A surge in borrowing tied to artificial intelligence is converging with expanding US debt supply, potentially pushing up borrowing costs across the economy, according to Bridgewater Associates founder Ray Dalio.
Dalio projects annual net US debt supply will climb to approximately $4.4 trillion in 2027 from about $3.5 trillion in 2025, not including maturing debt that must be resold. In a LinkedIn post on October 13th, he flagged the combination as a significant market challenge.
AI Companies Drive Record Bond Issuance
AI hyperscaler bond issuance reached approximately $220 billion in the first half of 2026, compared with about $108 billion for all of 2025, according to Dalio. The International Monetary Fund separately reports that hyperscalers now account for 10% of global investment-grade bond issuance, up from less than 1% in 2024.
Higher capital expenditure funded by increased borrowing and equity issuance is leading to declining free cash flow among these companies, which Dalio characterizes as "the classic first stage of a squeeze."
Growing Pressure on Borrowing Costs
If bond buyers demand higher yields to absorb the increased debt supply, borrowers ranging from the US government to technology companies will face elevated borrowing costs. Dalio describes this market dynamic as "the canary in the coal mine."
Supplier financing is also expanding, with chipmaker Broadcom agreeing to lend Anthropic up to $42 billion to purchase its chips.
Dalio notes that the concurrent growth of corporate debt in the AI sector and existing Treasury debt obligations creates a compounding challenge for the broader financial system.
