Will the AI Bubble Burst? What the Signs Actually Say

Nobody can tell you the date an AI bubble bursts, and anyone who claims to is guessing. What experts interviewed on Wealthion can offer instead is a clear read on the warning signs, the genuine disagreement about how serious they are, and what a correction would mean for a portfolio. As of September 2026, the debate is not whether AI is transformative, most agree it is, but whether the capital pouring into it has outrun what it can earn, and whether the unwind has already begun. Here is what the signs actually say, drawn from our full analysis of whether AI is a bubble.

Is the AI bubble already bursting?

One prominent voice says the process has started. Jesse Felder of The Felder Report argued on Wealthion in his August conversation that “I think it’s the beginning of the end,” pointing to the collapse of the leveraged momentum trade that had carried AI hardware stocks higher. His memorable framing is that the last buyers were the greatest of the fools in a greater-fool market, and now “there are no more fools” left to push prices higher. When the buyer of last resort is gone, he argues, price has nowhere to go but down, regardless of how good the underlying technology is. That is one expert’s read, not a consensus, but it captures why the question has moved from hypothetical to live: the character of the market changed when the momentum trade broke.

What are the warning signs experts watch?

Three signals come up repeatedly, and none requires predicting a date. The first is earnings quality. Felder argues that hyperscalers “have literally never reported lower quality earnings than they are today,” because reported profits stay high while free cash flow deteriorates under the weight of capital spending. The second is where the spending itself sits. Peter Boockvar of Bleakley Financial Group put it precisely on Wealthion in his July interview: “the bubble is clearly in the capex side, not necessarily the technology.” The risk, in this view, is not that AI fails but that the hundreds of billions spent building it cannot earn an adequate return. It is the distinction between a technology bubble and a spending bubble: the internet was real and transformative, yet the companies that overspent building it in 1999 still went bankrupt when the returns did not arrive on schedule. Boockvar’s point is that AI could follow the same pattern, world-changing technology and ruinous capital losses at the same time. The third is financing: much of the buildout is funded by debt and circular arrangements, which makes it sensitive to interest rates and to any wobble in the private AI giants that anchor the demand.

Is there a case that AI is not a bubble?

Yes, and it deserves equal weight. Michael Strain of the American Enterprise Institute offered the counterpoint on Wealthion in his July interview: “I am not worried that current AI companies are overvalued,” at least in aggregate, because corporate earnings have been genuinely strong. His concern is different, that government involvement could make the sector too big to fail and prevent a healthy shakeout, but on valuation itself he is not alarmed. Brett Rentmeester of WindRock Wealth adds nuance from his July interview: the AI story is “there’s this whole interconnected web” of real physical demand, which means even a valuation correction in the frontier names would not erase the underlying buildout. The honest picture is a real disagreement between serious people, which is itself a signal: consensus bubbles are rarer than confident predictions of them. When everyone already agrees a bubble exists, much of the risk is arguably priced in; it is the bubbles no one is debating that tend to do the most damage.

When will the AI bubble burst?

No expert on Wealthion offered a date, and that restraint is the point. Instead they name tripwires to watch: whether the private AI giants that anchor compute demand can raise the capital they need, whether hyperscaler free cash flow keeps deteriorating, and whether the bond market, covered in our debt analysis, keeps pushing the long-term interest rates that finance the buildout higher. A bubble bursts when the marginal buyer disappears and the financing tightens, not on a schedule. This is why the experts who study it talk in conditions rather than calendars: the setup can persist far longer than skeptics expect, then unwind faster than anyone positions for. Watching those conditions is more useful than any forecast, because it tells you the risk is rising without pretending to know the day.

What would an AI correction mean for the rest of the market?

More than tech investors might assume, because AI has become such a large share of index value and economic activity. A meaningful AI correction would pull down the broad indexes that most retirement accounts track, since a handful of AI-linked names now drive much of their value, the concentration risk covered in our analysis of whether the 60/40 is dead. It could also ripple into the real economy through the AI capital spending that has propped up growth. This is why several experts frame real assets, covered across our gold and commodities coverage, as diversification against exactly this scenario: assets whose value does not depend on AI valuations holding up. The takeaway is not to predict a crash, but to recognize that owning no individual AI stocks is not the same as having no exposure to an AI correction, because index funds carry that exposure whether an investor chose it or not.

FAQ: The AI Bubble in Brief

Will the AI bubble burst? No one can predict the timing. Jesse Felder argues the unwind has begun; Michael Strain argues valuations are not alarming. The honest answer is that the risk is real and contested, and the useful response is watching the warning signs rather than forecasting a date.

What are the signs of an AI bubble? Experts point to deteriorating earnings quality and free cash flow at hyperscalers, a capital-spending boom whose returns are unproven, and heavy debt financing that is sensitive to interest rates.

Is AI a bubble or is it different this time? Most experts interviewed on Wealthion agree the technology is real; the disagreement is about valuation and capital spending. Some see a capex bubble; others see strong earnings that justify prices. Both views are held by serious analysts.

When will the AI bubble burst? No date can be reliably forecast. The tripwires to watch are whether private AI firms can keep raising capital, whether hyperscaler cash flow keeps weakening, and whether financing costs keep rising.

How would an AI crash affect my portfolio? Because AI-linked names dominate the major indexes, a correction would hit broad-market funds that most retirement accounts hold, even for investors who own no individual AI stocks. It could also weigh on economic growth.

Which experts and interviews does this article reference? Wealthion interviews from July to August 2026: Jesse Felder on yields and the AI bubble, Peter Boockvar on the AI spending bubble, Michael Strain on the government’s AI bet, and Brett Rentmeester on AI’s $920 billion gamble.

Wealthion editorial content is for informational purposes only and is not investment advice. The views quoted belong to the named guests. If you want a professional read on how AI concentration risk affects your own portfolio, you can request a free portfolio review from an advisor who understands real assets at https://www.wealthion.com/advisors/.