New York, May 9: Michael Burry saw the 2008 mortgage collapse coming before almost anyone else on Wall Street. He built a trade around it, endured two years of investor fury, and ultimately returned $725 million to his clients when the system broke exactly as he predicted. Now, in the first week of May 2026, Michael Burry is sounding the alarm again. This time the target is not subprime mortgages. It is the artificial intelligence investment super cycle that has carried the Nasdaq to levels not seen since the dot com bubble peaked in March 2000.
- Michael Burry’s Numbers Behind the Dot-Com Bubble Warning
- Michael Burry Puts Real Money Behind the Bubble Warning
- Michael Burry on the Jensen Huang Moment
- Michael Burry’s Depreciation Problem Hiding Inside Big Tech Earnings
- Paul Tudor Jones Agrees With Michael Burry, But on a Different Timeline
- The Bull Case Against Michael Burry’s Warning
- What Founders Should Take From Michael Burry’s Dot-Com Bubble Warning
- 5 Frequently Asked Questions
On May 7 and May 8, 2026, Burry published a pair of posts on his Substack that shook financial media. He compared the current AI fueled rally directly to the final chaotic months before the dot com bubble burst, revealed he has been doubling down on short positions against the most celebrated stocks in the market, and backed every word with nearly $1.1 billion in notional options exposure. Michael Burry is not nervous. He is positioned.
Michael Burry’s Numbers Behind the Dot-Com Bubble Warning
In posts published across Substack and X this week, Michael Burry laid out a data comparison that is hard to argue against. The average returns of the top 10 performing Nasdaq 100 stocks in 1996 came in at 559 percent. That figure climbed to 622 percent in the year leading up to March 2000, the very peak of the dot com bubble. The average gains for the top 10 performers in the same index for the 12 months ended May 5, 2026 hit 784 percent, beating both dot com peak periods by a margin that no serious analyst can dismiss as noise.
The AI rally, judged by the performance of its top performers, has not merely approached dot com territory. It has already surpassed it.
The most extreme individual case Burry highlighted was SanDisk, trading under the ticker SNDK. During the dot com bubble years, Qualcomm held the record as the ultimate high flyer, boasting a peak rolling 52 week return of 2,620 percent. SanDisk has dwarfed that figure by surging an unprecedented 3,960 percent between May 2025 and May 2026, exceeding the dot com record by 1,300 basis points. Burry also noted, with deliberate irony, that SanDisk was already the second best performing stock back in 1999, rising 581 percent at that time. Same name. Two generations of speculative excess. The current one is more extreme by every measurable standard.
The Philadelphia Semiconductor Index has surged approximately 224 percent since the 90 Day Pause period. The S&P 500 is up 48 percent and the Nasdaq 100 has gained 69 percent over the same window. AI infrastructure stocks have outperformed the S&P 500 by 115 percent since December 2023, more than any other category, according to analysis from The Kobeissi Letter.
Michael Burry Puts Real Money Behind the Bubble Warning
Writing on Substack after a long drive spent listening to financial radio and television, Michael Burry was characteristically blunt. “Absolutely non stop AI. Nobody is talking about anything else all day,” he wrote, before adding, “Stocks are not up or down because of jobs or consumer sentiment. Feeling like the last months of the 1999 to 2000 bubble.”
That was not just commentary. It came with concrete positioning. Burry disclosed he has added to his Nvidia January 2027 puts at a strike price of $115 and March 2027 puts at a strike price of $125. He also added to his QQQ puts for January 2027 at a strike price of $550 and loaded up on SOXX puts at a strike price of $330 for the same expiry. These additions came while chip stocks were still climbing, with Intel, AMD, and SanDisk among the biggest recent gainers.
While Scion Asset Management’s filings do not include the exact premiums paid, the underlying shares tied to its put positions carried a combined notional value of nearly $1.1 billion. Burry paired his bubble commentary with a photo of Christian Bale portraying him in The Big Short. The theatrics are familiar. The stakes underneath them are not trivial.
