Wall Street Mobilizes $2.25 Trillion for AI Infrastructure Push

AI model training and inference are turning hyperscalers into capital-intensive businesses. Data centers require not only accelerators and servers but also power generation, grids, cooling, water and critical minerals. Gartner estimates hyperscalers will spend $6.3 trillion on AI infrastructure through 2030. The scale is drawing banks and private-capital firms deeper into financing while sharpening concerns that long-dated assets, rapid hardware obsolescence and uncertain AI returns could leave lenders exposed.

Between Aug. 10 and Aug. 12, 2026, Wall Street unveiled initiatives totaling more than $2.25 trillion. Morgan Stanley said it would facilitate $1.5 trillion of capital raising, financing and investment over a decade. Nvidia outlined about $500 billion of AI infrastructure financing with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR, while Bank of America launched a $250 billion, one-year Critical Infrastructure Finance Initiative on Aug. 12. The total reflects capital to be mobilized or facilitated, not direct balance-sheet spending by the banks.