AI Hyperscalers Accelerate Borrowing to Fund Infrastructure Amid Rising Treasury Spending

The borrowing activity among AI hyperscalers has surged significantly, with major tech companies issuing between $159 billion and $220 billion in bonds in 2026 alone. This figure has already eclipsed the total bond issuance for 2025, which was around $108 billion to $120 billion. Goldman Sachs projects that AI-related debt issuance could reach nearly $500 billion by the end of the year.

Leading the charge is Amazon, with approximately $53 billion in debt commitments aimed at enhancing its AI infrastructure. Oracle and Alphabet follow with around $25 billion and $20 billion, respectively. This borrowing spree is primarily driven by the need for capital expenditures that exceed internal cash flows, as building data centers and acquiring advanced chips for AI development requires substantial investment.

In 2026, the projected capital expenditure for the AI ecosystem is estimated to be between $725 billion and $740 billion, with forecasts exceeding $1 trillion for 2027. The bond market has shown signs of price weakness for some recent offerings from these hyperscalers, indicating a lukewarm reception from investors.

As competition intensifies among tech giants like Amazon AWS, Google Cloud, and Oracle, the rapid increase in borrowing—from just $13 billion to $17 billion in 2024 to a forecasted $500 billion in 2026—highlights the swift shift in the market following the rise of generative AI.

Investors are advised to monitor AI revenue growth closely, as disappointing returns could lead to significant debt burdens for these companies. The current softness in bond prices suggests that some investors are already factoring in potential challenges ahead.