AI debt is surging. A credit ratings agency has concerns
Add Axios as your preferred source to
see more of our stories on Google.
AI financing is getting bigger and more complicated, and the pristine credit ratings of the Big Tech hyperscalers are at risk, warns S&P.
Why it matters: The enormous sums these companies are spending — it's in the trillions of dollars, much of it borrowed, some invested in riskier businesses — to fund data centers and other AI-related infrastructure are fueling U.S. economic growth overall.
- Risks to the sector are risks for everyone.
What they're saying: "The credit quality of hyperscalers is gradually weakening," the S&P analysts wrote.
- "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."
- They forecast that the top six hyperscalers — Amazon, Microsoft, Alphabet, Oracle, SpaceX and Meta — will spend more than $7 trillion on data centers and AI capex through 2030.
The big picture: For the hyperscalers, the risks are not "existential," says Naveen Sarma, an analyst at S&P who coauthored the report. Most of these companies, with the exception of Oracle, are highly rated borrowers with strong cash-flowing businesses.
- The issue is that these established players are lending their reputations and credit ratings to smaller and untested companies — neoclouds, data center operators and unproven, yet giant startups like Anthropic and OpenAI.
- "When we get together and talk about where the risks are from AI, it's these smaller companies," Sarma says. "It's municipalities, banking on taxes from data centers and spending lots of money on infrastructure. It's utility companies building power plants. All of these ancillary things."
Between the lines: Those ancillary things are driving a lot of economic growth, yet there are "Rumsfeldian" unknowns in this increasingly massive sector, write the analysts. These include:
- The potential for cheap, open-source models to undercut Anthropic and OpenAI, bringing down the cost of AI and the return to investors.
- Exactly when will these massive investments pay off?
- How much the circular financing is inflating the revenue of everyone involved?
How it works: Much of the money companies are spending on AI is coming from debt — they're borrowing for themselves — Amazon alone has raised around $100 billion in the bond market this year.
The intrigue: They're also, critically, backing borrowing by all these other players. Essentially, hyperscalers are using their stellar credit ratings to help less well-known and riskier companies borrow money for lower interest rates.
- That's the more opaque part of the equation — a shadow lending market.
- That assistance might mean agreeing to sign a lease in the future, to backstop a loan or to buy chips.
- The big guys are also taking big stakes in OpenAI and Anthropic.
AI earnings are becoming more circular, writes Richard de Chazal in a note from William Blair Equity Research Friday morning.
- "If end-demand disappoints, these firms could be hit twice: first through slower revenue growth and then through lower valuations on their AI-related investments."
What to watch: If one of these lesser-known companies stumbles, that would hurt the giants, per the note. And that would ripple through the interconnected system in unclear ways.
- "The scale of overlap and interconnectedness is vast," the S&P analysts wrote. "In a downturn even the best capitalized and most profitable firms may incur substantial pain."
