An AI market crash could start somewhere investors aren't looking

An economist says that if the AI boom cracks, the trigger may come from outside the AI companies themselves. Dario Perkins, managing director of global macro at Global Data.TS Lombard, wrote in a note on Monday that investors may be overlooking weak spots beyond the big tech stocks. His point: the spark does not have to come from inside the boom.

Where the money comes from

Perkins argues that much of the industry's revenue growth still comes from companies spending within the AI world. It does not come mainly from businesses and everyday users paying for AI services. He calls this "capex recycling." Capex is money spent on big equipment and buildings, such as data centers.

His question is whether revenue is coming from outside the ecosystem, presumably to pay for tools that make workers more productive, or from that circular spending.

Why the worry is growing

The note follows a recent selloff in chip stocks. Last week the Philadelphia Semiconductor Index fell more than 11%. It then clawed back 2% on Monday. The drop has sharpened doubts about whether the huge sums being spent on AI infrastructure will pay off.

The weak spots he names

• Debt at smaller cloud providers, model developers and infrastructure suppliers
• Cash-rich giants choosing not to renew data center leases
• A chip glut if companies over-ordered to avoid shortages
• Higher interest rates, which hurt companies that borrow to expand

On debt, Perkins notes that the biggest cloud companies, Microsoft, Amazon and Google, have strong balance sheets. Smaller players have been borrowing more. He estimates debt across parts of the AI ecosystem could reach $2 trillion by the end of 2027. He does not see it as a risk to the whole financial system. But he warns that weak earnings and canceled orders could ripple outward.

On chips, he says they are "notoriously cyclical, and prone to bullwhip effects." In plain terms, small changes in demand can cause big swings in orders and supply.

The people problem

Perkins sees one more trigger that has little to do with earnings: political backlash. He says people who cheer for AI online live in a bubble. In his words, ordinary people do not find it useful, do not want to live near data centers and worry about their jobs. "And they vote," he wrote. He suggests one high-profile accident could turn that unease into wider opposition.

What to keep in mind

This is one economist's warning, not a forecast that a crash is coming. The material does not say when or whether the AI boom will break. For anyone who invests, saves for retirement or relies on AI tools at work, the useful takeaway is where to look: the lenders, smaller suppliers and chip makers tied to AI, not only the best-known tech names.