In his May 8 Substack post, Burry shared a chart showing the Philadelphia Semiconductor Index up more than 10 percent in a single trading week, pushing its 2026 gains to 65 percent. Headlines credited a strong jobs report. Burry’s reading of the driver was considerably less benign.
Michael Burry on the Jensen Huang Moment
By way of explaining his decision to double down on Nvidia puts specifically, Michael Burry reposted an interview with Nvidia CEO Jensen Huang on X in mid April 2026, where Huang was pressed with hard questions by podcaster Dwarkesh Patel. “Dwarkesh nailed it,” Burry wrote. “Jensen squirmed and obfuscated. This was shocking to me. Every NVDA bull needs to watch this with an unbiased eye.”
Nvidia shares hit a peak price of $216.61 as of late April and have risen more than 73 percent in the last year alone. That is the context for Burry’s unease. When the CEO of a company priced for perfection cannot cleanly defend the demand narrative under direct questioning, it raises a question the market has chosen not to engage with seriously. Michael Burry engaged with it, and he did so with material capital on the line.
Michael Burry’s Depreciation Problem Hiding Inside Big Tech Earnings
Beyond the chart comparisons and options positioning, Michael Burry has raised a structural accounting concern that the mainstream financial press has largely underreported.
His analysis suggests that major technology companies may collectively understate depreciation by $176 billion between 2026 and 2028, presenting investors with an earnings picture that does not reflect operational reality. He has estimated that Oracle could be overstating earnings by as much as 62 percent, Amazon by 31 percent, and Meta by 30 percent through a specific accounting practice.
The mechanism is straightforward. Several large technology firms have quietly extended the depreciation schedules for their AI hardware from the traditional 4 to 5 year range to 5.5 years or longer. By keeping aging hardware on the books as productive beyond its realistic useful life, companies push costs into future periods and make current earnings appear materially stronger than they are. Michael Burry called this “one of the most common frauds in modern finance.”
If this analysis holds, the earnings that justify today’s valuations are themselves overstated. Stocks priced at 40 to 50 times forward earnings become stocks priced at 60 to 70 times real earnings once the accounting adjustments work through. When markets price that reality in, the correction does not tend to arrive gradually.
Paul Tudor Jones Agrees With Michael Burry, But on a Different Timeline
Michael Burry is not alone in drawing the dot com bubble parallel. Billionaire macro investor Paul Tudor Jones told CNBC’s Squawk Box this week that the current environment feels similar to 1999, roughly a year before technology shares peaked in early 2000, and estimated the rally could run for another year or two.
Both men see 1999. Jones believes there is still meaningful runway before the correction forces its way into the conversation. Burry’s options positioning, expiring in January and March of 2027, suggests he believes the window is considerably tighter. When contrarian thinkers and macro legends start citing the same historical reference point independently, the debate shifts from fringe concern to a risk that mainstream portfolio management can no longer responsibly ignore.
To be fair, and serious analysis requires it, the case against Burry’s dot com bubble warning is not without merit.
The Bull Case Against Michael Burry’s Warning
The companies at the center of today’s AI trade are not the dot com era’s most notorious casualties. Nvidia generated over $130 billion in revenue in its most recent fiscal year. Microsoft and Alphabet produce hundreds of billions annually between them. These are cash generating businesses, not paper valuations built on page view projections. Palantir CEO Alex Karp has publicly dismissed the bearish thesis, arguing that AI companies are generating real money and reshaping industries in ways that are already measurable.
Some AI stocks have reached price to sales ratios exceeding 150, a level far beyond the 30 to 40 range that served as a ceiling during the dot com bubble peak. The performance gap between the S&P 500 and the S&P 500 Equal Weight Index is at its highest since March 2000. These are real concentration risks that even bull case analysts acknowledge privately.
The fundamental question is never whether a transformative technology is real. It almost always is. The question is whether the price already accounts for every optimistic outcome simultaneously, with no room left for execution risk, regulatory pressure, competitive disruption, or the commoditization that has historically followed every major computing wave without exception.
