At the WAIC 2026, while humanoid robots showcased impressive feats, a closed-door discussion titled 'Chain π Roundtable' addressed the gap between AI technology demonstrations and real-world applications. Wu Liming, Chairman of Huamao Technology, highlighted that Moore's Law has failed beyond 5nm, with chip performance now limited by physical constraints. Advanced packaging and optical interconnects are crucial, with advanced packaging sales expected to surpass traditional packaging by 2025.
The conversation also introduced the concept of 'native AI in factories' by Wei Jie, Chairman of Shanghai Jingzhi. He noted that many industrial AI projects excel in labs but struggle in factories due to issues like data fragmentation and insufficient validation scenarios. He emphasized that AI must grow from real manufacturing data to ensure operational stability.
The roundtable concluded with three key insights: the bubble in computing infrastructure will not burst before fall 2028, the speed of AI adoption depends on data loop iterations, and the focus should be on profitability rather than trends. The overarching theme was the need for AI to operate sustainably and profitably in real environments, underscoring the importance of stable revenue for technology development.
Editor's Note
The discussions at WAIC 2026 reflect critical challenges in the AI sector, particularly in bridging the gap between technology demonstration and practical implementation. As companies strive for sustainable AI solutions, understanding the nuances of data integration and operational constraints will be vital for successful deployment in various industries, including manufacturing and healthcare.
Leave a comment