Large technology companies are increasingly relying on mega bond sales, defined as $25 billion or more, to finance their aggressive expansion into artificial intelligence. This year alone, seven tech companies have reached this threshold, exceeding the total number of such sales in the previous six years combined. Amazon's recent $25 billion bond sale highlights this trend, even though some investors reportedly perceived it as a surprise given prior expectations about their borrowing for the year.
This surge in borrowing has led to concerns among financial experts. Morgan Stanley analyst Khanduja noted that credit risk is currently undervalued, pointing to issues like light investor demand for Amazon's latest offering and widening spreads on other Big Tech bonds. In the first five months of the year, major players like Amazon, Meta, Alphabet, and Oracle collectively issued $159 billion in debt, and analysis by the Kobeissi Letter indicates this corporate borrowing accounts for approximately 15% of total US corporate bond issuance year-to-date.
The Bank for International Settlements (BIS) has warned that the race to build AI infrastructure, fueled by this debt, could mirror past technological booms that resulted in severe market disruptions. The BIS highlights the increased risk of broader financial turmoil if the anticipated productivity gains from AI fail to justify the massive investments and borrowing. This sentiment is echoed by institutional investors, with Apollo Global Management reporting a significant drop in cover ratios for hyperscaler bonds, from 5x in February to just 2x in July, suggesting these companies may need to offer higher rates to attract investors in the future.
Investor caution is evident in the bond market, where the cost to insure Big Tech debt against default has surged. Credit default swap (CDS) spreads for Oracle, Amazon, Google, and Microsoft have risen to around 75 basis points, near their highest in at least seven years. Excluding Oracle, these spreads are still around 49 basis points, the highest since at least 2018. Both metrics have more than doubled since the beginning of 2025 and are now significantly above their 2022 bear market peaks, indicating a rising perceived risk associated with Big Tech's AI-driven debt.
Overall, while the stock market remains optimistic about AI, bond investors are becoming increasingly skeptical about the long-term returns on these investments. Global AI-related debt issuance reached $23 billion as of May 31, a fourfold increase from the prior year, and Morgan Stanley projects this could grow to $57 billion by the end of 2026. This aggressive debt-fueled expansion into AI infrastructure is fundamentally reshaping current credit markets and raising questions about its sustainability.
