Mark Cuban thinks the stock market is in a bubble, but he says he has no idea what will burst it. So the billionaire investor is doing what he says his heroes did: keeping cash on hand and waiting. "There's always going to be a bear market," he said. "And that's where the money was made."
Cuban built Broadcast.com, turned the Dallas Mavericks into NBA champions, was an investor on ABC's Shark Tank and now runs the online pharmacy Cost Plus Drugs. He laid out his case to host Steven Bartlett on The Diary of a CEO, in an episode published October 26, 2026. He started with stock prices that swing on a penny of earnings news, moved on to the borrowing behind AI data centers and finished with how he is protecting his own money.
Patience, learned the hard way
Bartlett brought up Charlie Munger and Warren Buffett, the longtime partners behind Berkshire Hathaway, and their obsession with compounding. What fascinated him, he said, was reading that there were stretches of a couple of years when they invested no money at all.
Cuban said he has tried to do the same, though not always. When markets fell during COVID, "I just started buying up everything when it hit and I rushed it," he said. With "AI related stuff just going to the moon and all this debt," his advice now is restraint. "You can invest a little bit for the fun of it, but be patient, have some cash because at some point it's going to turn," he said. "And when that happens, that's when the real magic happens and the real money is made."
A bear market is a long, broad fall in prices. In Cuban's view, it is the moment when buyers who have kept their cash can pick up shares cheaply.
Why he calls it a bubble
When Bartlett asked directly whether he thinks we're in a bubble, Cuban said yes, "but not the same as the internet bubble back in 2000, 2001." This time, he said, "everything's a meme stock." A meme stock is a share whose price is driven by crowd enthusiasm and fast trading more than by the company's underlying business. Cuban put Bitcoin, Micron, Google and Apple in that category.
His evidence was how violently prices react to small news. A company might miss earnings expectations by a penny, or beat them by only a penny, and its stock will be down 18%, he said, yet "the intrinsic value of that company hasn't changed 18%." Traders move in and out on anticipated information instead of "the net present value of the future cash flows," the standard way of valuing a business by what its future profits are worth in today's money. "We're trading them like baseball cards," Cuban said. "There's no intrinsic value to this. It's just perceived value. And what do you think somebody else will pay for it?"
The episode's opening clip cuts away just after the question of what will cause the bubble to burst. In the full exchange, Cuban asked it himself and answered: "I have no idea."
Waiting without a forecast has an old precedent. In Berkshire Hathaway's 1986 shareholder letter, Buffett wrote that bouts of fear and greed will keep recurring, but their timing, length and extent cannot be predicted. He also wrote that a shortage of attractive stocks is not a prediction about where the market is headed.
Winners, but which ones?
Bartlett said he was tempted to put a couple of dollars into NVIDIA, whose chips power much of the AI boom, "because you think maybe this is the last technology."
Cuban disagreed. "Something will always happen," he said, and he believes there will always be somebody smarter than AI. He pointed to Dell and Amazon, whose stocks "got crushed when the internet bubble burst." Amazon was "losing money hand over foot," and people gave it a hard time for that. Investors who stuck with it ended up holding a company worth trillions. "So there will be winners," Cuban said. "I just don't know which company. Is it Anthropic? Is it OpenAI? I don't know."
Borrowing to power AI
Cuban calls himself "the biggest proponent of AI that there is," but he argues that today's systems are "just brute force," needing "more power and more GPUs than the history of power and GPUs prior to last year," when to reach its potential AI should be doing more with less power.
When Bartlett suggested AI is going to get cheaper, Cuban agreed. It is getting cheaper, he said, "because those companies keep on borrowing and spending more money to get more power." He meant the hyperscalers, the handful of giant cloud companies that build and run the largest data centers. "We literally could get into a debt bubble in the United States of America," he said, "because there's so much debt being taken out by all these, you know, hyperscalers."
He described a chain of pressures. The federal deficit keeps rising, so the government has to borrow more. The hyperscalers "are borrowing more than they're making," and they are competing for lenders too. Rates keep climbing. "If you've got a $40 billion debt and your interest rates keeps on going up, at some point you can't raise money anymore," he said. "There's a lot of praying going on right now."
Cuban said he is "a huge fan" of NVIDIA and its chief executive, Jensen Huang, but that the chipmaker is "basically subsidizing half the industry because they have the cash to do it." He started to say that Meta was borrowing more than its cash flow by some multiple, then left the figure unfinished.
If AI were as brilliant as promised, he argued, the money should be flowing the other way. "At some point it may get there," he said, but until then the industry is in "this private credit debt environment where there's a lot of uncertainty and a lot of things can go wrong." Private credit means loans from investment funds and other lenders outside the banking system.
Others have worried about AI's debts from a different direction. In August, Alvin Graylin argued on Moonshots with Peter Diamandis that cheaper AI could undermine the income that is supposed to repay data-center borrowing. Cuban's case is about the wider economy: government deficits, interest rates and companies competing with the government for the same lenders.
A race against interest rates
Cuban also invoked the investor Ray Dalio. In Cuban's paraphrase, "Ray says, we've already crossed the curve": the United States is already past the point where it could become a debt-laden society whose currency eventually becomes worthless. In an October 13, 2026 essay, Dalio set out the mechanism behind that worry. Debt pressure builds when new borrowing exceeds what buyers are willing to absorb, and when interest payments squeeze out other spending. He tied this to government refinancing and to corporate borrowing for AI infrastructure. In his account, the pressure can ease through higher yields, less borrowing and spending, restructuring, or printing money that erodes the real value of debt. Strong enough growth in productivity and income can also improve the arithmetic.
That last possibility is the hope Bartlett raised. A lot of people, he said, see AI as the "life raft" out of the debt, pointing to talk of a coming age of abundance. "It's a race, right?" Cuban replied. The question is whether lenders will keep lending to AI companies as rates rise. By his count, the 10-year Treasury yield, the benchmark rate for U.S. government borrowing, was around 5.2%. He asked what happens if it goes to 6%, 8% or 10%. At current rates, he said, the country is "paying a trillion dollars in interest every year." If rates and government borrowing keep rising while the hyperscalers compete for the same money, "something gets crowded out."
He would not say what. "I don't know which or what," Cuban said. "I'm not enough of a financial expert to do that but something's going to give."
Bartlett said such a future was hard to imagine for America when you're 34 and have always seen American dominance. Cuban agreed that it is hard to imagine as an American, and said he is optimistic enough to think the country will figure it out, "but there'll be pain associated with that." He was less hopeful about Washington: "Our politicians are great at sloganeering, they're really bad at actually doing things."
How he is playing defense
Asked whether he is playing defense, Cuban said yes. He has "got money in this country and that country," and he hasn't invested nearly as much as usual. "It's the old Warren Buffett thing," he said. "I'll wait and see as things happen."
He accepts that waiting has a cost. "I might lose a little bit because I'm not as protected as I should have been," he said. "I want to give it some time to see how things play out."