Core Insights - The report highlights the ongoing growth and investment in the AI computing infrastructure, driven by both domestic and international demand for AI applications [3][6][10] - The report emphasizes the strong performance of Industrial Fulian, with significant revenue and profit growth in the first half of 2025, indicating a robust business model and market position [9][11] AI Computing Industry - The AI computing supply chain is expected to accelerate, supported by rising domestic advanced process yields and collaborative efforts among various industry players to build a robust AI computing foundation [3][6] - Alibaba plans to invest over 380 billion yuan in cloud and AI hardware infrastructure over the next three years, indicating a strong capital expenditure trend in the domestic AI sector [6] - The report notes that the yield of Huawei's Ascend 910C chip has reportedly reached 40%, reflecting significant improvements in manufacturing processes and management [7] - The Chinese government is actively promoting the "Artificial Intelligence +" initiative, aiming to enhance the commercialization of AI technologies across various sectors [7] Industrial Fulian - Industrial Fulian reported a revenue of 360.76 billion yuan and a net profit of 12.11 billion yuan in the first half of 2025, representing year-on-year growth of 35.58% and 38.61% respectively [9] - The company’s cloud computing business saw server revenue growth exceeding 50%, with AI server revenue increasing by over 60% year-on-year [11] - The report anticipates significant capital expenditure growth from major North American cloud service providers in 2025, particularly in AI infrastructure, which will drive demand for high-end AI servers [10]

Investment Rating - The industry investment rating is "Outperform the Market" [12] Core Viewpoints - The AI computing power supply chain is accelerating, driven by advanced processes and policies supporting domestic production. The report anticipates a rapid increase in domestic AI applications and a robust ecosystem [1] - The report highlights that with the improvement in the yield of advanced processes, the supporting supply chain is expected to see accelerated growth, maintaining the "Outperform the Market" rating [1] - The report emphasizes the importance of domestic AI computing power as a foundational support for AI technology, with significant capital expenditure expected in the domestic AI sector [5] Summary by Sections Investment Suggestions - Recommended companies to focus on include: - Computing Chips: Cambrian, Haiguang Information - Advanced Manufacturing: SMIC, Huahong Group, Yongxi Electronics - HBM: Huahai Chengke, Lianrui New Materials - PCB/CCL: Shenzhen Circuit, Founder Technology, Nanya Technology, Shengyi Technology - High-speed Backplane Modules: Huafeng Technology - Power Capacitors: Taijia Co., Oulu Tong, Jianghai Co. - Liquid Cooling: Feirongda, Shenling Environment, Shuguang Digital Innovation, Hongrid [3] Supporting Rating Points - The report notes that the yield of Huawei's Ascend 910C has reportedly reached 40%, indicating significant improvements in manufacturing processes and management, which is expected to lead to increased shipments of the 910C series products [5] - The report mentions the Chinese government's commitment to implementing the "Artificial Intelligence +" initiative, which aims to promote the large-scale commercialization of AI applications across various sectors, leveraging China's complete industrial system and large market scale [5]

Investment Rating - The report does not explicitly provide an investment rating for the green computing industry Core Insights - The report emphasizes the necessity of developing green computing to address the increasing energy consumption and carbon emissions associated with computing power, which is seen as a critical component of high-quality development in the digital economy [7][21] - It highlights the significant growth in computing power in China, with a total scale reaching 280 EFLOPS by the end of 2024, and an average growth rate of nearly 30% over the past five years [7][9] - The report outlines the integration of green computing into various sectors, including energy, manufacturing, retail, transportation, and construction, with over 86% of central enterprises adopting cloud services [9][21] Summary by Sections 1. Background of Green Computing Research - The report identifies the exponential growth in energy demand for computing infrastructure, driven by advancements in AI technologies and distributed computing [19][20] - It discusses the contradiction between the increasing demand for computing power and the limited environmental capacity, necessitating a transformation towards green computing [21][22] 2. Overall Progress in Green Computing - The report notes a tightening of global policies regarding energy consumption and carbon emissions in the computing sector, with various countries implementing stringent regulations [29][30] - In China, policies are shifting from intensive construction to a collaborative model focusing on efficiency, energy consumption, and application [30][31] 3. Development Trends in Computing Equipment - The report highlights the rapid growth of AI servers as a key driver of energy consumption in computing equipment, with global data center IT loads expected to increase significantly by 2030 [48][49] - It mentions that China's computing equipment efficiency is improving, with advancements in chip manufacturing and cooling technologies [51] 4. Industry Dynamics - The report outlines the green computing industry chain, which includes hardware suppliers, energy providers, infrastructure builders, and application customers [37] - It emphasizes the importance of energy efficiency technologies, such as liquid cooling, which is projected to grow significantly in the coming years [41][42] 5. Corporate Practices - The report details various corporate initiatives aimed at achieving carbon neutrality and enhancing energy efficiency, with major tech companies committing to renewable energy usage [43][45] - It highlights the proactive measures taken by Chinese telecom operators to promote green computing and energy collaboration [45][46]

Investment Rating - The report maintains an "Outperform" rating for the telecommunications industry [2] Core Insights - The report highlights significant advancements in AI infrastructure, particularly with the introduction of the CloudMatrix384 system, which marks a transition from "catching up" to "leading" in AI capabilities [10] - The global AI server market is expected to reach $298 billion by 2025, accounting for over 70% of the total server market, indicating a robust growth trajectory and increasing opportunities for domestic alternatives [11] - The report emphasizes the collaborative effects within the industry, suggesting that core suppliers are likely to experience performance elasticity due to rising demand for supporting equipment [12] Summary by Sections Investment Strategy - Technological breakthroughs are reshaping the AI computing landscape, with the CloudMatrix384 system setting a new industry benchmark [10] - The global AI server market is poised for explosive growth, with significant domestic replacement opportunities emerging [11] - The synergy within the supply chain is becoming more pronounced, leading to enhanced performance for key suppliers [12] Industry News - Major overseas companies continue to demand AI computing power, sustaining a high level of industry prosperity [13] - The U.S. has released a national AI strategy aimed at establishing dominance in the global AI landscape through innovation, infrastructure, and geopolitical positioning [14] - China Mobile's AI model has achieved top rankings in international assessments, showcasing advancements in human-computer interaction technology [16] - The acceleration of satellite internet construction is indicated by a recent tender for rocket launch services worth 1.336 billion yuan [17] