U.S. AI data center buildout could cost more than $10 trillion by 2032

Forbes examines the money going into U.S. AI data centers, from chips and construction to electricity and water. It cites estimates as high as $10.3 trillion by 2032. Goldman Sachs economists forecast U.S. AI investment reaching 2.8% of GDP in 2028, which Forbes says would exceed the late-1800s railroad boom’s 2.4% share. Whether enough power can be built to support the planned data centers remains unclear.