Jing Ji Ri Bao·2026-01-08 21:43Core Insights - The artificial intelligence (AI) core industry in China is projected to exceed 1 trillion yuan by November 2025, driven by policy support, market demand, and technological advancements [1] - AI technology is expected to undergo further upgrades in 2026, enhancing the supply capacity of industry elements and empowering traditional industries and emerging sectors [1] Industry Development - Key technologies such as multimodal, logical reasoning, and physical perception are expected to see breakthroughs by 2025, solidifying the foundation for AI industry development [2] - As of September 2025, the number of AI companies in China has surpassed 5,300, with over 400 specialized "little giant" enterprises [2] Company Innovations - Alibaba is significantly increasing investments in its Tongyi Qianwen model, achieving notable improvements in long text processing and visual reasoning capabilities [3] - The China GenAI market report indicates that Tongyi holds the largest market share in enterprise-level model usage in the first half of 2025 [3] - iFLYTEK has iterated its Spark model five times, with the latest version featuring personalized memory capabilities [4] Application Expansion - AI is increasingly embedded in various industries, with the number of smart factories in China exceeding 30,000, leading to a 22.3% increase in production efficiency [5] - The AI large model solution market reached 3.07 billion yuan in the first half of 2025, reflecting a year-on-year growth of 122.1% [5] - TCL is enhancing its manufacturing and R&D processes through AI, with its Star Intelligence model improving product development efficiency by 20% and material development efficiency by 30% [5] Value Chain Insights - Analysis of AI model applications shows a "smile curve" distribution, with higher value in R&D and marketing services, while production manufacturing is gaining traction [6] - The proportion of AI applications in production manufacturing increased from 19.9% to 25.9% in 2025, indicating deeper integration into core value creation processes [6] Supportive Elements - The scale of intelligent computing power in China is projected to reach 1,037.3 EFLOPS by 2025, becoming a mainstream carrier for large model training and inference [7] - Data production in China is expected to exceed 50 zettabytes by 2025, with significant contributions from data annotation bases [7] Data Utilization - High-quality data sets are crucial for AI applications, particularly in autonomous driving, where continuous data collection is essential for model training [8] - Recommendations include prioritizing the construction of industrial large models and enhancing data governance to ensure safe and compliant AI model optimization [8]