According to Goldman Sachs, on July 28, big tech hyperscalers are expected to finance over one-third of AI capital expenditures through debt issuance, with bond offerings projected to reach $250 billion in 2026 and $400 billion by 2027. Goldman Sachs estimates these debt issuances will represent 33% of total capital spending in 2026 and 35% by 2027, as firms like Meta, Microsoft, Alphabet, and Amazon face cash constraints amid soaring AI infrastructure investments.
In 2025, hyperscalers spent $405 billion on capital expenditures and issued $108 billion in investment-grade bonds, representing 26% of spending. Goldman Sachs expects this ratio to rise significantly, with capital expenditures projected to reach $750 billion by end-2026 and $1.2 trillion by 2027. The bank noted that debt financing may also take alternative forms including project finance and asset-backed securitization.