Burry likens next market crash to a super volcano and offers five signs for investors
The legendary short seller says a collapse in AI favorites could create opportunities in stocks beaten down before the crisis
The market resembles a brewing super volcano, Burry argues, with conditions appearing stable on the surface but imbalances brewing underneath / Photo: Shutterstock
A “super volcano” threatens Wall Street, argues Michael Burry, the famous investors who predicted the U.S. mortgage market crash almost two decades ago. Beneath the stock market’s apparent calm, he sees the threat of a crisis that could follow an unusual pattern: a decline in the benchmark S&P 500 accompanied by gains in stocks that have already suffered steep losses.
In early October, Burry warned of an imminent market crash and claimed investors were ignoring the warning signs. Now, after studying 10 market peaks over the last 40 years, including the potential peak on October 6, he has refined his forecast. In his Cassandra Unchained blog on Substack, Burry writes that rather than crashing abruptly, the S&P 500 could gradually drift lower, punctuated by steep selloffs. Financial problems in the AI industry would intensify the pressure, in his view. He does not claim to know exactly when the downturn will begin but believes the preconditions are already in place.
Seismic risks
The market resembles a brewing super volcano, Burry writes: conditions appear stable on the surface, but imbalances are building underneath. Above ground is a new S&P 500 record, driven by a handful of the largest companies; below are hundreds of stocks that have already lost 10-30% or more from their highs.
Only 15 stocks in the S&P 500 closed at a new high with the index on October 6, last week. A paltry 97 were within 10% of one. In March 2000, it was 15 and 87. That is some thin dirt.
Before the dot-com crash, the median S&P 500 stock was 31% below its high, versus 22% today, Burry writes. However, conditions are deteriorating quickly: since August, the share of stocks down more than 20% has risen from 47% to 55%. If the trend continues, market indicators could reach or even exceed their 2000 levels within months.
Thing is, we do not know where we are – perhaps we are at the beginning, perhaps we are four months in, perhaps we are seven months in, and perhaps it can blow any time.
The next bear market
Burry uses 2000 to illustrate how sharply the performance of today’s market leaders and already beaten-down stocks could diverge. In the 12 months following the S&P 500’s March peak, the former leaders suffered a median loss of 33%, while stocks already in bear markets gained 14%. After 18 months, the leaders’ losses reached 50%, while previously beaten-down stocks remained 8% above their March levels.
Burry believes the first signs of such a shift in leadership could emerge before the broader market begins to fall. In the year following the 1998 market peak, previously beaten-down stocks gained 7%, while market leaders rose just 4%, foreshadowing their relative performance after the dot-com bubble burst. He observed something similar in early 2025: following the release of an AI model by Chinese start-up DeepSeek, tech stocks fell sharply, while some previously beaten-down names delivered stronger returns.
This time, financial problems in the AI industry could put additional pressure on the largest companies. Among the warning signs Burry identifies are corporate earnings quality, off-balance-sheet liabilities, supply-chain risks, and reliance on outside financing. He also sees a potential weakening of the “mechanical bid” channeling money into the largest stocks, which currently supports the market, as another risk.
All this taken together with the other work I have done, the market will fail in a manner that lifts the median broken stock (the median stock among those down more than 20% before the crisis – Oninvest). This is my base case. It will perhaps not fail in a big crash, but will drift lower, sometimes violently, as fundamental news from the land of compression makes its way into the AI narrative.
Burry fears that a prolonged decline would be harder to arrest than previous crashes. Support measures that help when stock prices plunge could prove less effective if corporate fundamentals and market conditions deteriorate gradually, he argues. Moreover, previous crises have substantially reduced the U.S. government’s capacity to support the economy. If the private debt bubble bursts, the government would have a harder time cushioning the fallout.
Five rules of a bear market
Drawing on his study of 10 market peaks over the last 40 years, including the potential October 2026 peak, Burry identifies five key lessons about market reversals.
An index can keep setting records while most stocks are already falling. Gains in the largest companies can keep the S&P 500 at record highs even when more than half of its constituents are already deep in the red. Historically, this is an anomaly: by Burry’s calculations, at ordinary market peaks over the last 40 years, only 20-31% of stocks were more than 20% below their rolling 10-year highs. Today, that share is 55%, below March 2000’s 71.2%. These are the only two instances in his four-decade study when the index set records while a majority of its stocks were in bear markets by this measure.
A growing share of declining stocks is a warning sign. Burry fingers1995, when the S&P 500 reached a new high even though 46% of its stocks were down more than 20%. At that point, however, the share was steadily shrinking, allowing the market to continue its multiyear rally. Today, the situation is reversed: the share of stocks down more than 20% has already exceeded 50% and continues to expand alongside gains in the index. For Burry, this is the clearest sign of deteriorating market breadth.
A falling market does not necessarily mean all stocks will fall. Burry’s historical analysis shows that previously beaten-down stocks often continued to decline and, in several downturns, fell more than the market leaders over the following year. After the 2007 peak, for example, previously beaten-down stocks suffered a median loss of 45%, versus 38% for the leaders. However, Burry identified one exception: the 2000 dot-com crash. Twelve months after the market peak, the index’s leaders had plunged 33%, while the median “broken” stock had gained 14%. Even a steep S&P 500 drawdown can coincide with gains in neglected sectors.
The leaders of an overheated market can take years to recover. After the dot-com crash, many companies that had previously powered the S&P 500’s gains struggled to regain their highs for years. Some had still not recovered by the start of the 2008 financial crisis. “When the big bubbles pop, the leaders can go down and stay down a very long time. Japan may have been the biggest bubble of the modern era, and we all know how that went. Decades of futility for the largest,” Burry writes.
The market can go decades without delivering returns. After major crashes, the U.S. market has sometimes taken decades to regain its previous levels in inflation-adjusted terms. Burry cites three such periods: 1929-1949, 1966-1982, and 2000-2011. Reinvesting dividends substantially reduced losses, however, and in some cases allowed investors to earn positive returns even after inflation.
