Entry Timing, Pre-Entry Knowledge Diversity, and Innovation Performance in AI Technology

With the breakthroughs in algorithms, computing power, and big data, artificial intelligence (AI) technology has garnered widespread attention in both academic and industrial spheres. The existing literature on entry timing predominantly focuses on the traditional manufacturing industry, treating knowledge as merely a resource or capability foundation, and often overlooks the dynamic knowledge capabilities specific to the AI industry, where existing knowledge can be self-learned, and new knowledge can be autonomously generated. Through a dynamic knowledge capability lens, this paper seeks to elucidate the relationship between entry timing and AI-based innovation performance within the AI industry, as well as the moderating impact of pre-entry knowledge diversity. In this study, we analyze data from 891 AI firms in China, obtained using an AI-related keywords database, and compile 2,101 panel data points from 2018 to 2020 for regression analysis. The findings reveal a U-shaped relationship between entry timing and AI-based innovation performance, which is negatively moderated by pre-entry knowledge diversity. This paper dissects the dynamic knowledge capability mechanism that underlies entry timing in the AI industry and provides a practical framework for firms with varying levels of pre-entry knowledge diversity.