With the rapid development of new generation information technologies such as artificial intelligence, computing power has become an important driving force for promoting technological innovation, industrial upgrading, and social progress. China attaches great importance to the development of the computing power industry. The outline of the 15th Five Year Plan proposes to comprehensively promote the construction of computing power facilities, the development of model algorithms, and the supply of high-quality data resources, in order to build a solid foundation for the development of digitalization. The 2026 government work report proposes to implement new infrastructure projects such as ultra large scale intelligent computing clusters and computer collaboration, strengthen national integrated computing power monitoring and scheduling, and support the development of public cloud. The Ministry of Industry and Information Technology promotes the construction of computing power infrastructure in accordance with the systematic approach of "point, chain, network, and surface", and has issued multiple policy documents such as the "Action Plan for High Quality Development of Computing Power Infrastructure", "Notice on Organizing the unveiling of a Strong Computing Power Foundation", and "Notice on Fully Carrying out Automated Monitoring of Computing Power Situation Awareness", providing important guidance for the construction of computing power infrastructure and the development of the computing power industry. Under the guidance of policies and the guidance of industrial development, the development of the computing power industry has achieved positive results, presenting a good development situation of continuous optimization of computing power layout, steady increase in computing power scale, gradual improvement of industrial ecological construction, deepening of computing network integration, and increasing empowerment of computing power applications. In terms of computing power infrastructure construction, as of the end of March 2026, China's intelligent computing power scale has reached 1882EFLOPS (FP16), ranking among the top in the world and strongly supporting the rapid development of China's artificial intelligence industry. At present, a multi-level computing power facility system and a national integrated computing power network have been preliminarily constructed, with national hubs as the core, key regions as the undertaking, and urban edges as the support. The eight major hub nodes continue to strengthen policy guidance, improve supporting guarantees, promote the construction of large and super large computing power center projects, and continuously optimize the overall computing power supply efficiency and radiation range. Promote lightweight computing power construction on the edge of cities according to demand, and accurately respond to local real-time computing power needs. In terms of the computing power industry ecosystem, the computing power industry chain is constantly improving, and the innovation of computing power core technologies is accelerating. Continuous breakthroughs have been made in fields such as power supply and distribution, refrigeration, whole machine manufacturing, and software research and development, reaching the international leading level. However, there is still a gap between the research and development of high-performance computing chips and the international advanced level. To further enhance the utilization and energy efficiency of green electricity in the computing power center, domestic energy and computing power enterprises are strengthening cooperation, promoting the application of green energy technologies such as "source grid load storage" and distributed photovoltaics, and significantly improving the utilization rate of green electricity in the computing power center. At the same time, computing power enterprises actively promote the research and application of technologies such as liquid cooling and intelligent refrigeration, accelerate the deployment and application of liquid cooled servers, and cope with the heat dissipation problems caused by the large-scale deployment of smart computing devices. Under the guidance of platforms such as the Computing Industry Development Matrix and the Open Data Center Committee (ODCC), market entities in the computing field continue to strengthen technical exchanges, project cooperation, and ecological construction. At the same time, relying on international multilateral cooperation organizations such as the China Shanghai Cooperation Organization Big Data Cooperation Center, Chinese enterprises actively participate in international cooperation in the field of computing power, orderly promote overseas computing power layout, and improve global service capabilities. In terms of the integration of computing and network development, there is a coordinated development of computing, inter computing, and intra computing networks. At the level of computing networks, edge intelligent application scenarios are rapidly growing, and promoting millisecond level low latency access has become a core demand. Telecom operators actively promote the application of edge optical transport networks (OTNs), optical service units (OSUs), network slicing and other technologies to provide users with a better computing experience. Multiple cities are accelerating the upgrading of computing networks and promoting the construction of 1ms urban computing networks. At the inter computing network level, telecom operators actively carry out the construction of direct connection networks between hub nodes, promote the application of wide area lossless technology, improve interconnection bandwidth and flexibility, and continuously optimize the computing power delay circle. At the network level, the demand for large-scale training