Michael Burry, the investor famous for betting against the U.S. housing market before the 2007-2009 financial crisis, is moving up the timeline on his bearish thesis for the artificial intelligence trade. In his Monday investment newsletter, reported by CNBC on September 28, Burry said he is switching from short positions to put options on key AI-linked stocks — a move that gives him cheaper, shorter-horizon leverage if the sector turns.

"Fundamentally, I am moving timelines up," Burry wrote in the newsletter. "As such, I want more leverage in my short positions. Better timelines make leverage more palatable. Nothing says leverage like options, in this case put options, which are relatively cheap due to exceptionally tight volatility measures such as the VIX."

His warning lands amid a broader pullback in sentiment, with fresh analyst caution about AI spending and continued scrutiny of circular financing deals across the sector. You can follow all the latest AI developments and market coverage on AI Buzz Wire.

The New Positions, in Detail

According to CNBC's report of the newsletter, Burry's swaps cover four major AI-exposed tickers:

  • Micron (MU): his short position was replaced with put options at a June expiration date, in the $500 strike price range.
  • Nebius (NBIS): his short was swapped for puts at the same June expiration, in what he described as the "double digit strike price" range.
  • iShares Semiconductor ETF (SOXX): his short position was replaced with September 2027 puts "in the low $400s."
  • Palantir (PLTR): he "replaced and rolled the Palantir short and put position into an enlarged put position" centered on a September 2027 expiration in the low $100s.

Burry noted that some of the restructuring was intended to reduce his tax liability, but said most of it reflected a simpler view: "the bubble in AI may burst sooner than later." CNBC observed that the new put-heavy positioning suggests he believes the AI trade could flip by next summer.

Why Burry Thinks the Timeline Has Moved Up

The core of Burry's argument rests on research from Ares Management that he cited in the newsletter. The report emphasized the precarity of AI business models that rely on unproven revenues while being structured with demanding legal agreements.

"It would take only a season in which AI revenue disappoints the capital expenditure underwriting it," the Ares report stated, according to CNBC. "In that scenario, a handful of boards, predisposed to redeploy capital toward the highest-conviction bet, would simply need to conclude that the highest-conviction bet has shifted. The legal documents contemplate that decision."

In other words, the enormous compute commitments propping up the AI boom — many structured as irrevocable contracts — could unravel quickly if revenue growth disappoints for even a single quarter, because the same legal machinery that locks in spending also provides an exit for boards that change their minds.

Burry also pointed to comments from Acer CEO Jason Chen, who told Taiwanese media that cyclicality will return to the memory chip sector as Chinese production capacity increases. "How could there be a continuous shortage? China's production capacity has been consistently increasing, and there is absolutely no shortage issue," Chen was quoted as saying, adding that contract prices are currently fluctuating at high levels, with some rising and others falling. That is a direct challenge to the memory-supply squeeze narrative that has driven chip stocks higher.

A Long-Standing Bear Who Has Been Early

Burry is not new to this call. Earlier in September he had increased his short positions on Micron, Nebius, and the SOXX semiconductor ETF. Back in May, he said equities were "feeling like the last months of the 1999-2000 bubble," and he has a public history of large bearish bets against AI names, including a $1.1 billion put position against Nvidia and Palantir disclosed in late 2025.

So far, the market has not agreed with him. The Nasdaq Composite closed at a record high just last week. Even so, the rally has narrowed: Micron is about 16% below its record level, and Palantir trades roughly 10% below its all-time high — signs that some of the froth may already be coming off individual names even as the indexes climb.

What It Means for AI Investors

Burry's shift from shorts to puts is a change in tactic, not direction. Short positions bleed when the market grinds higher; puts define the maximum loss and pay off sharply if the sector falls within a fixed window. Moving to options means Burry wants to be paid on a timetable, not eventually — and his timetables now center on June 2027 for memory and cloud-infrastructure names, and September 2027 for software names like Palantir.

Skeptics will note that calling the top of a speculative boom has humbled better-timed prognosticators than Burry, and AI revenue reports over the next two quarters will do more to settle the debate than any newsletter. But with put pricing kept cheap by historically low volatility, the market is effectively selling crash insurance at a discount — and one of the most famous bears in modern finance is buying it in size.

Whether the bubble bursts in 2027 or deflates slowly, the terms of the debate have now shifted: the question is no longer whether AI capex is extraordinary, but whether AI revenue can keep underwriting it.

Stay Ahead of AI

Markets, models, and regulation are all moving at once. Follow the latest AI news, analysis, and breaking stories on AI Buzz Wire.

Read more AI news →