What Founders Should Take From Michael Burry’s Dot-Com Bubble Warning
For the startup ecosystem, the practical implications extend well beyond public markets.
If the semiconductor complex corrects materially heading into 2027, the funding environment for AI adjacent startups will tighten in direct proportion. Venture funds that have marked up AI portfolios based on stretched public market comparables will face pressure to recalibrate. LP appetite for follow on commitments will cool. Bridge rounds that sustained 2024 and 2025 vintage companies will become harder to close on favorable terms.
That is not an argument to exit the space. It is an argument to build with the discipline that durable companies have always required. Treat unit economics as a survival condition. Manage burn with the seriousness of a company that cannot assume cheap capital is permanent. The founders who survive market dislocations are rarely those with the best technology alone. They are those who built sustainable operations during the window of easy money rather than assuming it would never close.
Michael Burry was early in 2005 before being definitively right in 2008. Markets have a documented ability to stay irrational longer than most short sellers can stay solvent. But the signals he is reading this week, return comparisons that have already exceeded dot com bubble levels, accounting concerns hiding inside Big Tech earnings, a SOX chart that mirrors 1996 to 2000 almost precisely, and a CEO who cannot comfortably defend his company’s demand story, collectively add up to something that deserves serious attention from anyone with meaningful exposure to the most crowded trade in a generation.
5 Frequently Asked Questions
- What exactly did Michael Burry say about the dot com bubble in May 2026?
On May 7 and May 8, 2026, Michael Burry published posts on Substack drawing a direct comparison between the AI rally and the dot com bubble. He shared data showing the top 10 Nasdaq performers in the past year have already exceeded dot com peak returns and stated plainly the market feels like “the last months of the 1999 to 2000 bubble,” as reported by CNBC.
- What short positions has Michael Burry taken against AI stocks?
Michael Burry has added Nvidia January 2027 puts at a $115 strike and March 2027 puts at a $125 strike. He has also added QQQ January 2027 puts at a $550 strike and SOXX January 2027 puts at a $330 strike. The combined notional value of Scion’s put positions is reported at nearly $1.1 billion, according to Stocktwits and IBTimes UK.
- Why does Michael Burry consider SanDisk a bubble warning signal?
During the dot com bubble, Qualcomm held the record Nasdaq 100 52 week return at 2,620 percent. SanDisk surged 3,960 percent between May 2025 and May 2026, exceeding that dot com record by 1,300 basis points. Michael Burry’s point is that gains of this magnitude have historically appeared only at the peak of speculative cycles, not at their beginning.
- What is the depreciation fraud argument Michael Burry is making against Big Tech?
Michael Burry has argued that major technology companies including Oracle, Amazon, and Meta have extended AI hardware depreciation schedules from 4 to 5 years to 5.5 years or longer, collectively understating industry depreciation by $176 billion between 2026 and 2028. He estimates Oracle could be overstating earnings by 62 percent and Amazon by 31 percent, calling the practice one of the most common frauds in modern finance.
- How has Wall Street responded to Michael Burry’s dot com bubble warning?
The stocks Michael Burry is betting against have continued rising since his posts, consistent with his historical pattern of being early on major calls. Paul Tudor Jones echoed the dot com bubble comparison but believes the rally could extend for another year or two. Palantir CEO Alex Karp dismissed the bearish thesis publicly. Major investment bank analysts have not revised price targets in response to the warning.
Connect With Us On Social Media [ Facebook | Instagram | Twitter | LinkedIn ] To Get Real-Time Updates On The Market. Entrepreneurs’ Diaries Is Now Available On Telegram. Join Our Telegram Channel To Get Instant Updates.
Isabella is a global business journalist and former McKinsey analyst from Brazil. She brings sharp insights on economic shifts, policies, and founder journeys from around the world.
Tokyo-based CFA translating global markets into clear insights for modern entrepreneurs.
