The costs of building AI

The surge in debt issuance by hyperscalers for AI buildout is significant, exceeding $200 billion in 2026 alone. While this debt fuels economic growth and corporate earnings remain healthy, investors should monitor this trend closely. Current debt levels, though large, are not excessive relative to the companies' size and strong revenue growth.

However, risks are emerging. Widening credit spreads and rising credit default swap levels signal potential stress, especially if free cash flow expectations decline. Additionally, aggressive customer and supplier financing by hyperscalers could obscure leverage and create single points of failure.

Watch for any further deterioration in credit pricing and the potential for a shift back to traditional credit markets if circular financing proves problematic. Competition from Asia and potential political backlash against data centers in the US also present external risks.