JPMorgan has raised its forecast for AI infrastructure investment to $5.5 trillion, as giants like NVIDIA are turning to debt financing

J.P. Morgan strategist Tarek Hamid and his team have raised their forecast for total investment in artificial intelligence infrastructure by 2030 to $5.5 trillion in a recent research report, an increase of $400 billion from their prediction last November. The bank noted that in this investment race for super-scale data centers, approximately $4.1 trillion will come from debt financing, with loans covering an average of 85% of total project costs, indicating that AI capital expenditures have shifted to a debt market-centric financing model.

Since last November, global bond issuance related to AI and data centers has exceeded $300 billion. The latest typical case comes from chip giant NVIDIA, which completed the pricing of a $25 billion investment-grade bond issuance this Monday, marking its return to the bond market for the first time in five years. This issuance was conducted in seven tranches (with maturities ranging from 2 to 30 years) and attracted oversubscription of up to $85 billion, ultimately increasing the issuance size by 25% from the initial target.

The research report emphasizes that although tech giants like NVIDIA, Alphabet, and Amazon are generating substantial cash flow from the AI boom (with NVIDIA estimating free cash flow exceeding $200 billion this fiscal year), these giants still choose to issue hundreds of billions of dollars in bonds. This indicates that such bond issuance is not due to a "lack of financing," but rather that the credit market is confirming the pricing of AI assets.