AI high-end dialogue: Is the agent “cold”? What is hot is a concept, what is cold is an industry

Zhidongxi Author | Yang Jingli Editor | Xinyuan Zhidongxi reported on July 7 that on July 2, the 2026 China Intelligence Conference was held in Hangzhou. In the high-end dialogue session of the opening ceremony, Zhang Guoren, co-founder and editor-in-chief of Zhixixi, held discussions with Hao Jianye, professor at Tianjin University and founder of MemoraX AI, Xie Jibao, vice president of technology for Alibaba’s Qoder product line, and Fang Shaoxia, founder and CEO of Original Semiconductor, focusing on "from individuals to enterprises, building and crossing bridges of intelligent agents." ▲High-end dialogue: From individuals to enterprises, building and crossing the bridge of intelligent agents. This high-end dialogue started from the changes in the popularity of the intelligent agent industry in the first half of 2026, and discussed core issues such as the transition of intelligent agents from personal use to enterprise implementation, the industrial division of labor between large manufacturers and start-up companies, the key threshold for the inclusiveness of intelligent agents, the essential difference between Agent Computers (intelligent computers) and AI PCs, memory systems and multi-agent collaboration, and changes in working methods in the AI ​​era. All three guests believed that the intelligent body industry is moving from a conceptual craze to a more pragmatic industrial stage. Xie Jibao said that customers no longer only look at demos, but are concerned about whether the task can be completed end-to-end and how much efficiency improvement it can bring to the enterprise; investors have also begun to ask about product ROI, business monetization logic and single-task end-to-end costs. He said: "What is hot is actually called a concept, and what is cooled down is called an industry." ▲Xie Jibao, Vice President of Technology for Alibaba's Qoder product line In terms of the division of labor between large manufacturers and startups, all three guests mentioned "complementarity." Hao Jianye believes that the future market space for Agent is "at least trillions, or even larger." Start-ups can start from segmented industry tracks and continue to improve their competitiveness through data barriers. Xie Jibao gave a more intuitive metaphor: Large manufacturers are more like making engines, and start-up companies can build a variety of cars based on engines. Regarding the inclusiveness of intelligent agents, Hao Jianye believes that large models as tools have reached the general public, but higher-end personal assistants and productivity tools have not yet truly become available.