近日，由中国高技术产业发展促进会主办的" 2025 AI智能体赋能产业增长暨创新创业发展峰会 "在杭州隆重召开。 未可知人工智能研究院院长 杜雨博士 受邀出席大会，并发表题为《 AI产业革命与具身智能崛起 》的主旨演讲，深入剖析人工智能产业趋势，分享前沿 研究成果，引发广泛关注与热议。 AI产业革命进行时 中国AI进入2.5阶段 杜雨博士指出，中国AI产业正经历第三次发展浪潮。 继"AI四小龙"和"AI六小虎"之后，以 DeepSeek 为代表的新兴力量推动中国AI进入"2.5阶段"，即 从通用大模型向具身智能、 AI for Science等纵深领 域演进 。 "大语言模型的崛起不仅重塑了AI产业格局，也催生了 具身智能、 AI硬件、AI for Science 等万亿级新赛道。"杜雨博士表示。 具身智能 下一个万亿级赛道 在演讲中，杜雨博士重点分析了 具身智能与人形机器人 的产业机遇。 他指出，随着"中国制造2025"战略推进，智能制造、医疗、服务等领域对 人形机器人 的需求将爆发式增长。 到2030年，全球人形机器人市场有望迎来爆发式拐点。 | | | 非人形机器人 | 人形机器人 | | --- | ...

Investment Rating - The report does not explicitly provide an investment rating for the AI for Science industry. Core Insights - The AI for Science industry leverages artificial intelligence to accelerate scientific research and discovery, utilizing data-driven and model-driven approaches to enhance efficiency and accuracy in scientific endeavors [9][10][12]. Summary by Sections Industry Overview - AI for Science is defined as the use of AI technologies to expedite scientific research and discovery, employing big data and machine learning to uncover hidden patterns [9][10]. - The evolution of scientific paradigms has transitioned from direct observation to AI-assisted research, marking significant advancements in scientific methodologies [24][26]. - The current stage of AI for Science is characterized by a deep integration of AI technologies into scientific research, enhancing predictive capabilities and fostering innovation [28][30]. Technical Analysis - Core technologies in AI for Science include high-performance computing, data management infrastructure, scientific computing software, pre-trained large models, and high-throughput experiments, all of which facilitate accelerated scientific research [32]. - High-performance computing is crucial for processing large datasets and training complex machine learning models, significantly improving research efficiency [35][38]. - High-throughput experimentation enables rapid execution of complex experimental designs, generating vast amounts of data for machine learning model training [42][45]. Industry Development Practices - AI for Science is a cross-disciplinary field that applies AI technologies to traditional scientific domains such as physics, chemistry, biology, and medicine, showcasing its potential to drive scientific research and technological innovation [46][51]. - In the life sciences, AI is transforming drug development, optimizing genomic research, and enhancing personalized medicine through data analysis and predictive modeling [53][56]. - The application of AI in earth sciences improves data analysis and predictive modeling, aiding in climate change research and geological disaster prediction [62]. - In materials chemistry, AI enhances data analysis and predictive modeling, helping scientists understand and address complex material systems [65].