Big Tech’s record debt fuels AI infrastructure concerns

Major U.S. technology companies have raised over $120 billion in debt this year for artificial intelligence infrastructure, shifting from their traditional cash-funded strategies.

The change has generated concerns among investors regarding market stability and expected returns.In September, four prominent hyperscaler companies issued nearly $90 billion in public bonds. Alphabet raised $25 billion, Meta $30 billion, Oracle $18 billion, and Amazon $15 billion in its first U.S. dollar bond sale in three years.

Only Microsoft has avoided borrowing from debt markets recently. Total debt issuance has increased substantially from an average of $28 billion over the past five years.

The surge in debt has exacerbated U.S. investment-grade credit spread widening from 74 basis points in mid-September to 84 basis points in mid-November. Alphabet and Meta both paid around 10-15 basis points over their existing debt for these new offerings, indicating investor caution. Oracle’s long-term bonds have faced scrutiny, declining by about 8% and trading at 65 cents on the dollar, with some analysts predicting a potential downgrade to junk status.

Market Pressures

AI capital expenditure is expected to hit $600 billion by 2027, up from $200 billion in 2024. Despite significant borrowing, around 80-90% of planned expenditures rely on cash flows. The mounting debt concerns have contributed to market volatility.

Following Nvidia’s earnings report on November 19, stocks initially rallied before a sharp reversal, with the S&P 500 declining 3% this month due to doubts about AI investment returns.