of artificial intelligence models promotes innovation in high-speed interconnection technology between multiple nodes and devices within the computer. Domestic manufacturers continue to promote the research and development of high bandwidth and low latency interconnection technology. At the level of network scheduling, basic telecommunications operators and research institutions actively promote technological innovation and standard setting in the integration of network perception, measurement, orchestration, and scheduling, and carry out collaborative computing power and transaction platform construction, achieving positive results. At the same time, actively promoting the construction of China's computing power platform, achieving comprehensive integration of "platform, subject, resources, ecology, and scenarios", effectively supporting the automation monitoring of national computing power situation awareness, computing power supply and demand docking, and transaction matching. In terms of enabling the application of computing power, we will continue to strengthen the innovative leadership of computing power application, drive new quality productivity with computing power, further release and give play to the key role of computing power in industrial digital transformation, and comprehensively help the construction of new industrialization and Chinese path to modernization. The Ministry of Industry and Information Technology has carried out a series of work to deepen the application of computing power, promote the deep integration of computing infrastructure with various industries, empower thousands of industries, and assist in the high-quality development of the digital economy. At present, the empowerment of computing power applications is gradually deepening, from advantageous industries such as the Internet, finance and government affairs to traditional industries such as industry, energy and transportation. The Internet, government affairs, finance and other industries directly provide services to a large number of end users. In order to improve user satisfaction and constantly promote the construction of digital and intelligent service capabilities, computing power applications are relatively high. In order to better cope with fierce market competition and meet users' demands for fast, real-time, and high-quality response to products and services, traditional industries such as industry, energy, transportation, and healthcare actively promote the construction of computing power facilities and computing power platforms. They use digital technology to strengthen intelligent control, process optimization, and cost reduction and efficiency improvement, improve product and service quality, and continuously enhance the level of computing power application. With the deepening development of the digital economy, the increasing competition in technology, and the rapid iteration of artificial intelligence models, emerging computing power application service models continue to emerge, and computing power is accelerating its transition from the basic support base to the core driving force of new quality productivity. Facing the overall trend of a new international competitive landscape, accelerated rise of intelligent agent economy, and accelerated iteration of computing technology, the future development of computing power will present the following trends. The diversified computing power system is gradually improving, and the comprehensive supply capacity of computing power is steadily increasing. In the traditional computing power supply model, general-purpose computing power undertakes massive data throughput and routine business operation tasks; Intelligent computing power, relying on high parallel computing capabilities, has become the core support for large model training and inference; Supercomputing focuses on cutting-edge scientific research in aerospace, biomedicine, and other fields, continuously breaking through the limits of computing power performance; Edge computing power focuses on meeting the low latency computing needs of industrial sites, connected vehicles, and other scenarios. With the large-scale application of artificial intelligence models, the deepening of the strategy of integrating air, space, and earth, and the urgent demand for global computing power supply in the digital economy, the development of computing power is moving from a single form and local layout to a new stage of diversified collaboration, global coverage, and cloud edge integration. The construction of a national hub node for the integrated computing power network is accelerating, promoting efficient cross regional scheduling and optimization of various computing resources such as general, intelligent, and supercomputing, and solving the problem of mismatch between computing power supply and demand. At the same time, the iteration of computing power technology and the integration of scenarios are constantly deepening, and emerging forms of computing power such as space computing are gradually moving from pilot verification to a systematic development stage. The domestic independent ecosystem is accelerating its maturity, and the innovation level of the computing power industry continues to rise. Driven by multiple forces such as national strategic guidance, market demand, and technological innovation, the localization process of the computing power industry chain is accelerating comprehensively, and industrial development is moving from single point technological breakthroughs to a new stage of full stack system collaboration and from basic usability to high-performance usability. From the perspective of key links in the industrial chain, the pace of technological breakthroughs and product iterations in key areas such as infrastructure, core chips, complete equipment, and software