Abstract: Against the backdrop of deep integration between "Data Elements×" and "AI+" initiatives, AI models have emerged as a high-order form for unlocking the value of data elements. Exploring their rise and evolutionary dynamics in the data element trading market holds both theoretical value and practical significance for improving China's basic systems for data. Taking highly representative cases such as Hugging Face, OpenAI, AWS Marketplace, Alibaba Cloud, and Shenzhen Data Exchange as samples, this study adopts an exploratory multi-case study approach to conduct an in-depth analysis, systematically deconstructing the evolutionary patterns and trading ecosystems of typical AI model transactions within the data element trading market. On this basis, through a horizontal cross-case comparison, five distinctive evolutionary paradigms are identified: open-source ecosystem, closed-source service, platform integration, platform-community synergy, and institution-led types, thereby revealing the global evolutionary logic and driving mechanisms of AI model trading. The study further identifies differentiated development paths for AI model trading between China and the rest of the world, and accordingly puts forward targeted policy recommendations in light of the current status of AI model trading in China's data element trading market, with a view to propelling the evolution of China's data element market toward an intelligence-driven, high-order form.

Key words: AI Model trading, Data element trading market, Multi-case study, Evolutionary paradigms, Trading ecosystem, Basic systems for data