S&P Global warns of less-visible AI debt at hyperscalers

S&P Global has warned that major technology companies could face more AI-related debt exposure than their disclosures reveal, but sees no immediate need to change their credit ratings. The Economic Times Markets, citing Reuters in an October 19 report, said Mark Habib, S&P’s head for digital infrastructure, raised the concern at a London roundtable.

Habib pointed to complex financing arrangements involving data-centre projects and smaller AI companies that could create liabilities for hyperscalers without making those obligations readily apparent on their balance sheets. He said S&P had no estimate of the additional debt these arrangements might generate, and that its visibility was limited where obligations were not disclosed or transactions were not rated by the agency.

The agency expects Amazon, Microsoft, Alphabet, Oracle, Meta and SpaceX together to spend more than $7 trillion on data centres and other AI-related capital expenditure during 2025–2030. Its central ratings concern remains whether that investment produces the expected profits. Further spending increases or more intense competition for revenue could alter its assessment.