Ray Dalio, the billionaire founder of Bridgewater Associates, has issued one of his most direct warnings about market froth, telling a podcast audience that AI-driven enthusiasm has inflated equity valuations to levels comparable to the crashes of 1929 and 2000. The warning lands at a moment when some of the largest IPOs in history are hitting the tape and capital spending on AI infrastructure has reached figures that would have seemed absurd even two years ago.

When the founder of the world’s largest hedge fund agrees that markets are in a bubble rivaling 1929, and then keeps buying AI stocks anyway, the signal is not just about valuations. It is about the distorted incentive structure that defines every late-cycle mania, and about why hard assets matter most when the music is loudest.

Appearing on The Diary of a CEO podcast hosted by Steven Bartlett, Dalio was asked about a prior guest’s claim that American markets face the biggest investment bubble in their history. Dalio’s response, as Fortune reported via Yahoo Finance, was blunt: “He’s right.”

That prior guest was Jeremy Grantham, the cofounder of GMO and one of the few institutional voices who called the Japanese bubble before its early-1990s collapse, flagged the dot-com mania before it burst, and wrote in Fortune in September 2007 that U.S. housing was in bubble territory, months before the financial system nearly came apart.

Related Posts
Copper hits all-time high as tariff fears drain supply

A Bubble Within a Bubble

Grantham’s framework, laid out in an April 2026 Fortune interview tied to his memoir The Making of a Permabear, describes what he calls a “bubble within a bubble.” The original super-bubble, in his telling, was already inflating dangerously by 2021. The S&P 500 then fell roughly 25% from January through October 2022, a decline that might have been the start of a proper unwinding.

Then ChatGPT arrived.

Grantham described the effect in vivid terms: “The day after Chat came out, the Mag Seven lifted the market on its broad shoulders and staggered forward.” In his view, AI did not resolve the prior overvaluation. It “deferred it while making it larger.”

A January 2026 paper coauthored by Grantham and financial historian Edward Chancellor found that the market’s price-to-book ratio and cyclically adjusted earnings multiples had reached extremes surpassed only in 1929, 1972, 1999, 2000, and 2021. Each of those prior episodes was followed by a severe correction. The pattern is not subtle, and it is not new. As we noted in our coverage of Grantham calling this the priciest U.S. market in history, the valuation data has been flashing red for some time.

The Four Horsemen Framework

The bubble question is not just about price levels. Owen Lamont of Acadian Asset Management has outlined what he calls the “Four Horsemen of the Bubble Apocalypse”, a checklist of conditions that historically precede the worst market unwindings. The four conditions: extreme overvaluation, widespread “bubble beliefs” among investors who know prices are too high but expect them to keep rising, a surge in equity issuance, and a flood of new market participants.

The current market appears to be checking those boxes. SpaceX has already gone public in what the Fortune article describes as the largest IPO ever. Anthropic and OpenAI are described as approaching trillion-dollar valuations. The IPO pipeline is open, the retail crowd is engaged, and the capital expenditure numbers border on the surreal.

That last point deserves emphasis. Yahoo Finance reported that Amazon, Alphabet, and Microsoft alone are set to spend $580 billion in capital expenditures in 2026, with the bulk directed toward AI infrastructure. That is a bet of historic proportions, not a rounding error, on a technology whose revenue model, for most companies, remains unproven at the scale required to justify the spending.

The valuations now embedded in U.S. equities echo patterns we have tracked in our analysis of stock market valuations hitting extremes only seen before the dot-com crash. The historical rhyme is hard to miss.

The Bubble Believer’s Dilemma

What makes this moment unusual is the self-awareness. Dalio is not simply warning from the sidelines. In a separate Bloomberg Television appearance, he acknowledged that AI is showing “typical signs” of a bubble pattern, consistent with how all great technological changes play out. Yet his own fund was aggressively buying AI stocks in the first quarter of the year.

This reflects the defining tension of every late-stage mania, not hypocrisy. Dalio himself described the dilemma facing corporate decision-makers: companies are being forced to choose between overspending on AI and the risk of losing out entirely by investing too little. That logic, the fear of being left behind, is precisely what Lamont’s “bubble beliefs” condition describes. Investors and executives alike know prices are stretched, but the perceived cost of stepping aside feels even higher.

“All great technological changes create bubbles and AI was showing the typical signs of this pattern.”

, Ray Dalio, Bloomberg Television

The parallel to the late 1990s is instructive. Plenty of smart money knew dot-com valuations were irrational. Many stayed invested anyway, because the penalty for being early was career-ending. The technology turned out to be real, the internet did change everything, but the valuations assigned to individual companies were not validated by actual cash flows for years, and in many cases, never.

That same dynamic appears to be playing out now. AI may well be transformational. But transformational technology and sound investment are not the same thing, and the gap between the two is where capital gets destroyed. As we explored in our look at why a 50% drop from late-1990s-style valuations would be historically normal, the math of mean reversion does not care about the quality of the underlying innovation.

Wealth Is Not the Same as Money

The headline lesson Dalio attached to his warning, “wealth is not the same as money”, cuts to the core of what metals investors already understand intuitively. In a credit-driven system, asset prices can rise for years on liquidity, leverage, and narrative momentum. The nominal wealth on a brokerage statement can look enormous. But when the bubble breaks, the distinction between paper wealth and durable purchasing power becomes painfully clear.

This is the lesson that gets relearned in every cycle. In 1929, in 2000, in 2008, each time, a generation of investors discovered that the wealth they thought they had was contingent on conditions that could not hold. The money was real only as long as the music kept playing.

For capital-preservation-minded investors, the signal is not necessarily to panic or to try to time the top. The signal is to understand what kind of wealth you hold and how it behaves when credit conditions tighten, when liquidity reverses, or when the narrative that justified extreme valuations simply runs out of new buyers.

Other veteran market observers have sounded similar alarms. Michael Burry has drawn direct comparisons between current conditions and the final months before the dot-com crash. The warnings are not coming from the fringes. They are coming from people who manage billions and who have been right before at exactly these kinds of inflection points.

What This Means for Gold

Gold tends to perform its most important function not during the bubble itself, but in the aftermath. When equity valuations reset violently, when credit stress emerges, and when policymakers respond with the only tools they have, rate cuts, liquidity injections, fiscal expansion, gold’s role as a monetary asset outside the credit system comes into sharp focus.

The current setup is worth watching carefully. If $580 billion in AI capital spending does not generate returns sufficient to justify the investment, the write-downs and margin compression that follow could ripple through corporate balance sheets, credit markets, and eventually the banking system. That is a description of how capital misallocation has resolved in every prior cycle, not a prediction.

The Nasdaq options market has already flashed warning signals not seen since 2008. Whether those signals prove prescient or premature, the risk profile of concentrated equity exposure at these valuations is not a matter of opinion. It is arithmetic.

For readers holding physical gold, silver, or positions in the metals complex, the Dalio warning is less about timing a market top and more about understanding the regime. In a world where the largest hedge fund on the planet acknowledges bubble conditions and keeps buying anyway, because the incentive structure demands it, the case for holding assets that do not depend on someone else’s balance sheet only gets stronger.

When the people running the money tell you the game is rigged but they cannot stop playing, that is a reason to own something that does not need the game to continue, not a reason to join them at the table.