Closely watched credit ratings of the Great & Mighty

AI Generated by Gemini

This week in AI Finance September 6, 2026 – Here I am reading the Financial Times, meandering through the boggeyman land of Standard & Poor's ratings measures, which surely have lost some punch over the years. Perhaps not expecting a lot, FT’s Toby Nangle began to drill down into a recent report from the credit authority, and found S&P searching for measures that illuminate the real state of six Big 7 companies' outlooks. The backdrop: AI spending still on the rise.

We think we know AI investors are beginning to demand larger concessions as debt issuance volumes reach record levels, despite now baby-butt-fresh balance sheets at the hyperscalers and Nvidia. [see AI Apocalypse Now ... Pay Later]

What we know we don’t know is where the pay-off is. It’s the same for S&P, which writes:

Every time we take a deep dive into this sector, we find that capex is rising faster than we anticipated, financings are becoming more complicated and less transparent, and that returns on investment will take years to realize.

A basic criticism from the accountants is lack of shared evidence of this glowing payback. The complaint: “the big six hyperscalers don’t quantify their returns on investment on AI.” [Note: The Big 7 rundown these days regularly places Nvidia in the ‘hyperscaler’ category. They've become far more than a supplier, and even become a banker.]

There’s more to the S&P worries too. One central concern is disclosed in the report's litany of circular financing that encases this technology revolution. How about direct equity investments, chip financing, lease liabilities and lease guarantees, power purchase agreements (PPAs), take-or-pay agreements and residual-value guarantees, and backstop guarantees?

Plenty much more in “Are credit rating agencies getting fed up with hyperscalers?” – this story  includes an FT' "Rumsfeldian quadrant" that parses the S&P thinking on the issue. What are unknown unknowns? It is a question that may have an influential inflection on the economy in the near future.

Side Note: So many ways to dissect this trending. For example: The hyperscalers [some of them] built their clouds by taking on the customers’ capex burden... and now, with non-commodity GenAI tech generation, the capex is possibly busting their balloon.

I’ve yet to find anyone who can accurately and consistently forecast the market behavior with any degree of accuracy, beyond short-term trend following. -Barry Ritholz