Michael Burry says the stock market is stuck in denial, the first stage of grief, and he expects that phase to run another six to nine months before the AI bubble finally pops. The investor, known for his bets against the housing market before 2008, has moved from short positions to put options on Micron, Nvidia, and Palantir as of late September 2026, according to his own disclosures.
The warning lands as the Nasdaq Composite climbed 1.05% to 27,477 on October 5, a day after touching an intraday record. Underneath that headline number, the market looks far weaker: 75% of S&P 500 stocks fell in September even though the index closed the month slightly higher.
What Burry means by 'denial'
Burry is borrowing from the psychology of loss. Denial is the first stage, and he argues that's where investors sit right now — unwilling to accept that the AI trade has run its course. He expects this stage to last six to nine months, which would push the reckoning into mid-2027 if he's right.
He points to two precedents. After the dot-com peak in March 2000, the S&P 500 eventually lost nearly 50%. In 2008, stocks rallied into May before Lehman Brothers collapsed in September and triggered the worst sell-off of the crisis. In both cases, early resilience gave way to deeper losses. Burry sees the same pattern forming now.
The macro backdrop isn't helping
The economy is sending mixed signals at best. US employers added just 29,000 jobs in September 2026 — far below the 84,000 forecast. Meanwhile, the 10-year Treasury yield is holding above 5%, which keeps borrowing costs elevated for companies and consumers alike.
That combination of weak hiring and expensive money would normally weigh on stocks. Instead, the major averages are near records. For Burry, that disconnect is the point: the market is pricing in a perfect AI future while the labor market and bond market tell a different story.
Why he wants OpenAI and Anthropic to stay private
Burry has argued that markets should fall hard enough to stop OpenAI and Anthropic from going public. He hasn't spelled out exactly how that would work, but the implication is that a sharp sell-off would close the IPO window and keep those companies private — denying public investors the chance to buy into the AI hype at what he considers inflated valuations.
Anthropic's planned IPO looms as a key test. If Burry's timeline holds, the denial stage would end by mid-2027, overlapping the June expiry of his Micron and Nebius puts and Anthropic's planned listing. That's a tight window, and it suggests Burry is betting on a specific collision between market sentiment and corporate action.
Not everyone is buying the bear case
Wedbush Securities analyst Dan Ives continues to back Nvidia as the main engine of the AI rally, directly opposing Burry's warning. Ives has consistently argued that the AI infrastructure buildout has years of runway left, and that Nvidia's chips remain the bottleneck everyone needs.
The split between Burry and Ives is a clean one: Burry sees a bubble about to burst, Ives sees a revolution still in its early innings. Both can point to the same data — record highs, narrow breadth, strong AI demand — and draw opposite conclusions.
What to watch before mid-2027
The next few months will test Burry's thesis. If the denial phase really does run six to nine months, the market should keep grinding higher or at least hold steady into early 2027, even as fewer stocks participate. A narrowing rally — like September's, where three-quarters of S&P 500 names fell while the index rose — is exactly the kind of setup that precedes a reversal.
Burry's put options on Micron and Nebius expire in June 2027. Anthropic's IPO is expected around the same time. If stocks are still near records then, Burry's warning will look early at best. If the AI trade has cracked, his timeline will look prescient. Either way, the next two quarters should reveal whether the market is still in denial — or finally moving on.
