Year: 2026
Type: dissertation
Abstract: This study examines whether meaningful AI disclosure in IPO prospectuses is associated with a valuation premium at the time of listing and whether any such premium persists in the aftermarket. Using a sample of 527 US IPOs completed between 2016 and 2026, two OLS specifications with year and industry fixed effects are estimated. The primary variable of interest is a threshold-based dummy variable capturing substantive AI disclosure, defined as more than two distinct AI-related keyword matches in the prospectus. The results indicate that firms meeting this threshold command a valuation premium of approximately 37% at the time of listing relative to non-disclosing firms, while subsequently underperforming the S&P 500 by approximately 12 percentage points over the six months following their IPO. The effect is threshold-based rather than continuous, as the raw count of AI mentions carries no significant coefficient in either specification. The findings are consistent with a narrative-driven mispricing mechanism, whereby substantive AI disclosure inflates initial valuations beyond fundamentally justified levels, with the premium subsequently corrected as a broader set of investors can participate in trading. The results are robust to alternative variable constructions and temporal subsamples.
Authors: Mettävainio, Isak, Bergo, Erik
Language: English
Subfield: Finance
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7221843059
