Debt is a growing concern in the bear case for AI, and it's enough of a worry that a top economist says it should be considered a risk to the broader financial system if things go awry in the tech sector.
Mark Zandi, the chief economist at Moody's Analytics, has been bearish on the US economy all year. He's been focused on the growing threat of a recession, but recently, the economist flagged another concerning trend in the tech world.
Zandi noted that AI hyperscalers are taking on high levels of debt through record bond issuance in 2025, eclipsing the borrowing by tech companies in the years before the dot-com bubble popped.
"Issuance so far this year by all technology companies, including AI and telecommunications companies that did the bulk of the borrowing around Y2K to expand the internet, dwarfs that done around Y2K," he wrote on X.
He emphasized that companies aren't simply refinancing existing debt, but increasing their borrowing as they fight for a share of the booming AI market. Zandi argued that, even adjusting for inflation, tech-sector borrowing is at records. What's more, surging debt is a major risk to broader markets and the economy, he said.
"Borrowing by AI companies should be on the radar screen as a mounting potential threat to the financial system and broader economy."
Zandi's posts reiterated a bearish argument about the AI market and the economic outlook he made earlier this month, warning of misguided optimism driving the investment frenzy.
"There are soaring AI stock prices, which discount this optimism, and arguably, a lot more," he said in an earlier post. "There are the massive (over) investments being made in data centers and other AI infrastructure."
His concerns echo those of other commentators who have attempted to sound the alarm on colossal spending on AI infrastructure and data centers. Skeptics fear that, even for industry-leading companies, the payoff from AI is uncertain, and the economics of the technology could easily be upended as they were for early internet companies in the 2000s.
Zandi argues that if companies can't keep pace with expectations, it could send share prices down, negatively impact other tech stocks, and shake confidence in the entire industry.
Zandi's thesis on the AI boom and dot-com bust also includes the warning that the 2000s crash was mostly felt by equity investors. However, with debt levels rising, an AI-fueled market event would be felt across markets more broadly, increasing systemic risks.
