Nvidia CEO Jensen Huang projected on May 20, 2026 that AI capital expenditures could reach 4 trillion annually, far exceeding Wall Street estimates. [1] CFO Colette Kress added that AI infrastructure spending is on track to reach 3 to 4 trillion annually by the end of the decade. An analysis by Laura Martin at Needham shows consensus estimates for hyperscaler capex at $1.03 trillion in 2028, a fraction of Huang’s forecast.
Cloud revenue growth supports Huang’s optimism, with Alphabet jumping 63%, AWS 28%, and Microsoft 40% in the latest quarter. Synergy Research Group data shows Q1 2026 cloud spending hit 129 billion, with trailing 12-month revenues at 455 billion. [2]
Despite strong spending, returns remain uncertain. JPMorgan estimated that AI would need $650 billion in annual revenue by 2030 to deliver a modest 10% return on investment. A Federal Reserve study found a mismatch between executive-reported productivity gains and measured outcomes, with AI adoption showing substantial heterogeneity across firms. [3]