This study develops a cyclical model focusing on the data network effects in AI-enabled platforms. For this purpose, we attempt to connect the concept of data network effects and the previous research of the integrated approach for the platform theory which is a new trend. From the summary of previous studies, we classify the network effects into four categories and confirm the differences in their properties. Based on these studies, we present a cyclical model of data network effects as a multiple loop structure model that combines network effects related to the scale and scope of data. As a virtuous cycle of the data network effects, we point out the mechanism that the deepening of data learning and the expansion of platform boundaries proceed together in AI-enabled platforms. As a vicious cycle of data network effects, we point out the danger that the customer experience provided by AI-enabled platforms through machine learning would overfit existing customers and hinder machine-based innovation for new customers.