Burry's Bubble Call Meets a 150-Year Valuation Record

Michael Burry has warned that today's AI boom resembles the most destructive bubble in recent memory, a claim given weight by a valuation indicator now close to breaking a 150-year record, as the S&P 500 tracker SPY closed at $776.34 on Aug. 14, 2026.

Michael Burry has spent two decades being asked to repeat the trick that made him famous. This time the target is artificial intelligence. The investor who positioned against the American housing market before it broke is drawing a direct line between today's AI boom and the most destructive bubble in recent memory — and, according to 24/7 Wall St, one long-run valuation indicator is close to breaking a 150-year record, which is the sort of data point that forces even sceptics to give the argument a hearing.

That is the whole of the claim, and it is worth separating what is asserted from what is demonstrated. Burry's warning is a warning. The valuation record is a measurement. The two are being presented together because they point the same direction, not because one causes the other.

What a 150-year record actually measures

Long-history valuation gauges exist because a single year's price-to-earnings ratio tells you almost nothing. Earnings swing with the business cycle; a recession can make stocks look expensive at the exact moment they are cheapest. The fix, pioneered in academic work on long-run equity returns, is to smooth earnings over a full cycle and compare today's price against that average, adjusted for inflation. Because reliable American equity data reaches back to the nineteenth century, these series can be run over roughly a century and a half — which is why a "150-year record" is even a coherent phrase in finance.

The important property of such measures is what they are good at and what they are not. Historically, extreme readings on long-run valuation gauges have been informative about the following decade's returns and close to useless about the following year's. A record reading is not a sell signal with a date attached. It is a statement that the price paid for a dollar of smoothed earnings has rarely, if ever, been higher — and therefore that the return earned on that dollar over a long holding period has rarely started from a worse place.

That distinction matters for how an investor should treat Burry's warning. A bubble call and a valuation extreme can both be correct and still leave the index higher a year from now.

The tape is not panicking

Nothing in the most recent session suggests the market shares the alarm. As of the last trade on Friday, Aug. 14, 2026, the SPDR S&P 500 ETF Trust (NYSEARCA: SPY) closed at $776.34, down 0.20% on the day from a prior close of $777.88, with a day range of $775.43 to $778.80. The Invesco QQQ Trust (NASDAQ: QQQ), the tech-heavy Nasdaq 100 vehicle that carries the largest concentration of AI-linked megacaps, finished at $731.07, off 0.14% from $732.07, inside a $728.32 to $734.39 band. The SPDR Dow Jones Industrial Average ETF (NYSEARCA: DIA) closed at $536.80, down 0.21% from $537.91.

Three benchmarks, three declines of roughly a fifth of a percent, and intraday ranges narrow enough to fit inside a normal morning's noise. That is the signature of a market drifting, not one repricing risk. Notably, the Nasdaq proxy fell slightly less than the broad-market and industrial trackers — the opposite of what you would expect on a day when investors were actively unwinding AI exposure.

Why the dot-com comparison is the one being made

The analogy Burry is reaching for is the obvious one. The late-1990s technology boom shared several features with the present: a genuine general-purpose technology, a plausible story about productivity, enormous capital spending justified by future demand, and a narrowing set of index leaders doing most of the heavy lifting. The bust that followed did not disprove the technology. It repriced the equity claims on it.

What the current cycle has that the last one lacked is cash flow. The largest AI spenders are, for the most part, profitable incumbents funding capital expenditure out of operating earnings rather than serial equity raises. Bulls treat that as the decisive difference. Bears counter that profitable companies can still overbuild, and that depreciation on hardware bought at the top has a way of arriving late and all at once.

Neither side of that argument is settled by a valuation record. What the record does is set the price of being wrong. If earnings compound into the multiple, an expensive market simply delivers mediocre returns for a while. If they do not, the drawdown starts from a level with little historical precedent to cushion it.

What would confirm or kill the thesis

Three things are worth watching, and none of them is a headline about Burry.

Breadth. If index gains keep narrowing to a handful of AI-linked names while the average stock stalls, the bubble framing gets stronger. Broadening participation weakens it.
Capital-expenditure guidance. The AI trade is ultimately a bet that spending converts into revenue. Any sign that the largest buyers are slowing orders or stretching depreciation schedules is the tell.
The gap between QQQ and SPY. Sustained underperformance by the Nasdaq tracker against the broad-market fund would mark the point at which the market started discounting the AI premium rather than paying it.

How to hold both ideas at once

Burry's record earns him attention; it does not confer timing. The honest reading of the current setup is that a well-known investor has made a directional argument, a long-run valuation series is at or near an extreme not seen in a century and a half of data, and the market as of its most recent close is behaving as though neither of those facts is news. All three statements are true simultaneously.

For long-horizon investors, the useful response to a valuation extreme is rarely a wholesale exit. It is a review of concentration — how much of a portfolio's outcome depends on the same handful of AI-exposed names that dominate both major index trackers — and a realistic expectation for returns from here. That is an unglamorous conclusion. Bubble calls usually are, right up until they are not.

Key facts

S&P 500 tracker (SPY): $776.34, -0.20%, last trade Aug. 14, 2026, 20:00 GMT
Nasdaq 100 tracker (QQQ): $731.07, -0.14%, same session
Dow 30 tracker (DIA): $536.80, -0.21%, same session
Indicator cited: A long-run valuation gauge close to breaking a 150-year record

Frequently asked questions

What exactly is Michael Burry warning about?

Burry is warning that the current equity market is in a bubble, and he is drawing a direct comparison between today's artificial intelligence boom and the most destructive bubble in recent memory. His credibility on the point rests on having anticipated an earlier financial crisis when very few other investors did.

What is the 150-year indicator being referenced?

The reference is to a long-run valuation measure built on roughly a century and a half of American equity data — the kind of gauge that smooths corporate earnings across a full economic cycle and compares the current index price against that inflation-adjusted average. It is reported to be close to breaking its record over that history.

Does a record valuation reading mean stocks will fall soon?

No. Historically, long-run valuation extremes have carried useful information about returns over the following decade and very little about the following twelve months. A record reading raises the odds of poor long-term returns from that starting point, but it does not supply a date, a trigger, or a magnitude for any decline.

How did the major indexes close most recently?

As of the last trade on Aug. 14, 2026, the S&P 500 tracker SPY closed at $776.34, down 0.20%. The Nasdaq 100 tracker QQQ closed at $731.07, down 0.14%. The Dow tracker DIA closed at $536.80, down 0.21%. All three moves were small and the intraday ranges were narrow.

How is the AI boom different from the dot-com bubble?

The clearest difference is cash flow. The largest AI spenders today are mostly established, profitable companies funding capital expenditure from operating earnings rather than repeated equity raises. Critics respond that profitable firms can still overbuild capacity, and that the cost of hardware bought at a cycle peak shows up in depreciation later.

What signals would confirm the bubble thesis?

Three things: narrowing market breadth, where index gains depend on fewer and fewer AI-linked names; any slowdown or stretching of capital-expenditure plans among the largest AI buyers; and sustained underperformance of the Nasdaq 100 tracker against the broad S&P 500 tracker, which would show the market discounting rather than paying an AI premium.

Sources

Michael Burry Warns of a Market Bubble. This Indicator Close to Breaking a 150-Year Record Says He May Be Right — 24/7 Wall St

Photo: Rafael Minguet Delgado · Pexels Licence — source