Banks and debt investors scrutinise AI financing after market selloff

Banks and credit investors are examining artificial-intelligence financing more cautiously after a sharp stock-market selloff, according to banking and credit-market sources. Some lenders that recently sought such business are reported to be less willing to add exposure, even where large technology companies provide guarantees.

The shift concerns appetite for new credit risk, not a confirmed halt to all AI lending. Investors are examining how borrowers' cash flow would hold up in a downturn and when major customers could reduce or terminate contracts. Those conditions can affect repayment even when demand for computing capacity is high.

Sources also cited concern about greater scrutiny of data centers after the US midterm elections. Construction delays could affect borrowing growth and investment-bank fee income, but the report does not establish that every project will be delayed or lose funding.

Demand for debt remains substantial. Goldman Sachs Research forecasts global borrowing by large computing-infrastructure companies of $420 billion in 2027, compared with its $250 billion estimate for 2026. These are forecasts, not completed financing totals.

Higher-risk computing-rental businesses and independent data-center operators face particular pressure. The lending debate increasingly centres on the revenue available to repay debt and the exact terms of customer contracts, rather than relying on the broader AI investment story or a corporate guarantee alone.