AI investment announcements and stock market performance A comparative study across sectors in the United States

Sammanfattning

This thesis examines how the U.S. stock market responds to firm’s investment announcements in artificial intelligence (AI), focusing on the sectoral differences. This study is grounded in the Efficient Market Hypothesis, Modern Portfolio Theory and Signaling Theory. The study applies an event-study methodology to 20 S&P 500 firms. The sample is equally divided between the healthcare sector, characterized by extensive AI integration (“high-AI”), and the business/consumer services sector with comparatively lower AI adoption (“low-AI”). The expected returns are estimated using a restricted version of the market model, the abnormal returns are examined across two event windows [-1, 0, +1] and [-5, 0, +5]. The findings show a significantly negative market reaction for high-AI firms, while low-AI firms experienced a smaller decline, which was statistically significant only in the longer [-5, 0, +5] event window. These findings suggest that investors remain cautious about the long-term implementation risks in sectors where AI is already deeply integrated. Welch’s t-tests were conducted to assess statistical differences between the sectors and confirmed that high-AI firms experienced significantly more negative abnormal returns compared to low-AI firms. Overall, the study shows that the level of AI maturity affects the market reaction. The announcements perceived as costly or complex tend to be received negatively. The findings also suggest that uncertainty around new technologies might limit market efficiency.