Jeremy Grantham has been right about bubbles three times in a row. He saw the Japanese collapse coming in the late 1980s. He bet against the dotcom frenzy before the Nasdaq crashed 78% in 2000 and 2002. He warned about the US housing market before 2008. Each time, he was early, uncomfortable, and eventually correct.

Today, the 86-year-old co-founder of Boston investment firm GMO is saying it again: we are inside a bubble. Not a potential bubble, not an emerging concern - a confirmed one, by his definition, which has been consistent for 25 years. And in a new interview with Reuters Breaking Views, published today, he adds a second argument that is more interesting than the first. The AI spending race, he says, is not just inflating valuations to unsustainable levels. It is actively destroying the very thing that made Big Tech so profitable in the first place.

That is not a story about a bubble bursting. It is a story about an entire business model changing, probably permanently, whether the bubble bursts or not.

The Background

To understand why Grantham's argument matters, it helps to understand what made the largest US technology companies so extraordinarily valuable for the past two decades.

The Magnificent Seven - Microsoft, Apple, Alphabet, Amazon, Meta, Nvidia, and Tesla - did not get rich by competing fiercely with each other. They got rich by not having to. Each company dominated a separate market. Microsoft owned enterprise software. Google owned search. Amazon owned e-commerce and cloud services. Meta owned social networking. Apple owned the premium end of smartphones. There was overlap at the edges, but at the core, each firm was a monopoly - a company with no serious rival in its primary market.

A monopoly, in economic terms, is a company that can set its own prices rather than accept whatever the market dictates. Economists call this being a price setter rather than a price taker. If there is only one seller of something people need, that seller can charge more, spend less on competing, and harvest fat profits indefinitely. The result, over the period from roughly 2000 to 2025, was a steady transfer of wealth - from consumers and workers to corporate shareholders - that quietly reshaped the US economy.

Grantham is blunt about the systemic effects of this. During those 25 years, according to the Reuters interview, every industry in the US saw an increase in monopolistic tendencies. Fewer big players. Less competition. And as a consequence, slower GDP growth - because monopolies protect profits at the expense of productivity - alongside a widening gap between what ordinary workers earn and what shareholders accumulate. US GDP growth, he argues, has underperformed the 20 years before that, which themselves underperformed the 20 years before those. The era of easy monopoly money was very good for stock market returns and very mediocre for the economy at large.

AI is now changing that arrangement, and not in the way the market seems to believe.

What Is Actually Happening

The AI investment boom started, in Grantham's telling, in late October 2022, the month ChatGPT launched. Until then, a perfectly ordinary bubble - the kind that follows a decade of cheap money and rising tech valuations - had been deflating. The S&P 500 was down about 25% from its January 2022 peak. Growth stocks had fallen further. The Mag 7 had dropped collectively by around 35%. The market, he says in the interview, was heading toward a mild recession that would have completed the cycle and reset valuations.

Instead, ChatGPT arrived. Within weeks, AI investment began its escalation. Capital expenditure - the money large companies spend on infrastructure like data centers, chips, and computing power - surged from 2023 onward, strong enough to pull corporate earnings back up and keep the economy from sliding. Rather than finishing the correction, the market inflated a second, larger bubble on top of the first one.

The numbers are hard to grasp at full scale. According to Yahoo Finance's April 2026 earnings analysis, the four largest hyperscalers - Alphabet, Microsoft, Meta, and Amazon - are collectively on track to spend roughly $725 billion on AI infrastructure in 2026 alone. Microsoft expects to spend close to $190 billion this calendar year. Alphabet has guided to between $180 billion and $190 billion. Meta is budgeting $125 billion to $145 billion. Amazon's full-year target approaches $200 billion.

To put that in perspective: the combined GDP of Singapore and the United Arab Emirates is around $700 billion. These four companies are spending the equivalent of two mid-sized national economies, in a single year, on data centers and computing chips.

That spending has kept earnings growing and the market elevated. But Grantham's January 2026 paper, coauthored with financial historian Edward Chancellor and titled Valuing AI: Extreme Bubble, New Golden Era, or Both?, laid out why the current situation cannot resolve cleanly. According to Fortune, Grantham's answer to that question is mostly the first option. The AI boom deferred the original correction while making the eventual unwinding larger.

