从跟跑，到并跑：中国AI这五年｜“十四五”规划收官

Core Insights - The "14th Five-Year Plan" period is crucial for China's AI industry to transition from "catching up" to "keeping pace" and "leading" globally, marking a significant phase for the emergence of new productive forces [1][5] - The narrative of China's AI development has shifted from merely "catching up" to showcasing a collective rise, with China now holding 60% of global AI patents and achieving significant milestones in AI model development [1][6] Industry Growth and Trends - The AI industry in China is experiencing explosive growth, with a notable increase in the number of AI companies from 1,454 in 2020 to over 5,300 by September 2025, representing 15% of the global total [6] - By 2024, the AI industry in China is projected to exceed 700 billion yuan, maintaining a growth rate of over 20% annually [6] - The number of large models with over 1 billion parameters in China has surpassed 200, with a total of 1,509 models expected by July 2025, accounting for 40.2% of the global total [5][6] Technological Advancements - Significant advancements in technology autonomy have been observed during the "14th Five-Year Plan," with breakthroughs in foundational models, domestic computing power, and algorithm optimization [5] - Chinese AI models are increasingly competitive on the global stage, with nine out of the top ten open-source models listed by Hugging Face being from China [6] - The performance gap between Chinese and American AI models has narrowed dramatically, with the difference in performance shrinking from 17.5% in 2023 to just 0.3% [9] Applications and Impact - AI is becoming a foundational infrastructure across various sectors, including government, finance, manufacturing, and healthcare, with applications rapidly expanding [11] - The integration of AI in agriculture has led to breakthroughs in crop breeding, while AI is also enhancing operational efficiency in factories through intelligent scheduling and visual AI technologies [12] - The "AI+" initiative has resulted in the establishment of over 1,000 industry data sets and significant investments in AI infrastructure by central enterprises [10] Future Potential - The next decade is expected to see AI deeply integrated across various sectors, with a goal of achieving over 70% penetration of new intelligent terminals and applications by 2027 [13] - The concept of "AI twins" is emerging, where intelligent terminals will evolve into personalized partners, enhancing user interaction and experience [13] - The commercialization of embodied intelligence is gaining momentum, with significant contracts and collaborations indicating a growing market for intelligent robotics in industrial applications [14]