Michael Burry, the investor whose bet against the US housing market became famous through The Big Short, is bringing forward his timeline for a potential collapse in the artificial intelligence boom, warning that mounting debt, aggressive data-center spending and higher interest rates could expose weaknesses across the AI investment chain.
In a Substack post on Monday, Burry said he was “moving timelines up” because he now believes “the bubble in AI may burst sooner than later.” He said he is “more confident than ever before” that his bearish thesis will play out over the next year, compared with his previous base case of 2028.
The change has also altered how Burry is positioning his trades. Rather than maintaining outright short positions, he said he has shifted toward put options, which can provide greater leverage if the underlying stocks fall sharply.
Register for the next Tekedia Mini-MBA.
Register for Tekedia AI in Business Masterclass.
Join Tekedia Capital Syndicate and co-invest in great global startups.
Burry said he replaced outright shorts with puts on Nvidia, Palantir, Micron, Nebius, Oracle, Caterpillar, the iShares Semiconductor ETF and the Nasdaq 100. The contracts expire between June and December next year. He has also closed his short position in CoreWeave and plans to purchase put options on the company when he considers their prices attractive. Separately, he said he bought new puts on MetLife.
The structure of the trades indicates the scale of the downside Burry is positioning for. He said his Nvidia puts expire next September and have strike prices in the mid-$100s, compared with Nvidia’s $229 closing price on Monday.
His Nasdaq 100 puts have strike prices in the $24,000s and expire next September. With the index around 30,300 at the time, the positioning implies that Burry is preparing for a decline of roughly 20% within a year. That is a trading position, not a forecast that the index will necessarily reach those levels. The value of the options will also depend on timing, volatility, and other factors.
Burry’s latest argument focuses less on AI enthusiasm itself and more on the financing structure supporting the industry’s enormous infrastructure expansion.
AI companies and their infrastructure partners are committing vast amounts of capital to data centers, computing capacity, networking equipment and chips. Burry argues that the economics become vulnerable if spending growth slows before those investments generate sufficient returns.
“If the spending stops or slows, it all comes apart,” he wrote, arguing that the cash financing the expansion is becoming “increasingly debt, with strings attached.”
The AI boom has increasingly moved beyond technology companies simply spending their own cash on research and development. The construction of data centers and associated power and computing infrastructure requires substantial external financing, creating exposure to borrowing costs and credit conditions.
Burry argues that higher interest rates could therefore pressure several layers of the AI ecosystem simultaneously.
“Higher rates stress every part of that chain,” he wrote, pointing to the role of private equity, private credit and insurers in financing AI infrastructure.
The argument creates a potential feedback loop. Higher rates increase the cost of financing data centers and computing capacity. More expensive capital can make infrastructure projects less attractive. Slower investment can then reduce demand for chips, cloud capacity and other AI infrastructure, potentially putting pressure on companies whose valuations assume sustained spending growth.
Burry has also warned that the major technology companies are showing signs of strain.
“There are signs of strain at each of the big hyperscalers,” he wrote in a separate post, saying that their public statements and regulatory filings contain clues about how much pressure they are facing.
“I think there are many ways these companies are starting to fray,” he added.
His broader thesis is that the extraordinary spending required to build the AI infrastructure boom must ultimately be supported by sustainable economic returns. If companies continue increasing capital expenditure but AI revenue and productivity gains fail to keep pace, the gap between investment and returns could become more difficult for investors to ignore.
The timing of Burry’s warning is notable because the market has so far absorbed a series of potential challenges without ending the AI rally. AI-related stocks have remained major drivers of the broader equity market even as Treasury yields have risen, inflation concerns have returned, and geopolitical risks have increased.
That resilience is one of the major complications for Burry’s thesis.
A bubble can remain inflated longer than a bearish investor expects, particularly when companies at the center of the boom continue reporting strong revenue growth, and investors remain willing to finance large infrastructure projects.
Burry’s move from outright shorts to puts also highlights that timing risk. An outright short position can lose money as long as a stock rises, while an option can expire worthless if the expected decline does not occur before the contract’s expiration date.
His decision to use options therefore gives him greater potential leverage to a rapid decline, but it also makes the timing of his thesis more important.
Burry has been warning about excessive AI investment for some time. He has argued that major AI companies are masking slowing growth, spending excessively on chips and data centers, using accounting practices that flatter short-term earnings, issuing large amounts of stock to compensate employees, and entering financing arrangements that can reinforce the appearance of strong demand.
Those claims remain part of his investment thesis rather than established evidence that the AI industry is approaching a collapse. His current positioning nonetheless provides a clear window into what he considers the most striking vulnerability: the amount of capital being committed to AI infrastructure and the increasingly complex financing structure behind it.