On July 28th, Tianjin Mobile's TPU intelligent computing center was officially lit up, significantly reducing the unit computing power cost; On September 26th, the first application of Moore Thread was successfully approved, and the research and commercialization progress of the company's AI chip products are expected to accelerate; On October 11th, China Mobile announced that it will build the largest and technologically advanced intelligent computing infrastructure in China by the end of 2028... In recent times, competition for domestic computing power has become increasingly fierce. The industry believes that in the context of national policies promoting the construction of a nationwide integrated computing power network and continuous technological breakthroughs, it is expected to drive the construction of domestic computing infrastructure, and listed companies in the industry chain that have taken the lead in layout are worth paying attention to. Technological breakthrough computing power refers to the ability to process data, which is a new type of productivity that integrates information computing power, network carrying power, and data storage power. It is the fundamental support for digital industrialization and industrial digitization. In recent times, domestic computing power has continuously made breakthroughs in areas such as production capacity, key components, and large model adaptation. On July 28th, at the 2025 World Artificial Intelligence Conference (WAIC), Zhonghao Xinying, Tianjin Mobile, Taiji Co., Ltd., and Bozhong Shuzhi jointly held the Tianjin Mobile TPU Intelligent Computing Center lighting ceremony. According to Zhonghao Xinying's disclosure, the self-developed "Taize" GP TPU artificial intelligence server provided by Zhonghao Xinying has provided the core AI computing power efficiency for the project. Among them, the TPU chip "Instant" carried by "Taize" was successfully developed by Zhonghao Xin Yingli Shi in the past 5 years. It is reported that the AI computing power performance of "Instant" surpasses famous overseas GPU products by nearly 1.5 times, reducing energy consumption by 30%. Based on comprehensive price, computing power, and energy consumption calculations, the unit cost is only nearly half of similar products. In addition, the unique 1024 chips of "Instant" enable high-speed interconnection between optical modules, resulting in system cluster performance that far exceeds traditional GPUs by several times, and can support the computing needs of over 100 billion parameter large models. The competition of computing power is actually a competition of chips, architecture, energy consumption, and efficiency. When selecting the underlying computing hardware, the computing performance, power consumption, and usability of the chip are the core factors that Tianjin Mobile focuses on. Energy consumption determines long-term operating costs, cost-effectiveness determines whether it can be widely popularized, and usability determines whether developers are willing to stay. ”Gao Kang, the project construction leader of Tianjin Mobile TPU Intelligent Computing Center and the deputy general manager of Tianjin Mobile City Third Branch, said. On July 30th, Ziguang Hengyue announced that it had successfully delivered the UNIS Galaxy Large Model Integrated Machine equipped with the "dual core" architecture of Haiguang to a domestic intelligent computing platform project. This not only marks the further verification of the technology integration and scenario landing capabilities of both parties, but also provides key support for the demanding computing requirements of complex scenarios in the aviation industry with independent computing solutions. In addition, on September 26th, Moore Thread, which focuses on the research and development of fully functional graphics processing unit (GPU) chips and related products, successfully applied for the first release, which is expected to accelerate the research and commercialization process of AI chip products. Since the beginning of this year, policies have been continuously strengthened, gradually building a full chain and multi-level computing power development support system, injecting strong momentum into the industry from scale expansion to quality and efficiency improvement. On August 21st, the State Council issued the "Opinions on Deepening the Implementation of the" Artificial Intelligence+"Action", proposing to optimize the layout of national intelligent computing resources, improve the national integrated computing power network, fully play the role of the "East West Computing" national hub, and increase the coordination of resources such as data, computing, electricity, and network. On September 26th, the National Development and Reform Commission and other departments issued the "Several Measures to Strengthen the Cultivation of Innovative Enterprises in the Digital Economy", which emphasized the need to strengthen the supply support of computing resources, deeply implement the "East West Computing" project, accelerate the construction of a national integrated computing power network, and guide various computing resources to efficiently and accurately connect with the needs of digital innovation enterprises. On October 11th, the Ministry of Industry and Information Technology and seven other departments publicly issued the "Implementation Plan for Deepening the Innovation and Development of Service oriented Manufacturing (2025-2028)", which clearly stated the need to consolidate the foundation of the development of service oriented manufacturing. Strengthen the construction of new information infrastructure, deepen the