Core Insights - Guangdong is leveraging its extensive market and diverse scenarios to promote the intelligent upgrade of manufacturing enterprises through artificial intelligence (AI) [1] - The province has introduced 16 policy measures to shift support from "providing policies" to "providing opportunities," aiming to create a globally influential "AI + manufacturing" integration development demonstration zone [1] Group 1: AI and Manufacturing Integration - Guangdong has a complete industrial system with all 31 manufacturing categories, accounting for approximately 1/8 of the national manufacturing scale, creating a "testing ground" for AI applications [2] - The province hosts over 1,600 core AI enterprises, forming a full-stack industrial chain represented by companies like Huawei and Tencent, which provides a "new engine" for the transformation of manufacturing [2] - To address the challenges of unclear demand and mismatched supply in AI applications, Guangdong has established a national-level digital transformation promotion center for manufacturing, offering one-stop services for intelligent upgrades [2][3] Group 2: Infrastructure and Support Mechanisms - Guangdong is building a "1+1+N" embodied intelligence training ground system to facilitate the development and verification of embodied intelligence models, promoting data flow and standardization [3] - The province is focusing on enhancing innovation centers in the fields of intelligent agents and intelligent terminals, leveraging the technological advantages of cities like Guangzhou and Shenzhen [4] - Guangdong has established a national-level AI application pilot base in the consumer mobile terminal sector, with plans to accelerate the construction of more pilot bases [4] Group 3: Financial Investment and Policy Support - Since 2018, Guangdong has launched five batches of "new generation AI" initiatives, focusing on key areas such as intelligent manufacturing and quality inspection, achieving multiple breakthroughs [6] - The provincial government has allocated approximately 15 billion yuan for AI-related projects since 2018, with a budget of 1.32 billion yuan for 2025 to support cutting-edge technology initiatives [8] - The province aims to invest over 20 billion yuan in AI-related projects by 2027, driving the scale of related industries to exceed 100 billion yuan [9] Group 4: Communication and Computing Infrastructure - Guangdong has built a leading information infrastructure network, with 431,000 5G base stations established by August 2025, ensuring connectivity across the province [7] - The province is implementing the "Yueqi Wanzhao" initiative to create 100 10G optical network projects, with over one-third focused on factory scenarios [7] - A unified computing power service platform has been established in the Greater Bay Area, integrating 18 computing resource pools with a total capacity exceeding 14,000P [7]

Core Insights - Guangdong plans to establish 100 10G optical network projects within this year, with over one-third of these projects focused on factory scenarios [1][2] - The province aims to enhance the intelligent manufacturing foundation through three main aspects: building information infrastructure, promoting collaborative computing and networking, and integrating data with reality [1][2][3] Group 1: Information Infrastructure - As of August 2025, Guangdong has built 431,000 5G base stations and 2,529,000 10G PON ports, leading in scale across China, achieving "gigabit access in every town and 5G coverage in every village" [1][2] - The network construction includes a 400G all-optical network connecting cities in the Guangdong-Hong Kong-Macao Greater Bay Area, targeting ultra-low latency of 1 millisecond within Shaoguan, 3 milliseconds across the Bay Area, and 6 milliseconds province-wide [2] Group 2: Collaborative Computing and Networking - The province is advancing the establishment of a computing hub in Shaoguan, with telecom data centers capable of supporting a capacity of 10,000 cards [2] - The Qianhai integrated computing service platform has been developed, connecting 18 computing resource pools with a total computing scale exceeding 14,000 P, offering services such as computing scheduling, trading, application, and security [2] Group 3: Data and Reality Integration - By August 2025, Guangdong has established over 7,600 5G industry virtual private networks, with nearly 27 million 5G IoT terminal users [2] - The province is promoting the "Hundred-Thousand" action for 5G factories, with 79 factories recognized in the national 5G factory directory over three years [2] - The upgrade of the national top-level node for identification resolution in Guangzhou aims to standardize key processes and ensure interconnectivity of upstream and downstream data [2][3] Group 4: Future Initiatives - Future initiatives will focus on accelerating the deployment of 5G-A and 10G optical networks in manufacturing parks, optimizing computing infrastructure layout, and providing comprehensive AI services to support intelligent manufacturing [3]