The SpaceX IPO, expected to price in mid-June at a valuation that has already been revised down from $2 trillion to approximately $1.8 trillion, offers a live example of the mood Grantham is describing. Investors are reportedly falling over themselves to buy in, regardless of governance concerns or the fact that the company's S-1 filing showed slowing growth and widening losses. It is, in the Reuters interview's framing, another case of an obviously compelling idea attracting too much money.

The Money Trail

Here is the part of Grantham's argument that goes beyond standard bubble-watching. It is not just that valuations are too high. It is that the companies at the center of those valuations are now actively competing away the profit margins that justified the valuations in the first place.

For years, the Mag 7 operated like a cartel without needing to coordinate. Each firm sat in its own lane. The cloud market - Amazon, Microsoft, and Google - was the partial exception, but even there, according to Grantham, the three firms managed it like a "well-behaved oligopoly," sharing the market and avoiding the kind of brutal price competition that would have destroyed margins. An oligopoly is a market with only a small number of large sellers, who are often able to keep prices high without explicitly colluding.

AI has ended that arrangement. Every one of these companies is now competing for the same prize: dominance in AI platforms, AI infrastructure, and AI applications. According to Fortune, Grantham recently described the new dynamic directly: "We have gone from a monopoly world to a brutal competitive world. And we will stay there for years and there will be blood in the streets."

The economic logic is straightforward. When a company holds a monopoly, it can set prices above costs and pocket the difference as profit. When many well-funded competitors enter the same market, they bid prices down to near the cost of production. The product might get better. Consumers might benefit. But the fat margins disappear. The firms that spent hundreds of billions building infrastructure end up competing on price to fill it.

Grantham draws a direct parallel to the railroad bubble of the 19th century. Multiple rail lines were built between the same cities. Everyone lost money. Eventually, one or two operators emerged from the wreckage and built real businesses - but the original investors who chased the obvious idea were largely wiped out. The same happened with the internet. Amazon fell 92% from its 1999 peak to its 2001 low, before eventually inheriting the world that the internet had promised. The technology was transformative. The investment frenzy around it was not a good way to make money.

AI will be transformative too, Grantham believes. But transformation does not automatically translate into profits for the companies spending most heavily to build it. In the Reuters interview, he makes the comparison to computers: when GMO bought an early computer in the 1970s, the firm had a two-year analytical advantage over smaller rivals. Then everyone had a computer. It became a cost of doing business. Competitive advantage competed away. "If you look out 10 years and then turn around and look back," he says, "everyone will have AI."

The implication is that massive infrastructure spending today may produce moderate returns at best - and losses at worst for the firms that overspent competing for a prize that gets shared rather than won.

The market is beginning to pick this up. According to Wealth Management, Microsoft, Amazon, Alphabet, and Meta are collectively spending an estimated 100% of their operating cash flow on capital expenditure in 2026, compared to a historical average of 40%. Free cash flow - the money left after investment spending, which is what investors actually own - is being squeezed or turning negative at several of these companies.

What People Are Doing About It

Financial institutions face a structural dilemma that Grantham describes with uncommon clarity: no big commercial firm can afford to tell its clients to get out of the market, because if they are wrong for even two years, they lose half their clients - and those clients rarely come back.

GMO itself lived through this during the dotcom boom. The firm went from managing $30 billion to $20 billion as clients abandoned it for managers riding the rally. When the crash came and competitors fell from $50 billion to $30 billion, GMO barely recovered from $20 to $22 billion. Vindication arrived later - assets grew from $22 billion to $165 billion in the four years after the crash - but the business risk of being right and early is severe enough that most large fund managers simply will not take it.

The result is a structural bias toward bullishness throughout the financial system. Independent managers - small firms without commercial clients to lose - can afford to be contrarian. Large banks and asset managers cannot. They stay invested, stay bullish, and move only when the market has already moved against them.

