Moody’s Analytics chief economist Mark Zandi is warning that the rapid rise in debt issuance by major artificial intelligence companies is becoming a potential threat to the financial system.
His latest analysis shows that AI-related borrowing has accelerated so sharply that it now far exceeds the levels reached by internet and telecommunications firms during the Y2K era, Zandi said in an X post on December 7.
To this end, the economist noted this scenario raises concerns that the sector’s swelling leverage could amplify economic risks if conditions shift.
Zandi noted that non-refinancing bond issuance by technology companies in 2025 is running at record highs in nominal terms. After adjusting for inflation, the volume stands near historic peaks previously seen only when interest rates were exceptionally low.
This year’s surge is being driven overwhelmingly by AI companies, whose financing needs have grown alongside soaring valuations and intensifying competition in the race to build advanced computing infrastructure.
More aggressive borrowing patterns
He emphasized that these borrowing patterns are more aggressive than those seen around Y2K, when tech firms borrowed heavily to expand the early internet.
Even then, total issuance did not reach today’s levels, and Zandi sees little evidence that bank lending or alternative credit sources filled a larger gap for those earlier companies.
While he does not expect the rising debt to immediately undermine the AI sector, Zandi cautioned that the risks escalate meaningfully if companies fail to meet investors’ lofty expectations.
A downturn in stock prices would make their rapidly growing balance-sheet obligations harder to manage, potentially turning today’s expansion-driven borrowing into tomorrow’s strain.
Overall, Zandi’s assessment places AI-sector leverage firmly on the radar as a developing vulnerability, one that, if left unchecked, could spill over into credit markets and weigh on the broader economy.
Featured image via Shutterstock
