The artificial intelligence boom, requiring vast amounts of borrowed money for infrastructure, is increasingly competing with the US government for the same pool of cash. This competition has led to a phenomenon some are calling "reverse crowding out," where AI hyperscalers are issuing a flood of bonds, impacting Treasury yields.
Ruchir Sharma, chair of Rockefeller International, suggests the AI boom is approaching a bubble, rating it a "seven to eight" on a scale of one to ten. He identifies four characteristics of a bubble: overvaluation, over-ownership, over-investment, and overleverage. According to Sharma, AI stocks and the broader US market are currently as expensive as they have ever been, exhibiting signs of overvaluation.
US investment-grade corporate bond issuance, largely fueled by AI companies, reached approximately $1.7 trillion year-to-date through July, on track to surpass $2 trillion for the first time. This surge in corporate debt, particularly from AI firms, has led to higher Treasury yields as capital flows into corporate bonds instead of government debt. While other factors like federal budget deficits, higher oil prices, and a robust US economy also contribute to rising Treasury yields, the significant capital demands of the AI sector are a notable factor. Some analysts, including Ed Yardeni, suggest that the AI revolution is creating a classic crowding-out effect, causing Treasury yields to rise.
