Ray Dalio and Jordi Visser agree that markets have a bubble. They just disagree on where it is.
Last week, Dalio, the founder of Bridgewater Associates, described artificial intelligence as a classic bubble.
"I think we're close to that," Dalio said of the bursting point at the Forbes Global CEO Conference in Singapore.
Visser, head of AI Macro Nexus Research at 22V Research, answered in a video published Sunday on his YouTube channel. He said the AI bubble is not the one investors should fear.
"How many times has Ray Dalio called for a bubble over the course of since 2020? A lot," Visser said.
Dalio's AI Bubble Case Rests on Debt and Rates
Dalio highlighted the large amount of debt being raised to fund AI investment as the main risk for the bubble's burst. As interest rates climb, that debt becomes harder to carry.
The 10-year Treasury yield closed Friday at 5.24%. That is up 108 basis points, or 1.08 percentage points, since the end of 2025. The 30-year yield stood at 5.60% and the 2-year yield at 4.79%.
Dalio also said the need to turn paper wealth into cash, for example, to pay a wealth tax, could force investors to sell.
Visser Says AI Demand Still Exceeds Supply
Visser said every bubble needs supply to run ahead of demand. That applies to speculative bubbles, capacity bubbles and credit bubbles.
In past cycles, people were the buyers. This time, he said, the buyers are AI agents. These are software programs that carry out tasks on their own and consume computing power as they work.
"The demand side is coming from agents," Visser said. He expects their number to grow from thousands today to billions next year.
He pointed to Lumentum Holdings Inc. (NASDAQ: LITE) as evidence. Chief Executive Michael Hurlston told Bloomberg Television last week that its optical components are effectively sold out through early 2029. The stock is up 191% this year.
Valuations are his second argument. Nvidia Corp. (NASDAQ: NVDA) trades at about 20 times expected earnings for its next fiscal year. Visser said that is nowhere near dot-com multiples.
The Bubble Visser Sees Is in Rate-Dependent Companies
Instead, Visser said the excess sits in older businesses that survived because borrowing costs stayed near zero for years.
"There is a bubble in companies that would have been gone a long time ago around the globe if it wasn't for the fact that rates kept them in business," he said.
He described them as "zombie companies that grow 2 to 3% a year in revenue."
The term usually refers to a company that earns too little to cover its interest costs. It stays alive by replacing old debt with new debt. When rates rise, each refinancing costs more.
Visser named utilities, consumer staples, household durables and autos as the most rate-sensitive areas.
What Is the Market Pricing In?
Market breadth supports part of that view. Market breadth, the number of stocks taking part in a rally, points the same way. One common test is whether a stock trades above its 200-day moving average, the average closing price of the past 200 sessions.
Fewer than half of S&P 500 stocks currently trade above the 200-day moving average, even with the index near a record. Only 4 of the 32 utilities in the S&P 500 pass that test. None of the four homebuilders do.
"That's why 400 stocks go down and 100 go higher," Visser said.
Returns show the same split. The median S&P 500 stock is up 3.5% this year. Meanwhile, the index is up 13.6%.
Only two sector funds are ahead of the index. State Street Energy Select Sector SPDR ETF (NYSE: XLE) has gained 45.5% and State Street Technology Select Sector SPDR ETF (NYSE: XLK) has gained 36.5%.
Polymarket traders put a 6.1% chance on an AI bubble burst in 2026, down 1.3 percentage points in a week. However, the odds rise to 18.5% for a burst by June 30, 2027.
