Nvidia’s 1,047% AI surge is a dot‑com‑style red flag, warns crash‑calling legend

For AI builders, researchers, and startup founders banking on sky‑high valuations, Grantham’s warning is a reality check: even transformative technology can be priced for perfection. If the bubble bursts, it won’t just erase paper wealth—it could slash R&D budgets, shrink hiring, and force a painful re‑evaluation of which AI ventures actually have a path to profitability.

Veteran bubble watcher Jeremy Grantham is sounding the most dire alarm yet on the artificial intelligence stock boom, warning that the AI trade has entered the same speculative danger zone as the great market manias that preceded Japan's 1989 crash, the 2000 dot-com bust, and the 2007-2008 U.S. housing collapse. Appearing on the latest episode of Steven Bartlett's The Diary of a CEO podcast, the GMO Investments co-founder framed the AI frenzy as the clearest market peak signal he has seen since the dot-com era.

“

As of late June 2026, the stock trades around $193.75—a gain of 1,047%, or roughly 11.5 times its pre-AI boom level.

The numbers are staggering. AI poster child Nvidia, whose GPUs power most large language models, closed at just $16.89 on Nov. 30, 2022, shortly after ChatGPT's public launch. As of late June 2026, the stock trades around $193.75—a gain of 1,047%, or roughly 11.5 times its pre-AI boom level. This single stock's ascent has accounted for a disproportionate share of overall market gains, pulling mega‑cap tech, chipmakers, and growth funds to new heights. But Grantham sees not a sustainable revolution priced-in, but a speculative euphoria built on stretched narratives about future profits, addressable markets, and transformative potential that may take far longer to materialize than current valuations imply.

Grantham's track record grants his words unusual weight. He built his reputation not as a perma‑bear but as a disciplined historian of excess. He warned of the Japanese asset bubble before its 1989 peak, called the 2000 dot‑com top, and cautioned against the U.S. housing market years before the subprime implosion of 2007–2008. Each time, his early warnings were dismissed as the party roared on, only to be vindicated when the music stopped. "The market can stay irrational longer than you can stay solvent" is the caveat Grantham himself often quotes, and indeed his bearishness has sometimes arrived months or years before the actual turn—a nuance lost on those who treat his calls as a simple sell signal.

What makes this warning particularly pointed is the breadth of the economic damage Grantham envisions if the AI bubble bursts. He argues a crash would not be contained to tech portfolios; it would ricochet through wealth effects, consumer spending, corporate hiring, and the broader real economy. Nvidia's vertiginous rise has trapped within it the hopes not only of day traders but of retirement funds, sovereign wealth, and institutional mandates that have piled into the AI theme as a one‑way bet. A mean‑reverting move could spark margin calls, forced selling, and an evaporation of the paper wealth that has fueled a consumer spending tailwind since 2023.

Yet the AI story differs from past bubbles in meaningful ways. Unlike the dot‑com era, where many companies had no earnings, today's AI leaders—Nvidia foremost—are generating enormous real cash flows. Nvidia's data‑center revenue has grown more than 400% year‑over‑year in recent quarters, driven by hyperscaler build‑outs. The technology itself is undeniably transformative, already reshaping industries from drug discovery to software engineering. Grantham's critique is not that AI is worthless; it is that the market has extrapolated today's exponential growth far into the future, discounting the well‑established pattern that every technological infrastructure build‑out eventually confronts overcapacity, competition, or a lull in the adoption S‑curve.