AI Infrastructure Capital Commitments Reach Multi-Trillion Scale

Multiple financial institutions and research bodies have sized the AI infrastructure buildout at levels that dwarf prior technology investment cycles. The Brookings Institution projects total investment in data centers and AI infrastructure will reach $10.3 trillion between 2025 and 2032, averaging 3.6% of U.S. GDP per year. Moody's reports the five largest hyperscalers hold roughly $2.8 trillion in combined future financial commitments, up from about $350 billion in 2023, a roughly eightfold increase. That total includes over $1 trillion in uncommenced data center leases, about $1.57 trillion in purchase commitments, and about $137 billion in guarantees. Goldman Sachs estimates roughly $7.6 trillion could be invested across AI compute, data centers, and power infrastructure between 2026 and 2031. On the demand side, Truist projects AI cloud annual recurring revenue will grow from $310 billion in 2025 to nearly $2.1 trillion by 2030, at roughly 47% annually, with inference rather than model training becoming the dominant driver of infrastructure demand. Truist also projects AI data center capacity to grow from 24 gigawatts in 2025 to 100 gigawatts in 2030. In fixed income, AI-related bond issuances account for 18% of the investment-grade corporate bond index, the largest single sector represented, with data centers constituting an additional 2% separately.