Concept image of AI city File photo: VCG
2026 is a pivotal year for artificial intelligence (AI) to make the leap from cognitive intelligence to embodied intelligence, and more efficient and large-scale AI models are expected to be deployed in China, Wang Jing, vice dean of the School of Future Science and Engineering at Dongzhou University in Suzhou, eastern China’s Jiangsu province, told Global Times in an interview on the sidelines of the 3rd China Innovation Challenge on Artificial Intelligence on Sunday. Application scene of intelligence.
The event attracted a total of 113 teams, including more than 350 participants from across the country, to showcase their latest achievements in AI applications.
Mr. Wang predicts three major trends in China’s AI industry in 2026. First, the focus of competition among large-scale AI models has shifted from scale to efficiency and capabilities, with a focus on improving inference and efficiently building intelligent agents, driving the evolution of AI from generation to planning to execution.
More efficient large-scale AI models are expected to emerge this year, he said, and more enterprise-level applications will be equipped with task-oriented AI agents, acting as “digital employees” to improve the efficiency of basic tasks.
Additionally, spatial intelligence has become a research frontier, with large-scale AI models breaking through spatial understanding and powering areas such as unmanned systems and digital twins. Moreover, the integration of AI and industry has reached a deep psychological level, with AI deeply embedded in the core processes of various industries, driving the reshaping of industrial paradigms, Wang said.
In terms of new outcomes, embodied intelligence will see small-scale commercial use in scenarios such as industrial testing and home services. The development of large-scale scientific AI models will be accelerated, the scientific research process will be restructured, and R&D cycles will be shortened. New computing architectures and green data centers introduced to alleviate energy pressure will further evaluate AI innovation using green energy and provide new support for large-scale applications of AI, Wang said.
The number of AI companies in China is expected to exceed 6,000 in 2025, and the size of China’s core AI industry is expected to exceed 1.2 trillion yuan ($171.39 billion) in 2025. Meanwhile, AI applications are expanding to cover major industries such as steel, non-ferrous metals, power, and communications. According to the latest data from the Ministry of Industry and Information Technology, applications in product research and development, quality inspection and customer service are increasing.
However, AI development around the world still faces challenges and bottlenecks, such as the contradiction between the demand for computing power and the supply of resources, the lack of capacity for large-scale AI models and the requirement for accurate use, and the contradiction between rapid technological development and lagging ethical governance, Wang said.
To solve this problem, experts said, efforts are needed to improve the energy use efficiency of computing power training by restructuring the training framework and optimizing the underlying chip design, incorporating low-cost green energy into the computing power planning system, and mitigating uneven resource distribution by introducing edge computing power and promoting cross-domain scheduling of computing resources.
Furthermore, further efforts are needed to improve the training structure and data logic of large-scale models, enhance the model’s causal inference ability, improve the accuracy of the generated content, and at the same time overcome technical bottlenecks in model interpretability to increase confidence in decision-making. Third, efforts should be made to improve ethical norms, laws and regulations in the AI field, while strengthening international coordination and rule coordination in AI governance, Wang said.
