Goldman Sachs: The Macroeconomic Spillovers From AI Electricity Demand

Goldman Sachs examined the macroeconomic spillover effects of rising electricity demand driven by artificial intelligence. As of December 2025, US electricity prices surged 6.9% year-over-year, significantly outpacing the 2.9% increase in overall personal consumption expenditures (PCE) inflation. The bank expects robust demand from AI data centers to further accelerate electricity consumption, while slow supply growth tightens regional wholesale power markets.

Data centers currently account for roughly 7% of total US electricity demand. Goldman Sachs projects that data centers will contribute approximately 1.2 percentage points to overall electricity demand growth through 2030. Supply constraints persist due to prolonged regulatory approvals, with project timelines extending to 4.5 years, alongside equipment and labor shortages. Residential electricity inflation is expected to remain around 6% through 2027 before easing to 3% in 2028 as natural gas prices decline.

Higher electricity prices directly impact inflation and indirectly raise core PCE inflation through cost pass-through channels, with healthcare, transportation, and food services notably affected. The resulting reduction in real disposable income is projected to drag down consumption growth by approximately 0.2 percentage points in 2026 and 2027.

Ultimately, rising power prices are expected to shave about 0.1 percentage points off annual GDP growth. However, Goldman Sachs emphasizes that this macroeconomic drag is marginal compared to the substantial long-term productivity gains generated by artificial intelligence.