3 analysts on why AI debt isn't a bubble

Group 1: Corporate Bond Issuance - U.S. companies have issued over $200 billion in investment grade corporate bonds in 2023, with approximately 13% allocated for AI-related infrastructure projects as of end-October [1] - Five major companies, known as hyperscalers—Amazon, Google, Meta, Microsoft, and Oracle—account for the majority of this borrowing, collectively raising $121 billion in investment grade debt this year [2][3] - The borrowing surge is expected to push total U.S. corporate bond issuance to a record $1.8 trillion in 2024, according to JP Morgan [3] Group 2: Market Reactions and Concerns - Concerns are rising that hyperscalers may be overspending, as stocks of the Magnificent Seven have declined by about 7% in November [4] - Bond markets are experiencing uncertainty, with Amazon's recent $15 billion note sale seeing a 40% drop in orders after pricing, which is double the typical drop-off rate [5] - Spreads on bonds issued by major hyperscalers have widened, indicating increased perceived risk, with Amazon, Microsoft, Meta, and Alphabet's spreads increasing from around 50 to nearly 80 basis points [5] Group 3: Oracle's Unique Position - Oracle's bond spreads have widened by 48 basis points since September, with its $3.25 billion, 5-year bond trading at 104 basis points over Treasuries, doubling since mid-September [6] - Demand for credit default swaps on Oracle bonds has surged, with spreads more than doubling since September, reaching the highest cost since 2023 [7]