Rising Borrowing Costs Threaten AI Infrastructure Spending Growth
Financial analysts warn that rising interest rates and power shortages could decelerate AI capital expenditure, potentially impacting S&P 500 earnings and hardware suppliers.
Rising corporate bond rates and Treasury yields are creating a critical moment for the AI trade, as increased borrowing costs threaten to slow capital expenditure. James Russo of AllianceBernstein warns that soaring yields—driven by inflation and government debt—make companies less likely to fund AI projects unless returns are exceptionally high. This pressure is already visible in the market, with yields on CoreWeave bonds recently hitting 15%.
While investment banks project massive spending—with Morgan Stanley forecasting $1.58 trillion by 2027 and Goldman Sachs estimating $800 billion for 2026—the rate of growth is expected to decelerate. Goldman Sachs notes that nearly half of the S&P 500's 2026 earnings growth is driven by AI capex, meaning a plateau in spending could severely depress earnings. This volatility is compounded by a Federal Reserve interest rate hike on September 16 to a range of 3.75% to 4.00%.
Infrastructure constraints further complicate the outlook. Morgan Stanley identifies a projected 32-gigawatt power shortfall for U.S. data centers through 2028. Despite these risks, Morgan Stanley views Nvidia as the best risk/reward stock in the sector, projecting that nearly $700 billion of 2027 spending will go to the chipmaker. Analysts are now looking toward upcoming third-quarter earnings reports from major hyperscalers to determine if firms are adjusting their spending strategies in response to these financial and regulatory pressures.