Individual investors are responding in ways that reflect both the excitement and the anxiety. The SpaceX IPO has reportedly reserved 30% of its float for retail investors - three times the typical allocation for a large offering - a deliberate strategy to maximize day-one demand. Retail platforms including Robinhood and Fidelity are expected to receive allocations. The structure is designed to create the appearance of broad enthusiasm, which validates the valuation on listing day.

Meanwhile, some institutional investors have been quietly rotating away from US AI stocks toward international equities. According to Fortune, GMO's emerging market fund returned approximately 70% over the past 12 months, compared to roughly 25% for the S&P 500. Grantham describes this as textbook mean reversion - the statistical tendency for extreme outliers to return toward long-term averages - playing out in real time as international stocks catch up from a period when US stocks were historically expensive relative to the rest of the world.

The Bottom Line

Jeremy Grantham has been calling this a bubble for years, and the market has kept rising. That does not make him wrong - it makes him early, which is a distinction that matters enormously to anyone who has to stay solvent while waiting. The more interesting claim, though, is not the valuation argument. It is the structural one. For two decades, the biggest US tech companies made extraordinary profits by not competing with each other. AI has forced them into direct competition for the first time, at a scale - collectively over $700 billion in capital spending in 2026 alone - that has never been seen before. Whether the bubble bursts or not, the monopoly era that produced those profits is almost certainly over.

Timeline

  • Late 1980s - Jeremy Grantham identifies and bets against the Japanese equity bubble, then trading at 65 times earnings, the most extreme valuation in any developed market in modern history.
  • 1999-2000 - GMO exits growth stocks ahead of the dotcom crash; the firm shrinks from $30 billion to $20 billion in assets under management as clients chase the rally, then grows back to $165 billion in the four years after the bust.
  • 2006-2008 - Grantham flags the US housing bubble before the financial crisis; the housing market, which he describes as a three-sigma event, later collapses symmetrically over three years.
  • January 15, 2022 - Grantham publishes a quarterly letter titled "Let the Wild Rumpus Begin," declaring that the market has checked off every indicator of a super-bubble ending.
  • January-October 2022 - The S&P 500 falls approximately 25%; growth stocks fall 35%; the Mag 7 drops around 35% collectively. The bond market records its worst year in modern history.
  • October 2022 - ChatGPT launches, triggering a sharp rally in AI-linked stocks and reversing the market's decline.
  • 2023-2025 - AI-linked capital expenditure by the Mag 7 escalates each year, pulling corporate earnings higher and preventing the recession Grantham expected.
  • January 2026 - Grantham and Edward Chancellor publish a paper titled Valuing AI: Extreme Bubble, New Golden Era, or Both?, concluding primarily that the current situation meets the definition of a bubble.
  • May 2026 - GMO's emerging market fund returns approximately 70% over 12 months, compared to 25% for the S&P 500, as international equities revert toward historical norms.
  • June 2, 2026 - Grantham tells Reuters Breaking Views that the Magnificent Seven have moved from monopoly dominance into direct competition, and that "there will be blood in the streets."
  • Mid-June 2026 (expected) - SpaceX IPO expected to price on Nasdaq; valuation target revised down from $2 trillion to approximately $1.8 trillion following S-1 disclosure of slowing growth and widening losses.

Summary

Who - Jeremy Grantham, co-founder of Boston investment firm GMO, and the firms comprising the Magnificent Seven (Microsoft, Alphabet, Amazon, Meta, Apple, Nvidia, Tesla).

What - Grantham argues in a Reuters Breaking Views interview that the US stock market is in a confirmed bubble driven by AI investment, and that the AI spending race has simultaneously destroyed the monopoly structure that generated Big Tech's extraordinary profits for the past two decades.

When - The interview published today, June 2, 2026. The underlying bubble, by Grantham's definition, has been inflating for four to five years, with a second wave beginning after ChatGPT launched in October 2022.

Where - Boston, Massachusetts, where Grantham spoke with Reuters Breaking Views host Peter Thal Larsen.

Why - Because the Magnificent Seven, previously dominant in separate markets, are now competing directly for AI supremacy - spending a combined $725 billion in capital expenditure in 2026 alone - in a fight that Grantham says will compress profit margins regardless of which company wins.