Artificial intelligence disclosures and IPO underpricing

Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).

This study examines how artificial intelligence (AI)-related disclosures in IPO filings affect underpricing. Using a sample of 1,138 U.S. IPOs from 2013 to 2024, we identify AI disclosures in S-1 filings through textual analysis. We find that AI-referenced IPOs achieve significantly higher first-day returns, consistent with attention-based mechanisms whereby AI language attracts investors and amplifies speculative demand. This effect is concentrated among non-tech firms and during periods of high investor sentiment, suggesting that novelty and market mood influence the pricing of AI disclosures. However, following the public release of ChatGPT in 2022, the underpricing premium associated with AI disclosures declines sharply, consistent with the view that increased investor familiarity reduced the narrative novelty of AI disclosures and attenuated attention-driven pricing. Overall, our evidence is most consistent with declining narrative novelty and increased investor familiarity rather than definitive evidence of investor learning.