GoKawiilBrookings study: AI data-center buildout equals 3.63% of US GDP annually
A Brookings Institution study by Columbia Business School professor Stijn van Nieuwerburgh finds that AI infrastructure investment—chips, electricity and data centers—now averages 3.63% of US GDP per year, exceeding prior booms in canals, railroads, electrification and highways. The study estimates 183 gigawatts of new data-center capacity is planned over seven years, more than tripling the roughly 57 gigawatts currently installed, and projects this could double the electricity consumption of the entire US residential sector.
GoKawiil's interpretation of the reporting above, not reported fact.
The Wall Street Journal reports the buildout is already raising costs for labor, electricity, commercial real estate and chip-dependent consumer goods, while crowding out new housing construction. Van Nieuwerburgh suggests, per Reuters, that heavy reliance on outside financing—banks, private credit and real estate firms—has raised leverage and spread risk across the economy in ways reminiscent of the railroad and telecom bubbles, though he stops short of predicting a specific collapse.
- AI buildout equals 3.63% of US GDP annually, larger than past infrastructure booms
- Planned data-center capacity could hit 183 gigawatts within seven years, up from 57 gigawatts now
- Rising outside financing has increased leverage and spread risk across banks, lenders and real estate firms
Source: slashdot.org — Posted, 2026-09-27
Published there as: “Just How Big is the AI Buildout - and How Risky?”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.