Mark Habib, S&P Global Ratings’ global head for digital infrastructure, said financing tied to data centers and smaller AI firms may create obligations the agency does not capture, raising questions about its credit analysis.
An S&P outlook dated Sept. 3 said Alphabet, Amazon, Microsoft, Meta, Oracle and SpaceX together may direct more than $7 trillion to AI and data-center investment over 2025-30. The projected total raises questions about how credit analysts account for commitments tied to the expansion.
Companies fund the buildout through borrowing, equity issuance, leases and other arrangements. Habib pointed to cases where hyperscalers support major data-center construction or smaller AI companies. Credit review may be less straightforward when that backing runs through a joint venture, a special-purpose vehicle or a residual-value guarantee.
Habib spoke at a London roundtable on Oct. 19. He said, “It is safe to assume that there is more (debt) out there that we are not seeing.” He did not name the companies or offer a figure for the possible exposure. Habib said he saw no disconnect that would call for S&P to change ratings immediately.
In an Aug. 27 forecast, S&P put the group’s capital spending above $1.3 trillion for 2027. S&P projects every company in the group to have negative free operating cash flow in both 2026 and 2027, with recovery projected only by 2029. The agency is monitoring how firms turn AI spending into revenue, how durable demand proves to be, the likelihood of excess capacity and the treatment of contracts that resemble debt.
Habib has previously said some AI-related obligations may be missing from public company disclosures and transactions rated by S&P. He has also identified profits below expectations as a key risk to hyperscaler ratings. The remarks extend EconoTimes’ earlier reporting on AI-related commitments not visible in public disclosures.