integration, innovation and large-scale application of "5G+Industrial Internet", layout computing infrastructure as required, and accelerate the integration of computing and industry applications. In addition to promoting the construction of a national integrated computing power network, policies also guide the penetration of computing power resources into more real economic sectors, accelerating the application and landing of computing power in industries such as intelligent manufacturing, smart healthcare, and transportation. For example, on September 26th, the Ministry of Transport and other departments issued the "Implementation Opinions on 'Artificial Intelligence+Transportation'", proposing to coordinate and optimize the supply capacity of computing power. Relying on major transportation infrastructure such as highways, ports, and transportation hubs, we will strengthen computing power support according to local conditions. The national policy promotes the construction of a nationwide integrated computing power network, which is expected to drive the development of domestic computing infrastructure. ”Open source securities stated in the research report. Listed companies have taken the lead in layout, and Huajin Securities stated in its research report that domestic computing power is rapidly developing with a significant increase in growth rate. According to the "Assessment Report on the Development of Artificial Intelligence Computing Power in China" jointly released by International Data Corporation (IDC) and Inspur Information, it is expected that by 2025, the scale of China's intelligent computing power will increase by 43% compared to 2024, and the market size of artificial intelligence computing power will reach 25.9 billion US dollars, an increase of 36.2% compared to 2024. Against the backdrop of policy efforts and continuous technological breakthroughs, many listed companies have increased their layout in the field of computing infrastructure construction to seize industry growth opportunities. On October 11th, at the main forum of the 2025 China Mobile Global Partner Conference, China Mobile officially released the upgraded "AI+" action plan and the "AI+" ecological alliance. By the end of 2028, China Mobile will continue to increase its investment in the field of artificial intelligence, doubling its overall investment and building the largest and technologically advanced intelligent computing infrastructure in China. Recently, Teruide revealed on an interactive platform that in the field of computing infrastructure, the company has innovatively developed power modules for data centers based on their needs and characteristics. The prefabricated solution is used to promote standardized deployment of products by integrating transformers UPS、 Distribution cabinets and other equipment greatly shorten the construction period, reduce construction costs, and improve delivery efficiency. At present, the company has established good cooperation relations with many domestic head data centers, including China Mobile, China Telecom, Alibaba, Pross, ByteDance and other well-known enterprises. Dongfang Guoxin stated that the AI intelligent computing center independently developed and constructed by the company has formed a large-scale intelligent computing power service capability. At the same time, the company has strengthened its strength in the field of AI computing hardware and platforms through strategic investments such as Shituo Cloud (Zhongke Shituo), which can support the management of computing resources including various domestic AI acceleration chips. Li Ang Technology claims that relying on the "computing power+algorithm+application" ecosystem and high-performance clusters of newly built computing power bases in Jianyang, Sichuan, the company can provide computing power support for AI large model training, reasoning and other scenarios, and supply hardware facilities such as servers and storage devices. At the same time, the company's technical operation and maintenance services cover real-time traffic monitoring of data centers, cross regional computing power scheduling, and specialized operations such as installation, debugging, and troubleshooting of GPU servers, ensuring the stable operation of its business. ZTE Corporation stated that in the field of computing power, facing the new opportunities brought by AI big models, the company has strengthened the research and development of intelligent computing products and industrial ecological cooperation. Through self-developed chips, intelligent computing servers, ultra large scale clusters, intelligent computing all-in-one machines and other full stack computing infrastructure, as well as resource management platforms, training and promotion platforms, Nebula big models, intelligent agent factories and other software solutions, the company has collaboratively created an open and inclusive intelligent computing base, providing end-to-end, full stack full scene intelligent computing solutions to meet diverse AI scene needs. First Shanghai Securities stated that it is optimistic about the sustained high growth in computing power demand driven by AI applications, and that AI applications at home and abroad are entering a turning point of popularization. On the domestic computing power side, the bottleneck of computing capacity has been overcome, and it is expected that domestic chips will see a surge in volume by 2026. Overseas computing power, Open AI and other overseas AI giants are accelerating commercialization, promoting the widespread implementation of AI applications, and maintaining high demand for computing hardware. (New Society)