platforms continues to accelerate. The collaborative adaptation and innovation effects of upstream and downstream links are significantly enhanced, and the bottleneck problems in some key areas have been effectively alleviated. The level of industry autonomy and controllability is steadily improving. At present, a full stack localization capability system covering "infrastructure IT equipment technical services" has been preliminarily established. The collaborative innovation mechanism of the industrial chain is becoming increasingly mature, and the standard system is accelerating its improvement, promoting the development of the computing power industry towards scale, intensification, and high-end direction. The acceleration of "computing power going global" continues, and domestic computing power enterprises continue to deepen cooperation with neighboring countries and emerging markets, actively participate in international standard setting and open source community construction, injecting new impetus into global digital economy cooperation. The integrated development of deep integration of computing and network, and the acceleration of intelligent and inclusive supply of computing power. With the rapid emergence of scenarios such as edge distributed real-time inference and 10000 card level artificial intelligence training, users' demands for flexibility in computing power scheduling, network transmission latency, and efficient resource allocation continue to increase, driving the development of computing networks from physical interconnection and unified access to logical integration and intelligent collaboration, and helping to implement intelligent on-demand and inclusive supply of computing power. The computing network is more flexible and ubiquitous, relying on user computing network technologies such as OTN, Internet Protocol (IP) wide area lossless, and IPv6 based segmented routing technology (SRv6) to optimize the user computing experience and promote the universal coverage of high-quality computing services to various industries and scenarios. The inter computing network is faster, more flexible, and more reliable. The bandwidth of the inter computing direct link continues to increase and the latency is continuously optimized, laying a solid network support for the construction of the national integrated computing power network. The intra computing network is more lossless, stable, and highly scalable. Through high-speed interconnection technology optimization, it solves the communication bottleneck of traditional clusters, improves the collaborative efficiency of computing power nodes within data centers, and meets the efficient scheduling needs of large-scale computing power clusters and trillion parameter large model training. Breakthroughs in computing network integration technology have promoted the transformation of computing power supply from "passive response" to "active adaptation", achieving on-demand call of computing power and a leap in computing efficiency, making computing power truly a new type of infrastructure that is ready to use and shared for all. Computing power empowers the development of the economy, and intelligent agents activate new driving forces for economic growth. With the iteration and upgrading of artificial intelligence technology, the new generation of artificial intelligence applications is accelerating and popularizing, and the application scenarios of computing power are continuously expanding. Gradually, it is transforming from traditional information interaction and data processing to intelligent collaboration and the creation of lexical value, becoming the core basic support for supporting the autonomous operation of intelligent agents and promoting efficient circulation of lexical elements. Intelligent agents will gradually break through the limitations of artificial intelligence interaction, possess autonomous planning, tool calling, task execution, and continuous optimization capabilities, and can efficiently support complex scenario applications, thereby significantly increasing the scale of word element consumption and generating sustained intensive computing power demand. In this context, the application of computing power is accelerating its evolution towards an integrated model of "computing power+algorithms+data+intelligent agents+lexical elements", providing a solid guarantee for the independent operation and interactive collaboration of massive intelligent agents. The coordinated development of computing and electronics is deepening, and green computing power has become an engine for sustainable development. The explosive growth of artificial intelligence models and intelligent agent applications has led to a continuous increase in demand for intelligent computing. The construction of large-scale, high-density intelligent computing facilities will continue to advance, which will bring huge energy consumption pressure. Energy will become a key bottleneck affecting the upper limit of artificial intelligence and intelligent computing development, and computer collaboration will become an important way to solve the bottleneck of computing power energy. From the perspective of computing power, with the continuous maturity of computing power load scheduling technology, computing power centers are transforming from simple energy consumers to flexible and adjustable load participants. Through mechanisms such as source load interaction, storage load interaction, and grid load coordination, they participate in peak shaving and valley filling to improve the operational efficiency of the power system. From an energy perspective, the energy system is accelerating the use of intelligent terminals and computing power platforms to achieve intelligent perception of intermittent renewable energy sources such as wind power and photovoltaics