Intangible Capital Pricing in the Age of Artificial Intelligence: Evidence from U.S. Equity Markets

Sammanfattning

This thesis examines whether R&D intensity and AI-related disclosure explain cross-sectional variation in future stock returns in contemporary U.S. equity markets. R&D intensity represents the established case from the classical R&D anomaly literature, while AI disclosure intensity, constructed from Item 1 and Item 7 of annual 10-K filings, serves as a textual proxy for disclosed firm-level AI exposure. Using 755 U.S.-listed, R&D-active non-financial and non-utility firms from July 2018 to December 2025, the study combines S&P Capital IQ data, Fama-French risk factors, and 10-K-derived AI disclosure scores. The Fama-MacBeth evidence does not detect a linear R&D return premium, but portfolio sorts identify a significant risk-adjusted alpha concentrated in the highest R&D-intensity quintile. AI disclosure intensity is positively associated with subsequent excess returns, suggesting that AI-related 10-K language contains pricing-relevant information beyond conventional R&D expenditure. Overall, the findings suggest that the traditional R&D anomaly has weakened as a broad linear predictor while surviving in the extreme upper tail, and that AI disclosure carries cross-sectional pricing information consistent with an early-diffusion technology premium.