BIS Warns AI Financing Opacity Could Speed Market Crash

The Bank for International Settlements, in its annual economic report, has put AI investment risk on a short list of threats to global financial stability. According to South China Morning Post reporting, the concern is not just whether AI investment will disappoint, but how fast a correction could move. Zhang Tao, BIS Asia-Pacific representative, put it plainly: "the interconnectedness of the financial system and interplay of vulnerabilities could mean the speed of a correction could be much faster than previous banking crisis episodes."

The structural reason for that speed is where the money is sitting. AI investment has increasingly flowed through hedge funds, private credit vehicles, and other non-bank intermediaries that operate with less oversight than conventional lenders. The five largest hyperscalers are set to spend over a trillion US dollars on AI-related capital expenditure from 2025 through 2026, with those commitments already outpacing their earnings and free cash flow, prompting debt issuance. Beyond straight corporate bonds, the BIS flagged circular financing arrangements between chip makers, hyperscalers, and AI labs, where equity stakes are coupled with multi-year purchase commitments in deals where terms are reportedly "typically poorly disclosed, with risks of the same asset being pledged multiple times." Direct lending funds have quadrupled their AI and IT sector exposure to 15% of portfolios.

All of that makes the financial system less legible to regulators at the moment when legibility matters most. Non-bank intermediaries do not carry the same circuit breakers as regulated lenders. The BIS drew the historical comparison itself, listing canal mania in the 1830s, British railway mania in the 1840s, electrification exuberance in the late 1920s, and the dotcom boom of the late 1990s as precedents where technology investment surges attracted excess capital and ended in reversals and recessions.

The honest caveat is that the BIS also noted that immediate risks to financial stability appear modest, so this is a forward-looking warning rather than a current alarm. What the reporting does not give you is a precise account of how much of the trillion-plus in hyperscaler capex is flowing through opaque non-bank channels versus conventional debt markets, or what specific regulatory responses might follow.

The clearest near-term opportunity sits with regulators, who now have explicit BIS backing for pushing disclosure requirements on off-balance-sheet AI financing structures. For investors, the report is less a sell signal than a prompt to ask harder questions about where AI infrastructure spending is actually being financed and by whom.