Wall Street banks are scrambling to finance the explosive growth of AI data centers as deal sizes reach tens of billions of dollars. Regional lender Citizens has set a billion-dollar minimum for new conversations, with managing director Adam Lewis noting that what were once $100 million milestone deals are now "rounding errors." The scale represents a fundamental shift from traditional commercial real estate financing to large-scale infrastructure finance.
The surge in capital requirements stems from soaring costs of land and electricity needed for AI infrastructure. Major banks including Morgan Stanley, Goldman Sachs, and JPMorgan have formed integrated teams spanning lending, markets, and private capital over the past two years to understand the mechanics of data center construction. The financing challenge is pushing these projects beyond conventional lending limits.
Citigroup estimates the AI data center buildout could require $3 trillion by 2030, according to an internal memo from the firm's investment banking leaders. The bank has established a dedicated AI infrastructure group to evaluate "all pockets of capital" as deals become larger and more complex. The financing needs are beginning to exhaust even the cash reserves of the world's largest technology companies.
The unprecedented scale of AI infrastructure investment is forcing banks to break down traditional silos between investment banking, corporate banking, and financing divisions. As hyperscale technology companies cannot afford to fall behind in the infrastructure race, banks are positioning themselves to capture what could be Wall Street's largest-ever financing opportunity across multiple business lines.