Goldman says the stock market has already priced in the AI boom, with $19 trillion of market value running ahead of actual economic impact so far

Goldman Sachs finds the U.S. equity market has already priced a large portion of AI’s potential — estimating the present discounted value of generative-AI-related capital revenue at a baseline $8 trillion (plausible range $5–$19 trillion) and saying this justifies current AI capex — but notes market capitalization of AI-linked companies has risen by more than $19 trillion since November 2022, implying enthusiasm may be running ahead of macro fundamentals. The bank’s strategists judge valuations high but not at “bubble” extremes, warning investors against two key risks that can drive overpayment: the fallacy of aggregation (extrapolating individual winners’ gains to the whole ecosystem) and the fallacy of extrapolation (treating short‑lived profit spikes as permanent), with historical innovation booms as precedents. While Goldman still sees meaningful upside—AI could lift U.S. productivity by about 1.5 percentage points over a decade, roughly a 15% increase in GDP/earnings—limited current profits outside hardware mean disappointment is possible if the investment and macro backdrop falters.

Analysis

Goldman Sachs estimates the present discounted value (PDV) of generative-AI-related capital revenues at a baseline $8 trillion with a plausible range of $5 trillion to $19 trillion, and states those PDV figures are sufficient to justify current and anticipated AI-related capex. The bank also reports that market capitalization of companies directly involved in or adjacent to AI has risen by over $19 trillion since November 2022, including major gains in semiconductors and hyperscalers and almost $1 trillion of value for the three largest private AI model providers, and characterizes valuations as high but not yet at "bubble levels."

Goldman identifies two structural valuation risks: the fallacy of aggregation—extrapolating standout winners’ profits across the whole ecosystem—and the fallacy of extrapolation—treating transitory profit spikes as permanent—noting precedents in past innovation booms. The bank highlights that valuation gains concentrated in semiconductors and private model providers already exceed the $8 trillion baseline, implying market enthusiasm has sprinted ahead of macro-consistent outcomes.

Goldman further estimates AI could raise U.S. productivity by ~1.5 percentage points over ten years, lifting GDP and earnings by roughly 15% if investment and the macro backdrop remain on track, but cautions that outside hardware current AI profits remain limited and disappointment is possible if expectations are unmet. Forward-looking markets will price gains ahead of realization, which is a feature, but investors face upside if execution proves durable and downside if competition and earnings reversion materialize.

Key Decisions for Investors

• Consider trimming or avoiding large overweight positions in AI-adjacent names where cumulative valuation gains already exceed Goldman’s $8 trillion baseline—notably in semiconductors and the largest private model providers
• Shift incremental exposure toward companies with tangible AI-driven capex and hardware revenue streams that are more likely to realize near-term profits, while keeping position sizes measured
• Monitor capex realization, revenue and margin trajectories outside hardware, and set predefined reassessment triggers to reduce exposure if top-line gains fail to materialize or if competitive erosion appears to be compressing profits