Goldman Sachs says AI still not showing up in companies' bottom lines (www.businessinsider.com)

Goldman Sachs reports a striking disconnect: AI talk is everywhere, but earnings impact is scarce. In Q2 a record 58% of S&P 500 companies mentioned AI on earnings calls, and AI-exposed stocks are up 17% year-to-date after a 32% surge last year — yet few firms quantify AI’s contribution to profits. Goldman echoes a McKinsey finding that over 80% of companies say generative AI hasn’t meaningfully affected their bottom line. The market’s enthusiasm has pushed the S&P 500 toward historically high valuations, raising concerns that hype is outpacing measurable returns. To frame the market’s position, Goldman maps the AI trade into four phases: Phase 1 (Nvidia-led chip boom), Phase 2 (current hyperscaler-led infrastructure buildout), Phase 3 (AI-enabled product revenue for software firms), and Phase 4 (broad productivity gains across industries). Hyperscalers — Amazon, Microsoft, Google, Meta and Oracle — are driving Phase 2 with an expected $368B in capital spending in 2025 (vs. $239B in 2024 and $154B in 2023). But Phases 3–4 remain uncertain: AI could compress SaaS pricing or fail to deliver widespread productivity gains. Goldman warns that if AI spending fell back to 2022 levels, it could trim $1T from 2026 sales forecasts and shave 15–20% off the S&P 500, underscoring that the economy is still in the “early innings” of adoption.