#大空头加码做空AI芯片 Big Short investor Michael Burry makes his move, shorting Nvidia and warning that AI chip stocks could fall 30%
Being right about the direction does not mean being right about the price—what Burry is shorting this time has never been AI itself, but its overextended valuation
AI chip stocks could fall 30%.
This was not something a retail investor casually said, but a call from Michael Burry—the man who became famous for shorting subprime mortgages and served as the inspiration for The Big Short. The last time he made such a high-profile short bet was three years ago, the year Lehman collapsed.
The moment he showed his hand, the market panicked
On July 3, Burry publicly shorted Nvidia and also bet against chip ETFs, saying that “AI chip stocks could pull back 30%.” After the news broke, the Philadelphia Semiconductor Index fell 12% in two days, while Nvidia dropped 16% over the same period. Yet in the previous quarter, these companies had collectively added about $2 trillion in market value—on one side, the strongest gains in history; on the other, the most aggressive shorting in history.
This was no coincidence, but more like a signal before a reckoning.
It is worth noting that Burry’s view comes from someone who once correctly bet against a once-in-a-century bubble, but he has also made multiple bearish bets since then that failed.
The market’s strong reaction to his words was less about believing his conclusion than about the fact that many people had already been on edge—the rally had gone too far and lasted too long, and everyone wanted to be the one who exited early.
He is not shorting AI, but AI’s valuation
Many people will misread this. Burry did not say AI is a scam. His actual view is this: AI may be real, but stock prices have already discounted several years of future delivery into today’s valuations.
In other words, he is not shorting the technology itself, but prices that have been overextended. And the market has actually already begun adjusting downward.
There are reports that Meta may cut orders and that multiple manufacturers are reducing their orders. The signal being transmitted through the chain is that supply is catching up with demand.
The core narrative supporting valuations over the past two years was that “computing power is in short supply.” Once that narrative begins to weaken, the entire valuation chain—from chips to cloud services to applications—will be repriced.
This script is nothing new.
Looking back at the Internet bubble of 2000, the belief that “the Internet would change the world” was completely correct, but that did not prevent related stocks from losing 80%–90% of their value over the following two years.
Being right about the direction does not mean being right about the price; having a story does not mean the current valuation is reasonable.
What truly deserves attention is that this AI boom is beginning to seriously judge itself for the first time. When an inflated market capitalization is taken for granted as reflecting “perpetual growth,” risk is never hidden in the fundamentals, but in your imagination—the price you are willing to pay for “it will definitely be realized in the future” determines how large a drawdown you will have to endure.
Do you think this round of AI chip growth is real, or is it a valuation bubble? Will it really pull back 30%? Share your thoughts in the comments.
$NVDA
