Will The AI Bubble Burst in 2026?

It has not burst yet, and no one can name the day it will. But a growing list of respected investors now argue the AI boom looks like a bubble, and that 2026 is the year it could crack. The loudest voices are not anonymous bears on X. They are Michael Burry, Ray Dalio, Bill Gurley and Jamie Dimon, and they are getting harder to ignore.

These are not people who hate technology. Most of them say AI is real and will reshape the economy. Their warning is narrower and more uncomfortable for investors: spending has far outpaced profits, and the math only works if the buildout continues at full speed.

Why does Burry keep comparing Nvidia to Cisco?

Michael Burry (@michaeljburry), the investor known for shorting the 2008 housing market, has spent months building the most detailed bear case of the group. His core claim is that Nvidia is the Cisco of this cycle. In his May 24 Substack post, he wrote that there is once again a "Cisco at the center of it all," and named Nvidia as that company.

His point is about history. Cisco sold the gear that wired the late-1990s internet boom. Its stock rose around 3,800% from 1995 to 2000, then fell more than 80% and took two decades to recover. Burry thinks Nvidia is in a similar spot, a real company selling real picks and shovels during a buildout phase that will not last forever.

He has numbers behind it. By his count, the top 10 AI stocks have climbed 784% over 12 months, against 622% for the comparable group before the dot-com crash. The S&P 500 Shiller price-to-earnings ratio sat above 40 in May 2026, the highest since 2000 and more than double its long-run average of ~17. Burry is not just talking. He is leveraged short the Philadelphia Semiconductor Index through January 2027 put options, and holds bearish bets against Nvidia and Palantir. In early June he also called Nvidia's chip arrangement tied to Elon Musk's SpaceX a "fugazi." His worry is circular financing, where Nvidia helps fund the same companies that then buy its chips, so the money loops back and inflates reported demand rather than reflecting fresh, independent buyers.

What does Dalio mean by converting wealth into money?

Ray Dalio (@RayDalio), founder of Bridgewater Associates, brought the macro version of the argument on June 3 in a Bloomberg Television interview. He said every major technology shift produces a bubble because companies are forced to overspend to grab market share before anyone knows who will win. AI, he said, is following that path.

His most useful line was about how bubbles actually pop. "The pricking is the converting of wealth into money," Dalio said. A startup can be valued at a billion dollars after raising only a fraction of that amount, but valuation is not cash. When investors are forced to sell to cover debt, taxes, or fund withdrawals, paper wealth is matched against a much smaller pool of real money, and prices fall. He flagged Federal Reserve tightening, a stressed bond market and a possible hit to Taiwan's chip exports as triggers worth watching.

Dalio's firm put hard figures on the spending. Bridgewater estimates Alphabet, Amazon, Meta, and Microsoft could pour roughly $650 billion into AI infrastructure in 2026, up from about $410 billion in 2025. Like Burry, he separates the technology from the trade. AI will endure, but the investment bubble around it is real.

How big is the spending, really?

The scale is the part both bulls and bears agree on. Gartner (@Gartner_inc) forecasts that worldwide AI spending will reach $2.52 trillion in 2026, a 44% jump from the year before, with infrastructure accounting for the largest share. The debate is about the gap underneath that number. Bulls say the return on investment arrives once the training phase gives way to everyday business use. Skeptics say that is exactly the bet that could break, because the buildout depends on a small group of hyperscalers ordering at a record pace. If even one pulls back, the orders ripple backward through chipmakers and suppliers, what Burry calls a bullwhip effect.

Why is the rush to go public a warning sign?

The clearest test of the bubble case is arriving this year, as the three most valuable private AI companies all line up to sell shares in the same window. To skeptics, three giant offerings landing at once reads like a textbook late-cycle signal.

• OpenAI (@OpenAI) closed a $122 billion round on March 31 at an $852 billion valuation, the largest private financing on record, and is preparing a listing that could value it above $1 trillion. The catch is the income statement. The company was running near a $25 billion annualized revenue rate early in the year, but is on track to lose about $14 billion in 2026 and does not expect to turn a profit until around 2030. Its reported IPO target works out to more than 75 times 2025 revenue.
• SpaceX (@SpaceX) is even larger. After merging with Elon Musk's xAI in February, the combined company was valued at $1.25 trillion, and its IPO filing targets a valuation of roughly $1.75 trillion to $2 trillion while seeking to raise up to $75 billion. Its rocket and Starlink units make money on their own. The AI arm, SpaceXAI, is the part pulling the group into the red, and the filing implies a price near 95 to 107 times revenue.
• Anthropic (@AnthropicAI) is the outlier. The Claude maker closed a $65 billion round on May 28 at a $965 billion valuation, passing OpenAI, and then filed confidentially for its own IPO days later. Unlike the other two, it reported a revenue run rate of around $47 billion and says it expects to be profitable this year. Even so, its valuation nearly tripled in about three months, a jump critics read as more of the same speculative heat.

Who else is sounding the alarm?

The headliners have company. Bill Gurley (@bgurley), the Benchmark partner who backed Uber, said in March that a reset is coming. He noted the AI boom helped make the world's 500 richest people $2.2 trillion wealthier in 2025, and warned the spending cannot last. "One day, I just think we trip and run out of money," he said, adding that investors should buy beaten-down software stocks after the shakeout.

Jamie Dimon, the JPMorgan Chase (@jpmorgan) chief executive, has been steadier but no less pointed. In a May Bloomberg interview he said markets show "too much exuberance," singling out AI and the infrastructure being built around it. Months earlier he put the odds of a market correction near 30% and said some of the money going into AI now will be wasted, even as the technology itself proves useful.

What would actually pop it?

The skeptics do not agree on timing, and none claims to know the date. What they share is a read on the structure: a narrow base of hyperscaler buyers, valuations built on future promises, and financing that increasingly loops between the same handful of companies. Cisco survived the dot-com crash. The shareholders who bought at the peak did not. That is the precedent that the critics keep forcing everyone to look at.

Sources

• Bloomberg Ray Dalio's June 3 television interview on the AI bubble and how it bursts.
• Fortune Michael Burry's Nvidia-as-Cisco comparison and dot-com parallels.
• TheStreet Burry's 784% vs 622% figures, Shiller P/E, and SOXX short position.
• 24/7 Wall St Burry's "fugazi" comment on the SpaceX-linked chip financing.
• Gartner Forecast of $2.52 trillion in worldwide AI spending for 2026.
• CNBC Anthropic's $965 billion valuation, the SpaceXAI merger value, and OpenAI's $852 billion mark.
• Euronews The three-way IPO race, including SpaceX's $1.25 trillion to $2 trillion targets.
• Investing.com Revenue-multiple analysis for the OpenAI and SpaceX listings.
• Fortune Bill Gurley on an AI reset and the $2.2 trillion wealth jump in 2025. 
• Fortune Jamie Dimon's "too much exuberance" warning on AI valuations.