Goldman Sachs addressed a critical question regarding the future of the U.S. equity market: Are current valuations accurately reflecting the potential benefits of artificial intelligence (AI)? Their findings indicate a nuanced agreement, suggesting that while the market is not in a bubble, it may be overly optimistic about AI’s implications for company valuations.
According to the investment bank’s analysis, the U.S. equity market has likely already absorbed much of the anticipated long-term value that AI could generate. Analysts Dominic Wilson and Vickie Chang pointed out that a straightforward mathematical approach indicates that the market pricing for AI advancements appears to outstrip the broader macroeconomic impact, with valuation increases in AI-focused companies nearing the plausible ceiling of potential economy-wide benefits.
The team’s research estimates that the Present Discounted Value (PDV) of the capital revenue attributable to generative AI for the U.S. economy stands at a baseline of around $8 trillion. This figure, while inherently uncertain, suggests that the eventual range of capital revenues could fluctuate between $5 trillion and $19 trillion. Crucially, this range of projected benefits aligns with current and expected investments in AI-related capital expenditures, which have been the subject of extensive scrutiny in recent financial discussions. However, the analysts caution that market enthusiasm appears to have outpaced these baseline macroeconomic forecasts significantly.
Since the launch of ChatGPT in November 2022, Goldman Sachs reports that the valuation of companies engaged in or connected to the AI sector has surged by over $19 trillion. This remarkable increase encompasses notable gains in the semiconductor industry and among so-called “hyperscalers,” as well as nearly $1 trillion in valuations for the three largest private AI model providers. This upward trajectory places the market gains at the highest end of projected macro benefits, notably exceeding the baseline estimate of $8 trillion.
Goldman Sachs maintained that forward-looking markets typically price in gains ahead of time, which they view as an intrinsic quality rather than a flaw. However, the analysts also underscored two significant risks that could lead to “overpayment” for expected future profits. They referenced historical precedents from the 1920s and the 1990s, where investor enthusiasm surrounding genuine innovations resulted in inflated valuations.
The identified risks include:
-
Fallacy of Aggregation: Investors may mistakenly project excessive aggregate revenue and profit gains from the remarkable earnings growth of individual companies, leading to an inflated valuation for the overall sector that may not accurately reflect collective potential.
-
Fallacy of Extrapolation: Competitive pressures often erode initial profit gains from innovations, resulting in a market potential that may be overestimated if short-term profitability boosts are regarded as permanent.
Despite the inherent risks, the productivity potential of AI remains substantial. Estimates suggest AI could enhance U.S. productivity by approximately 1.5 percentage points over the next decade, ultimately increasing U.S. GDP and earnings by about 15%. Provided that both the broader economy and the AI investment landscape continue to progress as expected, market optimism seems likely to endure. Nevertheless, current profitability from AI—apart from hardware considerations—remains limited, posing significant risks if anticipated outcomes do not materialize promptly.
