Columbia real estate professor Stijn Van Nieuwerburgh analyzes the AI infrastructure boom as a physical capital buildout centered on data centers, power infrastructure, cooling systems, and specialized chips. He studies how this buildout is financed through hyperscalers, third-party developers, REITs, private credit, and structured finance, and discusses the implications for leverage, risk allocation, and financial stability.
Data Centers: Financing the AI Buildout
Columbia Business School
Stijn Van Nieuwerburgh
Research
65 Pages
Key Takeaways
Buildout rivals history: Planned U.S. capacity of 200 GW implies roughly $8.2 trillion of investment, exceeding prior railroad, highway and telecom infrastructure booms.
Leverage moves outward: The financing mix is estimated at 60% equity and 40% debt, with data center debt commonly carrying about 70% leverage.
Opacity compounds risk: Meta’s 2 GW campus sold 80% to Blue Owl, then raised $27.3 billion through an SPV at roughly 90% leverage.