Goldman Sachs Forecasts 24‑Fold AI Token Surge by 2030 Amid Chip Shortage

TL;DR: Goldman Sachs sees AI token consumption exploding to 120 quadrillion per month by 2030, while a persistent high‑end chip shortage may delay the rollout of autonomous AI agents.

Context Goldman Sachs Research released a report projecting that autonomous AI agents will drive a 24‑fold increase in global token usage by the end of the decade. Tokens—units of text processed by language models—are the metric for AI workload. The firm links this surge to falling compute costs and a shift toward “always‑on” background agents that act without human prompts.

Key Facts - Global token consumption is expected to reach 120 quadrillion tokens each month by 2030, a 24‑times rise from current levels. - High‑end semiconductor (chip) supply will remain constrained for 12‑18 months, with full alignment to demand taking roughly two years as manufacturers expand capacity. - By 2030, 12 % of knowledge workers—employees whose jobs rely on information processing—will use agentic AI, climbing to 37 % by 2040. - Enterprise uptake lags consumer adoption; only a minority of small‑to‑medium businesses currently deploy autonomous agents, though pilot projects are growing.

What It Means The projected token boom signals massive data center demand, potentially boosting revenue for cloud providers that can secure sufficient chip inventory. However, the lingering semiconductor shortage creates a bottleneck: firms may struggle to scale AI workloads without access to the latest processors. Companies that invest early in chip partnerships or diversify across hardware vendors could capture a margin advantage, as higher gross margins translate into stronger operating cash flow.

For knowledge workers, the gradual rise to 12 % AI usage by 2030 suggests early productivity gains will be confined to tech‑savvy firms. The longer‑term trajectory to 37 % by 2040 points to a deepening integration of AI assistants into routine tasks, reshaping job functions across sectors.

Watch for updates on chip fab expansions, pricing trends for high‑performance processors, and the pace at which enterprises move from pilot to production deployments of autonomous AI agents.