Core Insights - The Chinese government is initiating a "millisecond computing" network construction project to enhance the urban computing network by 2027, aiming for high efficiency and quality in computing services [1][2] Group 1: Network Infrastructure Development - The initiative focuses on building a high-capacity, low-latency computing network within urban areas, targeting interconnection delays of less than 1 millisecond between computing centers and less than 10 milliseconds from terminals to computing applications [2][3] - By 2027, the goal is to achieve a 50% deployment rate of high-speed optical transmission systems (400Gbps and above) and a 50% deployment rate of all-optical cross-connect devices at key sites [2] Group 2: Technological Innovation and Integration - The action plan encourages local regions to conduct network technology validation and pilot projects, emphasizing the development of key technologies such as all-optical high-speed lossless transmission and intelligent operation of computing networks [2][3] - The initiative aims to create a "national computing network" that facilitates low-latency and efficient data flow between cities and computing centers [2] Group 3: Application in Key Industries - The project will promote the integration of computing networks into critical sectors such as manufacturing, finance, transportation, healthcare, education, and entertainment, encouraging telecom companies to offer customized computing network services [3][4] - By 2027, the coverage rate of urban computing centers with a 1 millisecond latency will reach 70%, establishing a sustainable foundation for computing network access [3][4]

Core Viewpoint - The Ministry of Industry and Information Technology has launched a special initiative to develop a millisecond-level computing power network, aiming to enhance the responsiveness of computing resources within urban areas by 2027 [1][4]. Group 1: Millisecond Computing Power Network Development - The initiative focuses on creating a high-speed, low-latency network that allows for millisecond-level access to computing resources, addressing the gap between traditional network delays and real-time computing needs [1][4]. - By 2027, the goal is to establish a comprehensive urban millisecond computing power network capable of providing connectivity with less than 1 millisecond latency for computing centers and less than 10 milliseconds for application terminals [1][4]. Group 2: Actions to Achieve Goals - The initiative outlines three main actions: enhancing interconnectivity between computing centers, improving access to computing resources, and optimizing application responsiveness [1][4]. - Specific targets include achieving a unidirectional latency of less than 1 millisecond between medium and large computing centers by 2027, with at least 50% deployment of 400Gbps capacity [2][3]. Group 3: Technological Innovations and Infrastructure - The initiative emphasizes the need for advanced technologies such as task scheduling and intelligent network operation to facilitate dynamic coordination between computing and network resources [3][4]. - It also aims to enhance the performance of computing centers through the deployment of all-optical networks and innovative network protocols, improving communication efficiency for applications like artificial intelligence [4][5]. Group 4: Industry Impact and Applications - The millisecond-level computing power network is expected to drive commercial breakthroughs in high-value scenarios, such as autonomous driving and telemedicine, by significantly reducing response times [5]. - The initiative supports the "East Data West Computing" strategy by minimizing data transmission delays, thereby improving resource utilization efficiency across regions [4][5].

