The deep integration of artificial intelligence (AI) and scientific research is giving birth to an unprecedented technological revolution. With China's rapid layout and continuous investment in the field of AI for Science, the strategic direction and development path of AI driven scientific research have become increasingly clear. Recently, the State Council issued the "Opinions on Deepening the Implementation of the" Artificial Intelligence+"Action", proposing to accelerate the implementation of the "Artificial Intelligence+" scientific and technological action, accelerate the process of scientific discovery, take the lead in establishing a new research and development paradigm based on AI, and promote China's scientific and technological innovation to be at the forefront of the world. Looking ahead to the 15th Five Year Plan, AI for Science will become a key engine for China's technological system reform and innovation capability leap, laying a solid foundation for achieving the intelligent transformation of the overall scientific research and development paradigm, as well as the strategic goal of building a world science and technology power by 2035. To promote the intelligent transformation of scientific research paradigms, the essence of AI for Science is to use artificial intelligence to solve major scientific problems that could not be solved in the past and accelerate the pace of original innovation. For example, the "century old challenge" of protein structure prediction in the field of life sciences has made breakthrough progress through AI technology. Secondly, AI for Science is developing new algorithms and tools to drive fundamental changes in scientific research. The emergence of new generation AI based research tools has greatly improved the efficiency and accuracy of scientific computing and simulation. The deeper significance of AI for Science lies in promoting the intelligent transformation of the overall scientific research paradigm. This is not a transformation of individual institutions or teams, but a reshaping of the entire scientific research system and industrial landscape. It will promote the transformation of scientific research from a "workshop model" to a "platform model". However, the breadth and depth of the revolution triggered by AI for Science have not yet been fully recognized by the international community. Therefore, this is a rare opportunity. If China can effectively grasp, concentrate its efforts, and systematically promote, it is expected to take the lead in realizing the new paradigm of "platform scientific research" in the next five years, ensuring China's leading position in global scientific and technological innovation. The paradigm shift of building an intelligent scientific research "highway" relies first and foremost on the construction of a solid scientific research infrastructure. Just as highways are to modern economic society, intelligent scientific research infrastructure is the "foundation" for the development of AI for Science. In recent years, China has made initial progress in the construction of scientific research infrastructure. For example, basic scientific research platforms represented by the Bohr Space Station integrate various functions such as literature, data, computation, and experimentation, which can efficiently support the full process management of scientific tasks and are rapidly becoming popular tools among researchers. In addition, breakthroughs have been made in the development of general scientific research models and intelligent agents represented by Innovator+SciMaster. These intelligent agents not only possess interdisciplinary knowledge, but also have independent innovation and "dry wet closed-loop" scientific research capabilities. While maintaining their general abilities, they significantly enhance their scientific professional capabilities. The promotion and application of AI for Science, a typical scenario and highlight project, requires breakthroughs in key areas and typical scenarios, forming replicable and promotable "model projects", and driving the transformation and upgrading of the overall scientific research system. In the field of materials, promoting the use of AI in material genome engineering to model and predict the complex relationships between material structure, properties, and processes can significantly shorten the research and development cycle of new materials and reduce research and development costs; In the field of chemistry, organic synthesis is an important branch. AI driven organic synthesis pathway planning and reaction prediction can significantly improve the research and development efficiency of new molecules and drugs. Intelligent and efficient catalyst design and catalytic reaction research have also made gratifying progress; In the field of life sciences, AI for Science has shown great potential in protein structure prediction, gene editing, drug discovery, and other areas. The AI based molecular docking and drug screening platform has become a core tool for new drug development, which is expected to promote the development of new models such as personalized medicine and precision diagnosis and treatment. Of particular note is that AI provides the possibility to establish virtual cells with a wide range of practical applications. The focus of scientific research is gradually shifting from "model building" to "data building" in order to unleash the value of scientific data. In the past decade, the scientific community has mainly focused on model innovation, solving a series of problems such as curse of dimensionality, symmetry, training stability, and long-range dependencies by continuously improving the complexity and capabilities of models. Looking ahead to the future, the quality and diversity of data are becoming the core of further breakthroughs in AI for Science. How to make good use of existing data and layout incremental data has become the key to winning. Among them, the stock data mainly includes literature, proprietary corpora, professional databases, etc., providing a foundation for pre training and knowledge acquisition of AI models. Compared with general data, scientific data presents stronger professionalism, involving complex and deep disciplinary knowledge and relationships, and puts higher demands on data processing and annotation. Incremental data is the driving force behind the continuous breakthroughs of AI for Science. With the popularization of automated experimental platforms, high-throughput computing, and intelligent sensing devices, the incremental data output capability in the scientific field has greatly improved. Among them, data communities are becoming an important platform for incremental data acquisition and sharing. By building portals such as "Science Navigation", we aim to create a hub for effective collaboration among researchers, data resources, and innovative forces, promoting open sharing, precise labeling, and collaborative governance of scientific data in a community-based manner. Reshaping China's scientific and technological innovation system with AI for Science is not only a choice of technological path, but also a historic opportunity to reshape China's scientific and technological innovation system. In recent years, China has continuously invested in the field of AI for Science, and has made positive progress in basic tools and facilities, typical application scenarios, and laboratory intelligent transformation. Currently, the AI driven paradigm shift in scientific research provides unprecedented strategic opportunities for China to achieve its goal of becoming a world science and technology powerhouse by 2035. It is urgent to seize the opportunity and coordinate efforts on the existing foundation. To achieve this goal, it is necessary to go from top-level design and policy support to collaborative innovation among frontline researchers, deeply implement the "AI+" scientific and technological action, promote the intelligent transformation of the overall scientific research and development paradigm, and form a scientific research and innovation ecosystem driven by AI as the core. Only by seizing opportunities and continuously breaking through can we win the initiative in global competition in the new round of technological revolution. (New Society)