Tech giants and chipmakers are increasingly relying on debt to finance the expansion of AI infrastructure. Major companies, including Amazon, Microsoft, Alphabet, Meta, Oracle, Nvidia, and Broadcom, are borrowing heavily to fund data centers, servers, GPUs, and power capacity.
Estimates for AI-related debt issuance in 2026 range between $450 billion and $570 billion, significantly exceeding the $100 billion to $121 billion recorded in 2025. Goldman Sachs reports that gross corporate-bond issuance by hyperscalers in early 2026 has already surpassed the total for the entire previous year.
Capital expenditures for hyperscalers are projected to reach between $600 billion and $820 billion in 2026, with potential for exceeding $1 trillion in 2027. Amazon alone completed a $54 billion bond sale in March 2026 to support its infrastructure development.
Financial indicators show rising pressure, with leverage ratios for leading hyperscalers increasing from roughly 0.9x to 1.8x over six months. Aggregate lease commitments have climbed to approximately $1.5 trillion, compared to $200 billion five years ago, with $1 trillion in leases yet to commence.
The bond market is showing signs of fatigue as spreads on AI-linked debt widen and cover ratios decline. Investors are increasingly cautious as credit pressure mounts for companies like Oracle and various chip manufacturers, while actual debt levels may be higher than balance sheets indicate.
