How to pick out the good kind of tech debt
Add Axios as your preferred source to
see more of our stories on Google.
Illustration: Allie Carl/Axios
An abundance of AI-related bonds have come to market, yet investors still have an appetite for the debt of cash-rich hyperscalers like Google.
Why it matters: Not all tech debt is created equal. Investors will need to suss out the winners from losers to avoid getting burned if the AI promise fails to turn out as hoped.
What they're saying: "The really pristine, strong balance sheet, hyperscalers, they're not going anywhere for a very long time. So the risks there are pretty modest," says Guy LeBas, chief fixed income strategist with Janney Capital Management. "Not all debt is bad," he tells Axios.
State of play: Besides fueling talk of a bubble, the AI-related debt wave is reshaping the broader investment-grade bond market.
- Credit spreads are widening, a sign that investors are starting to demand more compensation for the risk of owning corporate debt, LeBas says.
- Spreads were unusually tight this summer, as global demand for high-quality U.S. dollar debt bumped up against limited supply.
- Now, tech issuance alone could increase net new supply by as much as 20%, reversing that imbalance and pushing spreads wider.
By the numbers: Google's capital expenditure as a share of its cash flow is the lowest among the major hyperscalers, at 56%, according to Goldman Sachs.
- Meta is at 85%, Oracle is at 132%, and CoreWeave is at 271%.
- The divergence explains why the cost to hedge against an Oracle default jumped to three times its September level, while the cost of Google debt remains the same.
- Google debt is rated AA+ by Moody's, while Oracle debt is rated BBB.
Reality check: LeBas cautions that investment-grade corporate bonds with exposure to tech are "a little more concerning" than private credit, even as that sector takes on more AI financing.
- The returns on private credit are higher than those on corporate debt, so "you're paid very effectively in the sector for taking the risk," he notes.
- Despite concerns that private-credit lenders are overexposed to AI, he says only 10% to 15% of private-credit lending is directly tied to tech.
- "The worst risk I ever take is when I'm not paid well."
The bottom line: AI is transforming tech. Now, it's transforming debt.
- For investors, the question is not whether to own it, but whose debt to own, and how much compensation is offered for the risk.
