Top hyperscalers likely to have more AI-linked debt than now known, S&P Global warns

S&P Global warns that the world's leading tech companies could be carrying more debt linked to AI than previously recognized. The analysis comes as concerns grow over the mounting borrowing by AI infrastructure providers and "neocloud" firms like Nvidia-backed CoreWeave. While S&P Global's top analyst for the sector, Mark Habib, has not altered any ratings at this time, he warns that the biggest risk to the ratings of the major hyperscalers lies in the potential shortfall of expected profits.

Habib also mentions the complex financing structures in place, where hyperscalers essentially guarantee the construction of vast data centers or smaller AI firms. Habib emphasizes that if such liabilities are not disclosed, the true financial position of the hyperscalers might not be accurately reflected. S&P expects the top six US hyperscalers, including Amazon, Microsoft, Alphabet, Oracle, Meta, and SpaceX, to invest over $7 trillion in data centers and other AI-related capital expenditures between 2025 and 2030. Currently, these firms project negative free cash flow until at least the following year.

Habib states that S&P is attempting to adopt a holistic approach in assessing the situation. Although they do not yet have an estimate of the additional debt that could stem from the interconnected financing, they are working towards a comprehensive evaluation. In July, S&P downgraded Oracle to the lowest investment grade, BBB-, due to concerns over its AI spending and uncertain profitability path.

While the recent market selloff of bonds associated with AI's aggressive buildout hasn't prompted a change in S&P's ratings, Habib notes that this could potentially happen if the expenditure patterns become unsustainable. Factors such as further increases in investment by AI firms or heightened competition intensity could influence this perception.