Rising debt costs in the AI space are raising eyebrows

You’re likely aware that many technology companies are spending epic sums of money on AI these days. Take tech titans like Oracle, Alphabet, Microsoft, Amazon, and Meta that are building out massive data centers and other infrastructure to power AI. These so-called hyperscalers are on pace to invest more than $750 billion in AI this year—an amount equal to more than 2% of the U.S.’s entire annual gross domestic product.

That AI-focused spending tsunami is changing the tech landscape in ways that investors may want to pay attention to.

In the past, tech companies generally spent relatively little on physical assets such as factories, in sharp contrast to the “old economy” stocks that have been out of favor for much of the last decade. That helped keep tech firms’ borrowing costs low, with the credit spreads on their debt about 14% lower than those of the broader universe of similarly rated companies. What’s more, the firms often self-funded new spending initiatives using the ample free cash flow they generated.

That dynamic has been flipped upside down, with tech companies today largely focused on expensive physical assets that, in some ways, look decidedly old-economy. That’s forcing some hyperscalers and others to look beyond their own free cash flow to a mix of debt and equity financing from the financial markets in order to keep funding their big AI plans.

The result: Tech companies are finding it more expensive than ever to access the money they need. Tech now pays a 16% premium to issue debt relative to the broader market (see the chart).