The development of artificial intelligence technology represented by ChatGPT has triggered further growth in global computing power demand. Welcoming the new wave of artificial intelligence, where is China's growth point in computing power? What kind of computing power is more accurate and efficient? The reporter recently interviewed industry experts to anticipate the development trend of computing power. Big models call for "big computing power". In the new wave of global artificial intelligence technology development, big models are a keyword. ChatGPT and others use big models to "learn" massive amounts of data and achieve breakthroughs through them. Chen Runsheng, an academician of the CAS Member, said that great changes are taking place in the application of AI big models, which will lead to changes in many fields in the future, and will also consume a lot of computing power. Zheng Weimin, an academician of the CAE Member, said that "great computing power" is necessary to make a big model. ChatGPT is the product of large-scale computing infrastructure. At the recently held Zhongguancun Forum, the New Generation Artificial Intelligence Development Research Center of the Ministry of Science and Technology released the "Research Report on China's Artificial Intelligence Large Model Map". The report shows that China ranks second in the world in the number of large models developed. Lu Zhonghua, a researcher of the Computer Network Information Center of the Chinese Academy of Sciences, said that the demand for computing power of the whole society is growing day by day, and the demand for computing power of artificial intelligence applications is more prominent. We should meet this trend. In recent years, the proportion of artificial intelligence computing power in China has been increasing. According to estimates, the average annual growth rate of artificial intelligence computing power in the past few years has been approximately 70%, while other computing power has increased by about 30%. Chen Wenguang, a professor of computer science at Tsinghua University, also observed that as the penetration of artificial intelligence deepens, the utilization rate of intelligent computing centers becomes higher compared to traditional supercomputing centers. Cao Zhennan, deputy director of the National High Performance Computer engineering Technology Research Center, said that although computing power is often divided into supercomputing and artificial intelligence computing, there are many similarities between the two. The biggest difference is that the requirements for computing accuracy are different. At present, there is a significant gap in the computing power of artificial intelligence, and some artificial intelligence applications have lowered their accuracy requirements in pursuit of computational performance. However, if computing power becomes easy to obtain and cheap, some artificial intelligence applications may have higher requirements for accuracy, "Cao Zhennan said. Experts believe that "universal computing power+specialized computing power" will become the key to the construction of artificial intelligence computing infrastructure. On the one hand, it needs to meet a wide range of application scenarios, have universality, and achieve universality; On the other hand, it can also support personalized application scenarios that require high computational accuracy and efficiency, with high efficiency and dedicated implementation. Building a computing power network "highway" is seen by experts as a way to connect existing computing power centers with different architectures through the network, which can reasonably configure, share, schedule, release more computing power, and lower application barriers. Under the leadership of the High tech Department of the Ministry of Science and Technology, the national supercomputing internet work has been launched. According to the plan, by the end of 2025, the national supercomputing internet will become a "highway" supporting the construction of digital China.