Goldman's Hatzius: AI's Boost to U.S. Growth to Peak by Late 2026, Rate and Wealth Risks Loom

AI's contribution to U.S. economic growth is expected to peak between late 2026 and early 2027, Goldman Sachs Chief Economist Jan Hatzius wrote in an Oct. 26 client note, even before recent market doubts over the investment boom's durability surfaced. The mechanism is a slowing base effect, not a collapse: AI spending keeps rising but its share of GDP will grow more slowly, while the fading boost from the "One Big Beautiful Bill" law's restored 100% equipment depreciation compounds the drag on capex. Hatzius flagged two channels to watch. Long-term developed-market bond yields have risen roughly 100 basis points on average since early 2026, driven partly by large fiscal deficits and "AI-related borrowing"; sustained yields at current levels would shave an estimated 0.6% off GDP cumulatively. Oil prices add pressure, with fuel and European and Asian gas prices near or above March-April highs on Middle East tensions. Goldman estimates the AI-driven equity rally added 0.4 percentage points to U.S. consumption growth over the past year via the wealth effect, about 0.25% of GDP. Hatzius expects that contribution to fade gradually, though a sustained AI stock decline would make the unwind sharper. He also warned that AI adoption's labor-market disruption carries near-term negative macroeconomic effects, citing a significant negative correlation between AI adoption and non-manufacturing employment growth.