Jeremy Grantham has been calling bubbles for half a century, and he has been right often enough that Wall Street cannot afford to ignore him. In a February 2025 interview with Bloomberg, the 86-year-old co-founder of Boston asset manager GMO declared flatly that “AI is a bubble,” ranking the current frenzy alongside the dot-com mania he famously identified before it collapsed. Now, more than a year later, the market has not cracked. But Grantham has not flinched, either.
A forecaster with receipts
Grantham’s credibility rests on a short list of very large calls. He warned clients about the Japanese asset bubble before it burst in 1989, steered GMO away from tech stocks before the Nasdaq cratered in 2000, and flagged housing-market excess well ahead of the 2008 financial crisis. GMO, which manages capital for pension funds, endowments, and sovereign wealth funds, built its brand on that willingness to be early and unpopular.
The record, though, comes with an asterisk. In January 2022, Grantham declared U.S. equities were trapped in a “superbubble” and predicted a severe unwind. Stocks did fall sharply that year, with the S&P 500 losing roughly 19%. But the index then surged through 2023 and 2024, driven largely by a handful of AI-linked megacaps. Investors who moved to cash on his word missed one of the strongest two-year rallies in modern market history.
That gap between identifying real danger and timing the exit is the central tension in every Grantham call. He tends to be right about the disease and early about the prognosis.
The IPO signal GMO is watching
Part of GMO’s case draws on patterns in U.S. initial public offerings. According to Bloomberg’s reporting, the firm circulated a chart titled “U.S. IPOs, 2000-2025” showing that listing activity has begun to echo the frenzied pace seen at prior market peaks. The underlying data draws on records maintained by the U.S. Securities and Exchange Commission, which publishes IPO statistics stretching back to the first quarter of 2000.
During the dot-com peak, IPO volume surged as companies with negligible revenue rushed to list. A similar acceleration in AI-adjacent offerings would, in Grantham’s framework, signal that speculative appetite has outrun fundamentals. The SEC data gives the argument a factual anchor that many market predictions lack, and anyone can cross-check GMO’s claims against the official record.
That said, IPO surges are not automatic crash indicators. New listings also climb during periods of genuine economic expansion. The pattern rhymes with history; it does not guarantee a repeat.
The AI trade in spring 2026
More than a year after Grantham’s warning, the AI trade has proven stubbornly durable. Nvidia, whose graphics processors remain the backbone of AI training infrastructure, reported record data-center revenue through its fiscal year 2026 earnings. Microsoft, Amazon, Alphabet, and Meta collectively earmarked hundreds of billions of dollars for AI-related capital expenditure across 2025 and into 2026, according to their public earnings disclosures and investor presentations.
Beneath the surface, though, the picture is messier. Dozens of smaller AI startups that went public or raised late-stage private funding during the 2024-2025 boom have struggled to convert hype into recurring revenue. The gap between AI leaders generating real cash flow and AI aspirants burning through capital has widened, creating what some analysts describe as a two-tier market rather than a uniform bubble.
The arrival of China-based DeepSeek in early 2025, which demonstrated competitive large-language-model performance at a fraction of the training cost reported by U.S. rivals, briefly rattled AI valuations and raised questions about whether the massive capital spending by American tech giants would deliver the returns investors expect. That episode faded quickly from headlines, but the underlying question it posed has not been answered.
Meanwhile, valuations remain stretched by historical standards. The S&P 500’s concentration in its largest constituents sits near levels not seen since the early 2000s, a point Grantham and other skeptics have repeatedly flagged. Whether that concentration reflects justified dominance by a few genuinely transformative companies or dangerous crowding into a single theme is the core disagreement dividing Wall Street as of spring 2026.
The bull case Grantham is betting against
Grantham’s critics argue that AI is fundamentally different from the dot-com era’s vaporware. Nvidia and Microsoft are posting tangible earnings growth, not just user counts. Enterprise adoption of large language models and AI-driven automation tools has accelerated across industries from healthcare to logistics, generating measurable productivity gains documented in corporate earnings calls.
Aswath Damodaran, a finance professor at New York University’s Stern School of Business and one of the most widely followed voices on valuation, has pushed back on the reflexive use of the word “bubble.” In his Musings on Markets blog and public lectures through 2025, Damodaran argued that some AI valuations, while aggressive, are supported by plausible growth scenarios, and that slapping the bubble label on a sector often shuts down analysis rather than sharpening it.
The counterargument does not require believing every AI stock is fairly priced. It simply holds that a broad bust is less likely when the sector’s largest companies are generating real profits at scale, unlike the telecom and internet firms of 1999 that were valued on projected revenues that never materialized.
Interest rates add another layer. The Federal Reserve’s policy path through 2025 and into 2026 has kept borrowing costs elevated relative to the near-zero environment that inflated asset prices in the prior decade. Higher rates typically compress speculative valuations, which could support Grantham’s thesis. But they have not yet triggered the kind of broad repricing he has predicted, partly because AI earnings growth has been strong enough to absorb the pressure.
What investors can actually do with this
For anyone with retirement savings parked in a tech-heavy index fund or concentrated positions in AI-linked stocks, Grantham’s warning raises a practical question: act, or wait?
His track record suggests he identifies real risks. His timing has been inconsistent enough to punish those who follow his calls with all-or-nothing portfolio moves. Selling everything based on a single forecast, no matter how credentialed the forecaster, is a strategy that has historically cost more in missed upside than it has saved in avoided losses.
A more measured approach, and the one most professional allocators tend to favor when confronted with credible bubble warnings, involves reviewing concentration. Investors whose portfolios are dominated by a handful of AI-linked names can ask whether that exposure matches their actual risk tolerance and time horizon. Trimming oversized positions, diversifying across sectors and geographies, and setting predefined rebalancing rules are ways to acknowledge the risk Grantham is flagging without betting everything on his prediction landing on schedule.
A prediction the market has not yet settled
The verifiable facts in this debate are narrower than either side might prefer. The SEC’s IPO data is real and publicly available. Grantham’s historical calls are documented. AI capital spending by major tech companies is reported in audited filings. Beyond those anchors, the question of whether the AI boom is a bubble or a lasting shift rests on assumptions about future earnings, adoption curves, and competitive dynamics that no dataset can resolve in advance.
Grantham is making a falsifiable prediction: that AI valuations will crack and investors who bought near the top will suffer significant losses. The market, through prices, is making the opposite bet. As of spring 2026, neither side has been definitively vindicated.
What is not uncertain is the value of stress-testing your own portfolio against a scenario where Grantham turns out to be right. If a 40% drawdown in AI-linked holdings would force you to change your retirement plans, that is worth knowing now, before the market decides the question for you.
