Goldman Sachs economists Joseph Briggs and Sarah Dong stated in a report published on August 2 that the widely cited figure of approximately $794 billion in capital expenditures by hyperscale cloud service providers systematically underestimates global AI investment by about $200 billion, while overestimating domestic US investment by about $200 billion. After adjustment, Goldman Sachs expects total global AI investment to reach $1.019 trillion in 2026, with approximately $581 billion invested domestically in the US. The report emphasizes that this recalculation implies a stronger scale and duration for the AI capital expenditure cycle than previously anticipated, and Goldman Sachs' leading indicators show that recent growth momentum remains strong. Goldman Sachs also predicts that US AI capital expenditure as a percentage of GDP will rise from 1.8% in 2026 to 2.8% in 2028, while the global figure will increase from 0.9% to 1.4%, consistent with the investment shocks observed during historical general-purpose technology buildout cycles.