Brookings projects US AI spending to total $10.3T between 2025 and 2032, averaging 3.6% of GDP per year, the largest single-industry build-out as a share of GDP

The financing behind AI infrastructure is becoming less transparent, potentially obscuring systemic risks before they materialize.

Brookings projects that U.S. investment in Artificial Intelligence infrastructure, including data centers, power systems, networking, and specialized chips, will reach $10.3 trillion between 2025 and 2032. This represents an average of 3.63% of the U.S. GDP annually, marking the largest single-industry build-out as a share of GDP in U.S. history, exceeding past infrastructure booms like canals, railroads, and highways. The report highlights that the financing of this AI build-out is increasingly shifting from transparent corporate on-balance sheet methods to opaque off-balance sheet structures such as joint ventures, private credit, and securitization. These arrangements rely heavily on AI companies' cash flows and collateral, which are subject to significant uncertainties. The paper suggests that improved measurement and transparency are crucial policy contributions as the industry's capital structure evolves to better understand potential correlated exposures before any potential downturn.

The projected $10.3 trillion investment in AI infrastructure signals a monumental economic build-out, dwarfing historical infrastructure projects in its proportional scale to GDP. This massive capital expenditure underscores the rapidly expanding physical and technological foundations required to support the AI revolution, indicating a significant reallocation of economic resources towards this sector.

The shift towards opaque off-balance sheet financing for AI infrastructure presents a new set of risks, as these structures can obscure correlated exposures and make them difficult to monitor. This evolving financial landscape necessitates enhanced transparency and measurement to mitigate potential systemic vulnerabilities that could arise from the concentrated dependencies on AI companies' financial health and collateral values.