Imagine watching your portfolio shrink day after day while everyone around you gets rich. That’s Michael Burry’s life right now. The man who bet against the housing market and won is now betting against AI—and he’s getting slaughtered.
Burry, the real-life Big Short protagonist, started shorting AI stocks like Nvidia and Palantir last year. Since then, he’s expanded his bets to the entire AI supply chain, even shorting Caterpillar because bulldozers build data centers. His conviction? Absolute. His pain? Also absolute.
Shorting a bubble is easy. Surviving it is the real test of will.
Most people miss the uncomfortable truth: bubbles don’t end with a bang. They end with a slow, agonizing leak. The AI frenzy isn’t going to crash tomorrow. It’s going to keep inflating, testing the psychological endurance of everyone who dares to bet against it.
Let’s rewind to 2006. Burry was right about subprime mortgages. He saw the fraud, the leverage, the absurdity. But he was also bleeding cash. His investors demanded their money back. He had to sell positions at a loss—$40 million in realized losses in 2006 alone. Anyone who stuck with him through 2007 suffered two years of red ink before the crash finally vindicated them.
Now fast forward. Burry sees the same pattern: valuations decoupled from fundamentals, a market that ignores bad news, and a euphoria that feels eerily like the dot-com melt-up. He points to the Philadelphia Semiconductor Index climbing exactly like the Nasdaq did from November 1999 to March 2000. He’s not wrong about the bubble. He’s wrong about the timing.
Being right too early is indistinguishable from being wrong.
Here’s the twist that nobody wants to discuss: AI might not crash at all. The technology is iterating faster than any bubble in history. Compute costs are plummeting. Practical applications are emerging. The bubble could deflate slowly—not with a crash, but with a gentle hiss. That’s Burry’s nightmare. Years of holding the bag while the market proves him correct in the long run but bankrupts him in the short run.
What makes this so brutal is the asymmetry. When you’re short, losses are linear and gains are sudden. You can lose 50% over two years and then make 500% in two days—if you survive to see that day. But most people don’t. They capitulate at the bottom, right before the rebound.
Burry knows this. He’s been here before. He’s willing to endure the pain. But are you? The question isn’t whether AI is a bubble. It’s whether you have the stomach to lose money for years while everyone calls you an idiot.
In a recent filing, Burry wrote: “I am not just shorting because it is overvalued. I am shorting the business model. I am shorting the entire premise upon which the company rests. I am shorting the CEO.” That’s conviction. But conviction doesn’t pay the margin calls.
So here’s the real takeaway: If you’re tempted to short AI, don’t. The market can stay irrational longer than you can stay solvent. Instead, learn from Burry’s playbook—not his shorts, but his patience. Wait for the moment when everyone is convinced the party will never end. That’s when you strike. And only with money you can afford to lose for years.
The easiest way to get crushed in a bubble is to try to call the top.
Burry might be right. But being right doesn’t mean you’ll profit. It means you’ll be tested. And most people fail that test.
FAQ
Q: Is Michael Burry wrong about AI being a bubble?
A: He's likely right about the overvaluation. But bubbles can take years to pop. The danger isn't being wrong—it's being right too early. The AI ecosystem has real technological momentum that could sustain high valuations longer than any short can endure.
Q: What's the practical lesson for investors?
A: Don't short bubbles unless you have infinite capital and ice in your veins. The asymmetry is brutal: you bleed slowly while the market laughs, then you might get one shot at vindication. Most people break before that moment. If you must bet against mania, do it with options or after clear signs of a top—not during the melt-up.
Q: Could AI avoid a crash entirely?
A: Yes. Rapid iteration and falling compute costs might allow AI to grow into its valuation gradually. If the technology delivers commercial breakthroughs in the next few years, the 'bubble' becomes a rational premium. Burry's dot-com comparison is valid, but AI's iteration speed is unprecedented—that's the wildcard that could turn his bet into a decade-long grind.