Abstract: This paper delves into the economic incentives of data trading by developing an endogenous growth model that incorporates data as a factor in combination with AI. It examines the application of corporate data in the training of generalized AI to create vertical AI. Furthermore, the study investigates the differences in data trading incentives between in-house training models and those trained models provided by specialized AI service providers, as well as their respective impacts on the economy. It is found that, on the one hand, the combination of data and AI reinforces the growth effect in the data economy. On the other hand, compared to the in-house training model, firms have a higher willingness to sell data when trained by the AI service provider, which in turn increases the scale of the data trading and ultimately drives an increase in the level of total output. Further analysis suggests that the key driver behind these differences is the contribution of specialized data. It is observed that only within the training model facilitated by specialized AI service providers, the willingness of firms to engage in data trading increases in tandem with the rising contribution of specialized data. This paper expands the endogenous growth mechanism of the data economy and provides a reference for the design of incentive mechanisms for data trading centers.

Key words: data trading, nonrivalry, specialized data, artificial intelligence