AI data center borrowing has rapidly emerged as a major credit concern for Wall Street investors, with 34% of global fund managers now citing AI hyperscaler capital spending as the most likely source of a future systemic credit event, according to a Bank of America survey conducted between May 8-14. This represents a doubling from 17% in April, highlighting a significant shift in investor perception of market risks. US private credit remains the top concern for 42% of respondents, though this is down from 57% last month.
Tech companies have borrowed more than $300 billion from US investors to fund AI spending since the beginning of last year, with bankers predicting billions more to come in the months ahead. As reported by Bloomberg, JPMorgan Chase managing director David De Boltz described the situation as "exponential" in terms of volume, noting that "everyone is working out where they're going to put that cash, and how much cash do they have to hold for these deals." The concern centers on firms loading up on unprecedented debt amounts for AI buildouts where payoffs remain uncertain.
Michael Burry's latest analysis reveals alarming trends in AI investment patterns, with 87% of venture capital funding now flowing into artificial intelligence companies according to figures from Apollo Global Management chief economist Torsten Slok. Burry compared today's AI investment boom to the dangerous dot-com bubble before the 2000 market crash, warning that rising AI debt, aggressive tech valuations, and risky corporate borrowing could trigger a major financial correction. His warning arrives as venture capital firms aggressively fund AI startups while investment-grade and high-yield debt issuance increasingly finances AI infrastructure expansion.
Burry sees similarities in today's data-center construction race tied to artificial intelligence, drawing parallels to the telecom and cable companies that aggressively expanded networks during the dot-com boom. He noted that nearly half of all investment-grade bond issuance is now connected to artificial intelligence, while 38% of high-yield debt issuance is linked to AI-related activity. Historical comparisons are particularly concerning, as tech, media, and telecom companies accounted for roughly 40-50% of high-yield bond issuance during the dot-com era, with more than $100 billion of investment-grade bonds issued during that period eventually downgraded to junk status by 2002.
The survey of more than 150 participants reveals broader tail risks, with 40% noting a resurgence of inflation as the biggest threat, followed by 20% citing geopolitical conflict and 18% an orderly rise in bond yields. Other concerns include an AI bubble at 11%, private credit at 6%, and Japanese government debt at 4%. Despite AI spending risks, De Boltz noted that lenders are becoming more cautious about financing software companies that could be disrupted by AI, instead directing capital toward businesses directly tied to the technology. Burry's warning focuses on how quickly investor confidence can reverse when growth expectations become unrealistic, particularly as enterprise software companies face potential disruption from AI agents.
