Michael Burry, the investor who called 2008, says the signs of an AI bubble are painfully obvious. Whether to sell everything is the wrong question.
Nobody can time the pop, including the people famous for calling the last one. The useful question is quieter: how much of your money already depends on AI spending never slowing down?
The short version
- Nobody can reliably time when the AI bubble bursts, so “when” is a question you cannot use.
- Your index fund, your pension and probably your insurer already hold AI exposure, whether you chose it or not.
- Check how concentrated you are and whether you could leave your investments alone for five years.
What did Burry actually warn about?
Michael Burry, the investor from The Big Short, posted an unplanned note in late September. He pointed readers to a map drawn by Ares Management, a firm that lends money for a living.
The map covers about $573 billion across 26 AI financing deals from the past year. Bonds, building loans, leases and chip backed borrowing connect Meta, Oracle, Microsoft, Amazon, Google, Nvidia, OpenAI and Anthropic.
Burry called it information that could not wait. He also said Wall Street is finally catching up to his view that the AI bubble is real.
Why does one web of deals matter?
Many of these deals come with guarantees. One company promises to cover the bill if another cannot pay. It is the same pattern as AI companies funding each other, now with lenders attached.
Think of friends who co-sign each other’s loans. It works well until two of them lose their jobs in the same month. Then every promise comes due together.
Every one of those promises assumes AI spending keeps growing. If revenue disappoints, the guarantees come due exactly when the companies behind them have the least cash to spare.
Where do insurers and chips come in?
Burry says much of this debt ends up with insurance companies. US and Bermuda insurers hold a little over $10 trillion to invest, by his count.
Some loans use Nvidia chips as collateral, the way a mortgage uses a house. Ares put the resale value of the H100 chip down 51% in three years. A house that loses half its value that fast makes a nervous lender.
Is the AI bubble about to burst?
Maybe. Nobody knows, and that includes Burry.
The honest answer has two parts. Something here looks fragile, and fragile things can stand for years.
The largest tech companies plan to spend hundreds of billions of dollars on AI this year. An AI bubble can keep inflating for as long as that money flows.
Being right and being on time are different things
Burry has bet against AI linked companies since 2025, so he profits if they fall. That does not make him wrong. It does mean his warning is also his position.
History is unkind to timing. In December 1996, the head of the US central bank warned about “irrational exuberance” in stocks. The market then more than doubled before it peaked in 2000.
A warning can be correct and still cost you years of growth. That is the trap inside most crash predictions.
What do past run ups tell you?
Take every US industry whose share prices doubled within two years. About half went on to crash, and about half did not.
Even the crashes arrived late. Prices often kept climbing for months first, so selling at the first warning meant missing a large gain.
Pip’s take
I have watched humans predict the end of things for nine hundred years. They are right eventually, which is not the same as being useful.
Why “when” is the wrong question about the AI bubble
Timing asks you to know the future. Exposure only asks you to read a list you already own.
That makes exposure the one part of this story you can check tonight.
It also takes the fear down a notch. An AI bubble you cannot predict is frightening, and a number in your own account is just a number.
You probably own the AI trade already
You may think you avoided all this because you never bought Nvidia. Look inside a plain S&P 500 fund, the kind that buys the 500 biggest US companies. The ten largest companies make up close to 40% of it, and most of them sit in that $573 billion web.
You do not have to pick AI stocks to be betting on AI. Your pension fund holds the same companies. If your insurer buys this debt, a slice of your premiums rides on it too.
The real risk is what you do during the fall
When the internet bubble burst in 2000, the tech heavy Nasdaq index fell almost 80%. It took 15 years to get back to its old high. The internet itself kept growing the whole time.
Both things can be true of AI. The technology can change everything while the stocks disappoint for a decade. Knowing how bubbles work helps you hold both ideas at once.
The fall hurt two groups most: people who needed the money soon, and people who sold near the bottom. Those who held a broad fund and kept waiting came out fine.
Frequent traders tend to trail the market badly. Across 66,000 household accounts, the busiest traders earned about 6.5 percentage points less per year than the market. Reacting to headlines is usually the expensive part.
What can you check this week instead?
You cannot audit a $573 billion web of loans. You can audit your own accounts in about twenty minutes.
Three questions cover most of what matters.
Three questions worth answering
First, how concentrated are you? Open your fund’s top ten holdings and add up the share in big tech. If one story drives most of your money, that is worth knowing.
Second, when do you need this money? Money you need within five years should not sit in stocks, bubble or no bubble. Money for 2050 can sit through a bad decade.
Third, could you cover six months of bills without selling anything? A cash buffer lets you ignore a crash instead of joining it.
Free · 2 minutes
Savings Goal Calculator
See how long it takes to build a cash buffer that lets you sit out a crash.
What you can safely ignore
You do not need Burry’s full breakdown when it lands. You do not need to guess the month, the trigger or the first company to wobble.
None of this means doing nothing forever. If the check shows too much riding on one story, sending new savings into a broader mix is a calm fix. Small, boring adjustments beat one dramatic exit.
If your answers to those three questions hold up, the AI bubble becomes something you watch. It stops being something you must act on.
Burry might be early, right, or both. Your plan should work in every version of that story. The pop is his question, and the cushion is yours.
FAQ
Is the AI bubble going to burst?
Nobody knows, including the investors warning about it. Past run ups of this size crashed about half the time, and the crashes often came months or years after the first warnings. The more useful question is how much of your own money depends on AI spending staying high.
What did Michael Burry say about the $573 billion AI web?
Burry pointed to an Ares Management map of about $573 billion across 26 AI financing deals from the past year. The deals tie the biggest AI companies together through bonds, loans, leases and guarantees. His concern is that every guarantee assumes AI spending keeps growing.
Do I own AI stocks if I only have an index fund?
Almost certainly. The ten largest companies make up close to 40% of a standard S&P 500 fund, and most of them are central to the AI buildout. Check your fund's top ten holdings to see your real exposure.
Resources
- Stocktwits. (2026). ‘Big Short’ Michael Burry drops ‘can’t wait info’ on $573B AI financing web. ‘Big Short’ Michael Burry Drops ‘Can’t-Wait Info’ On $573B AI Financing Web: What Happens If Spending Stops?
- Greenwood, R., Shleifer, A., & You, Y. (2019). Bubbles for Fama. Journal of Financial Economics, 131(1), 20-43. Bubbles for Fama
- Barber, B. M., & Odean, T. (2000). Trading is hazardous to your wealth: The common stock investment performance of individual investors. The Journal of Finance, 55(2), 773-806. Trading Is Hazardous to Your Wealth: The Common Stock Investment Performance of Individual Investors
- Greenspan, A. (1996). The challenge of central banking in a democratic society. Federal Reserve Board. Remarks by Chairman Alan Greenspan
- Kindleberger, C. P., & Aliber, R. Z. (2015). Manias, Panics, and Crashes: A History of Financial Crises (7th ed.). Palgrave Macmillan.
- Shiller, R. J. (2015). Irrational Exuberance (3rd ed.). Princeton University Press.
