JPMorgan Upgrades AI Infrastructure Investment Forecast to $5.5 Trillion as Tech Giants Shift to Debt Financing

ChainCatcher report: JPMorgan strategists Tarek Hamid and his team have raised their forecast for total global investment in AI infrastructure by 2030 to $5.5 trillion, an increase of $400 billion from their November 2023 projection. The bank noted that of this massive data center investment surge, approximately $4.1 trillion is expected to be financed through debt, with loans averaging 85% coverage of total project costs—indicating that AI capital expenditures have shifted toward a debt-market-driven financing model. Since November last year, global bond issuances related to AI and data centers have exceeded $300 billion. The latest prominent example comes from chip giant NVIDIA, which priced a $25 billion investment-grade bond offering on Monday—its first return to the bond market in five years. The offering, structured in seven tranches ranging from two to 30 years, attracted over $85 billion in oversubscription, resulting in a final issuance size 25% higher than initially planned. The report emphasizes that despite tech giants like NVIDIA, Alphabet, and Amazon generating substantial cash flows from the AI boom—NVIDIA is projected to generate over $200 billion in free cash flow this fiscal year—these companies are still issuing hundreds of billions in bonds. This suggests that such debt issuance is not driven by funding shortages, but rather reflects the credit market’s formal valuation and confirmation of AI assets.