Core Viewpoint - The Ministry of Industry and Information Technology (MIIT) has announced a special action plan to develop a low-latency computing network, aiming for a one-way interconnection delay of less than 1 millisecond between medium and large computing centers in urban areas by 2027, which is expected to enhance the importance of computing power in the AI industry [1][3][4]. Group 1: Policy and Infrastructure Development - The MIIT's notification emphasizes the construction of a high-speed, low-latency computing network, targeting a millisecond-level resource network capability within urban areas [3][4]. - By 2027, the plan aims for a deployment rate of at least 50% for 400Gbps connections between medium and large computing centers, with a similar target for all-optical cross-connects at key urban sites [3][4]. - The initiative includes the development of innovative network technologies and protocols to enhance the operational efficiency of computing networks [4]. Group 2: Industry Opportunities and Trends - Multiple brokerages express optimism about long-term investment opportunities in the computing power sector, particularly in light of recent favorable policies [6][7]. - The MIIT and other departments have outlined plans to strengthen new information infrastructure and promote the integration of computing power with various industries, including manufacturing and finance [7]. - The domestic AI computing ecosystem is evolving, with significant orders for servers from major banks and telecom companies, indicating a strong market for domestic computing power suppliers [8]. Group 3: Market Dynamics and Projections - Analysts predict a rise in domestic computing power driven by increased capital expenditure from large enterprises and a growing consensus on the need for localization [9]. - The light communication sector is expected to see substantial growth, with a projected 167% year-on-year increase in net profit for the third quarter [10]. - The demand for optical modules, particularly 800G, is anticipated to continue its rapid growth, benefiting leading manufacturers in the sector [10].

Core Viewpoint - The Ministry of Industry and Information Technology (MIIT) has launched a special action plan to develop a metropolitan "millisecond computing" network, aiming for a coverage rate of at least 70% by 2027, enhancing the efficiency and accessibility of computing resources across the country [1][3]. Group 1: Overall Strategy and Goals - The initiative focuses on developing a high-speed, low-latency metropolitan network that provides millisecond-level computing resource accessibility, with specific targets for interconnection and access times [4]. - By 2027, the plan aims to achieve less than 1 millisecond latency for interconnections between medium and large computing centers and less than 10 milliseconds for applications to computing centers [4]. Group 2: Action Content - The plan includes three main actions: 1. Promoting "millisecond interconnection" between computing centers, with a target of 50% deployment of 400Gbps connections by 2027 [5]. 2. Enhancing "millisecond access" to computing resources, aiming for a coverage rate of at least 70% for 1 millisecond latency circles by 2027 [5]. 3. Improving "millisecond accessibility" for applications, targeting a network latency of less than 10 milliseconds from application terminals to computing servers by 2027 [6]. Group 3: Implementation Requirements - Local communication management departments are tasked with selecting cities with high computing resource density and good network infrastructure to implement the action plan, with a maximum of three cities per province [7]. - A mid-term evaluation will be conducted one year after the action's initiation, requiring local departments to report network monitoring data to the national platform [7]. - At the end of the action period, a summary of the implementation will be required, with results evaluated and published by the MIIT [8].

Core Insights - The 2025 China Mobile Global Partner Conference showcased the latest achievements and future plans of AI technology empowering various industries, with a focus on domestic chip solutions from Haiguang Information [1][3] - Haiguang Information presented its CPU, DCU (AI accelerator), and AI cluster solutions, emphasizing deep collaboration with China Mobile to provide mature practices and advanced directions for domestic computing power [1][3] Group 1: Product Offerings and Applications - Haiguang's computing products, including CPUs and DCUs, serve as the foundation for numerous digital services such as online shopping, mobile payments, and real-time video processing [3] - The "Water Resources Big Model Integrated Machine," developed in collaboration with China Mobile, utilizes Haiguang's DCU for flood warning and disaster prevention, demonstrating practical applications in various sectors [3][4] - Haiguang has also launched multiple integrated machines in voice intelligence, robotics, and government writing, indicating the extensive application of its DCU in digital transformation across industries [4] Group 2: Technical Innovations - Haiguang's C86 architecture addresses compatibility challenges faced by traditional software based on x86 architecture, allowing seamless migration without performance loss [4][5] - The C86 architecture is fully compatible with the x86 ecosystem, enabling efficient adaptation across various operating systems and cloud platforms, thus supporting millions of mainstream software applications [4][5] - Haiguang's CPU includes an independent security module that supports national encryption algorithms and privacy computing, enhancing data security in communication scenarios [7] Group 3: Future Plans and Vision - Over the next three years, Haiguang plans to increase R&D investment to accelerate the iteration of CPU and DCU technologies while fostering an open ecosystem for deeper software-hardware integration [6][7] - The company aims to create a flexible computing network that supports various scenarios, from data center model inference to real-time edge computing [8] - Haiguang envisions a future where computing power is as accessible as utilities, facilitating the digital transformation of various industries through intelligent connectivity [8]

