than perfection.
And this is where Bur's thinking becomes relevant to anyone under 30. Because this isn't just about Nvidia or Palunteer, it's about how you think about money. When everyone around you is excited about an asset, the natural question is, "How much can this make me?" Bur's question is different. How much of that success is already reflected in the price?
That single question can completely change the way you invest. But Bur believes there is an even bigger problem underneath the AI boom. And this is where the story gets much more uncomfortable. Point two, the AI boom requires an enormous amount of capital. AI doesn't just require software. It requires physical infrastructure, data centers, semiconductors, energy, networking equipment, cooling systems, cloud capacity, and enormous amounts of financing.
Bur has argued that the scale of AI investment is creating a capital cycle that could eventually become excessive. And here's the basic economic problem. When an industry becomes incredibly profitable, everyone wants to build more capacity. Competitors rush in. Capital floods toward the opportunity.
Companies spend aggressively because they don't want to be left behind. At first, this creates growth. Then something strange happens. Everyone starts building for the future. At the same time, suddenly the industry doesn't have a shortage of capacity. It has too much. We've seen versions of this pattern before.
oil, telecommunications, railroads, housing, the.com infrastructure boom, the technology changes. The economic pattern doesn't necessarily change. And Bur's argument is that AI could eventually experience the same capital cycle problem. Not because AI isn't useful, but because investors may build capacity faster than the economy can generate returns from it.
And that creates a dangerous question. What happens when companies have spent trillions preparing for demand that doesn't arrive fast enough? That's when the story can shift. Growth slows. Margins compress. Debt becomes harder to service. Depreciation becomes more important. And suddenly investors stop asking who is winning AI?
They start asking who actually earns enough money from all of this infrastructure to justify the spending. And that's the question Bur keeps coming back to. But there is another reason this matters. Because Bur isn't only looking at what companies are spending. He's questioning how investors are interpreting the financial results produced by that spending.
And that's where things get even more interesting. Point three, the accounting problem. One of Bur's most important concerns involves something that sounds painfully boring. Depreciation. Yes, the accounting topic most people skip after approximately 3 seconds. But stay with it because this could be one of the most important parts of his argument.
Companies purchase expensive equipment. That equipment is expected to generate economic value over several years. So instead of recognizing the entire cost immediately, accounting rules generally spread that cost over the assets useful life. That's depreciation. Simple enough. But here's the problem Bur has raised.
What if the useful life of AI hardware is shorter than companies assume? AI infrastructure evolves extremely quickly. A chip generation becomes outdated. A newer architecture arrives, companies upgrade, data centers change, computational requirements increase. If expensive equipment becomes economically obsolete faster than expected, then the accounting assumptions surrounding depreciation become extremely important.
And Bur has argued that overly optimistic assumptions could make current earnings look stronger than the underlying economics really are. This is not the same as saying companies are necessarily committing accounting fraud. It's about assumptions. And assumptions matter because the market doesn't value companies based purely on today's cash.
It values them based on expectations about future cash flows. If those expectations change, valuations can change very quickly. And that's the pattern you should remember. Technology creates opportunity. Capital creates expansion. Accounting determines how that expansion appears on paper. And eventually cash flow determines whether the story survives.
But there is one more layer because even if Bur is correct about AI spending, there is still a massive obstacle standing in his way. The market. Point four. Being right is not the same as making money. This may be the most important lesson in the entire story. You can be right about an asset and still lose money.
Bur knows this better than almost anyone. In 2005 and 2006, he correctly identified weaknesses in the mortgage market. But being early was painful. His positions moved against him. Investors became frustrated and the trade looked wrong for a long time before it became spectacularly right. The same principle applies today.
In September 2026, Bur closed his December 2026 Nvidia and Palanteer puts, citing rapid time decay while retaining longerdated bearish exposure. He also said he was comfortable holding cash while watching the market develop. That's an important detail because options have an expiration date. You don't just need the market to eventually agree with you.
You need it to agree with you before the clock runs out. Imagine correctly predicting that a company is overpriced, but the stock stays expensive for another 2 years. Your analysis could eventually be correct. Your option could still expire worthless. That's the brutal mathematics of being early. And this is where Bur's current strategy becomes fascinating.
Instead of simply saying I'm short, he's changing the structure and timing of the trade. He has moved several bearish positions toward 2027 expirations as his research has changed his timeline. That's not just a prediction. It's a lesson in position construction because the market doesn't pay you for having an opinion.
