Goldman Sachs analysts report that the top five US hyperscalers—Amazon, Alphabet, Microsoft, Oracle, and Meta—are projected to invest approximately $800 billion in AI infrastructure throughout 2026. However, the bank estimates that only $70 billion in additional AI-related cloud revenue will be generated above pre-AI trends, creating a potential annual shortfall of $230 billion that may challenge current tech stock valuations.
The surge in capital expenditure is impacting the broader technology sector, leading to weaker free cash flow. To fund these initiatives, companies are increasingly turning to debt markets. As of mid-September 2026, AI-related borrowers accounted for 44% of the $135 billion in US convertible bond issuance. Furthermore, Goldman Sachs has increased its forecast for US investment-grade debt issuance by $200 billion, potentially reaching a record $2.3 trillion this year.
Goldman Sachs Asset Management has adopted an underweight stance on hyperscaler debt, signaling caution regarding concentration risk. While the Brookings Institution projects total US AI infrastructure investment could reach $10.3 trillion by 2032, investors are now closely monitoring whether cloud revenue growth can bridge the gap with capital spending and how bond markets will respond to the rising supply of corporate debt.
