Zhang Yaqin: The Internet of Intelligent Bodies is the next stop for AI

On September 24th, Zhang Yaqin, a foreign academician of the Chinese Academy of Engineering, a Tsinghua University Chair Professor, and the director of the Institute of Intelligent Industries (AIR), delivered a keynote speech titled "Intelligence Emergence: Reflections and Explorations in the Age of Artificial Intelligence" at the 2025 Snapdragon Summit China. He also engaged in a dialogue with financial writer Wu Xiaobo, discussing the technological evolution and industrial prospects of AI.

Zhang Yaqin proposed that the internet has entered the "Internet of Agents" stage, with its industrial scale expected to expand more than a hundredfold compared to the mobile internet era.

Artificial intelligence is becoming a universal technological cornerstone, like electricity, driving exponential economic and social development. He reviewed the concept of "AI" proposed a decade ago, noting that AI has become the core driving force of the Fourth Industrial Revolution.

He emphasized that the key characteristic of current AI development is the "emergence of intelligence"—the simultaneous advancement of technological breakthroughs and industrial integration. On a technical level, he summarized five major trends in the development of large models: multimodality, autonomous intelligence (agents), edge intelligence, physical intelligence, and biological intelligence.

The essence of the new generation of AI lies in the fusion of atoms, molecules, and bits, representing the synergy of information, physical, and biological intelligence.

It will also promote the integration of carbon-based and silicon-based life. He believes that AI will be deployed increasingly at the edge, achieving low power consumption, low latency, and high personalization. He also called for global attention to AI governance and risks.

In a conversation with Wu Xiaobo, Zhang Yaqin pointed out that autonomous driving, as a typical application of physical intelligence, requires end-to-end technical support from large models.

Comparing the development of AI in China and the United States, he believes that China has advantages in application implementation and young talent, but still faces challenges in chips and basic models.

He predicts that open source will become the mainstream of AI, and that China will develop three to four large-scale model ecosystems. Looking ahead, Zhang Yaqin emphasized that the current path of relying on scale expansion will peak in five years.

The next generation of AI needs to break through the boundaries of memory, reasoning, and cognition. The integration of intelligent agents with physical and biological intelligence will be the key innovation direction for the next decade.

Tsinghua University AIR, led by Zhang Yaqin, has already deployed in three major areas: information, physical, and biological intelligence, and is collaborating with industry to promote technological implementation. This summit was hosted by Qualcomm, and Zhang Yaqin's sharing provided participants with important insights that combined academic foresight with industrial practice.