It pays you for structuring risk correctly. And that leads directly to the question most people watching this video should actually be asking. nut. Should I short Nvidia? But what can I learn from someone who spends this much time thinking about being wrong? Point five. Bur's real bet is against certainty.
Look at the psychology surrounding AI. People aren't simply optimistic. They're becoming certain. Certain that AI will transform everything. Certain that today's leaders will dominate tomorrow. Certain that enormous infrastructure spending will generate enormous profits. certain that the next decade will look like the last two years.
And certainty is dangerous in investing because the future doesn't owe you consistency. The best investors aren't necessarily the people who know exactly what will happen. They're the people who understand what could make their thesis fail. That's why Bur's approach is so useful to study, even if you completely disagree with his bearish position.
You can believe AI will revolutionize the economy. You can believe Nvidia will remain dominant. You can believe Palanteer will keep growing and still ask, "What happens if I'm wrong?" That question changes everything. It changes how much you invest. It changes how diversified you remain. It changes whether you chase a stock after a huge rally.
And it changes whether you confuse a great company with a great investment because those are not the same thing. A great company can be an overpriced stock. Mediocre can occasionally become an attractive investment at the right price. Price matters. Expectations matter. Time matters and risk matters.
Which brings us to the final piece of Bur's current thesis. Because his latest move suggests he isn't simply predicting a crash. He's trying to position himself for a change in the capital cycle itself. Point 6. The real question is what breaks first? If Bur is right, what actually triggers the downturn?
It may not be one catastrophic event. It could be something much more ordinary. A company reduces AI spending. A major data center project gets delayed. Financing becomes more expensive. AI revenue growth disappoints. Hardware becomes obsolete faster than expected. Margins compress. Investors realize the expected returns aren't high enough.
One assumption changes, then another, and suddenly the narrative changes. This is how bubbles often behave. They don't necessarily collapse because everyone suddenly realizes the technology is useless. They collapse because people stop believing the future will be better than the price already assumes.
That's the key. And as of late September 2026, the market is still giving enormous weight to AI related growth. The NASDAQ had recently reached record levels even as Bur was moving parts of his bearish portfolio toward 2027 options. So there is a huge gap between Bur's thesis and the market's current optimism.
And that gap creates the opportunity or the danger because one of them is going to be early. Maybe Bur, maybe the market, maybe both sides are partially right. AI could genuinely transform the economy while some AI stocks become terrible investments. And that might actually be the most important possibility of all.
Point seven, the lesson for your money. You don't need to become Michael Bur. You don't need to short Nvidia. You don't need to buy puts. And you definitely don't need to make a billion dollar bet just because a famous investor did. In fact, copying famous investors is usually missing the point. The valuable part isn't the trade.
It's the thinking process behind the trade. Bur is forcing investors to separate three things: technology, business economics, stock valuation. They're connected, but they're not identical. AI can be revolutionary. Nvidia can be incredibly profitable. Palenterer can continue expanding and a stock can still be priced for unrealistic expectations.
That's why the smartest question you can ask isn't who is right. It's what does the current price assume? Then ask what happens if those assumptions are wrong. That's how you stop investing based on excitement. That's how you stop investing based on fear. And that's how you start thinking in probabilities instead of predictions.
Because wealth isn't built by predicting every crash. It's built by surviving the ones you don't see coming. Michael Bur's original $1.1 billion AI bet became famous because of the number. But the number isn't the most important part. The real story is what happens after the headline. He didn't simply place a massive bet and walk away.
He watched. He researched. He changed positions. He closed some options. He extended others. He moved risk. And as his timeline changed, his strategy changed with it. That is what sophisticated investing actually looks like. Not certainty, adaptation. Because the market doesn't care how confident you were 6 months ago.
It cares about what is true now. And if there is one lesson to take from Bur's latest move, it's this. Don't confuse a popular story with a guaranteed outcome. The AI revolution may continue. It may accelerate. It may disappoint. Some companies may become enormous winners. Others may discover that building the future is much more expensive than selling the dream of it.
You don't need to know exactly which outcome happens. You need a financial system strong enough to survive uncertainty. Because that's the difference between speculation and wealth building. And if you want to understand another investor who mastered the opposite side of this equation, knowing when to become aggressive instead of cautious, watch Dreen Miller's seven rules for knowing when to bet big.
I'll put it right here on the end screen because Michael Bur can teach you how to recognize when the crowd may be wrong. Ducken Miller can teach you something equally important. When the opportunity is real, how do you know when it's time to press the bet? Watch that video
Comments 0