Core Insights - The 2025 China Mobile Global Partner Conference was held in Guangzhou from October 10 to 12, showcasing Haiguang Information as a key partner with its CPU+DCU products and comprehensive solutions [1] Group 1: Product Highlights - Haiguang's AI cluster solution offers efficient support for super node systems, enhancing AI application storage performance and liquid cooling system design [3] - The high-performance interconnect bus protocol developed by Haiguang breaks down barriers between CPUs and accelerators, featuring high bandwidth, low latency, and cache coherence [3] - Haiguang's multi-dimensional security solution aims to safeguard the construction of the computing power ecosystem [1][3] Group 2: Strategic Collaborations - Haiguang has established a nationwide IT cloud, CT cloud, and public cloud computing base in collaboration with three major telecom operators [3] - The company plans to continue developing a more flexible computing-network integration system to provide reliable "core" solutions for communication business innovation [3]

Core Insights - The rapid growth of AI applications, such as large model training and smart driving, is driving the demand for ultra-high bandwidth, low latency, and high reliability in network transmission solutions, leading to the emergence of "all-optical capacity" as a core infrastructure for the digital economy [1][2][3] Industry Trends - The demand for computing power is expected to grow significantly, with China's intelligent computing power projected to reach 1,037.3 EFLOPS by 2025, a 43% increase, and 2,782 EFLOPS by 2028, with a compound annual growth rate exceeding 40% [2] - The integration of optical technology into the entire computing power chain is essential to overcome traditional network limitations in bandwidth, latency, and reliability [2][3] Company Developments - Major companies like China Unicom and China Mobile are actively enhancing their optical networks to support low-latency computing, with initiatives such as the "AINet" and "Jiuzhou" projects aimed at improving capacity and coverage [4][5] - Huawei is promoting the "Optical Intelligence Integration" concept, focusing on building a millisecond-level capacity foundation through advanced optical network technologies [5] Market Performance - The optical module sector is experiencing significant growth, with companies like Zhongji Xuchuang reporting a 46% year-on-year increase in optical communication module sales, driven by strong demand for high-end products [6] - NewEase achieved remarkable revenue growth of 282.64% year-on-year in the first half of 2025, indicating a robust market for optical modules [7]

Core Insights - The 8th Industrial Internet Innovation Development Conference will be held in Huangshi on October 24, 2025, focusing on the digital transformation of small and medium-sized enterprises (SMEs) and the integration of industrial internet technologies [1][5] - The conference aims to promote the deep integration of artificial intelligence and industrial internet, enhancing the quality and efficiency of manufacturing and SMEs [3][4] Group 1: Conference Overview - The conference will feature a theme of "AI Integration, Digital Intelligence Rebirth," focusing on three dimensions: "Artificial Intelligence + Industrial Internet," promotion of SME digitalization, and innovation in network integration [3][4] - A diverse activity matrix will be set up, including an opening ceremony, three thematic forums, and three major exhibition areas to showcase advancements in industrial internet and AI applications [3][4] Group 2: Industry Development - Huangshi has selected 346 provincial pilot enterprises for digital transformation, with 190 already completed, and has cultivated 36 national-level industrial internet pilot honors [1][4] - The city has initiated the construction of "10 Gigabit Cities" and "AI Cities," significantly accelerating the development of new productive forces [1][4] Group 3: Goals and Impact - The conference aims to establish a replicable model for SME digital transformation in Huangshi, contributing to high-quality development in China's manufacturing sector and new industrialization [5] - The event will gather experts and representatives from various fields to share the latest technological research and practical applications, fostering innovation and collaboration [4][5]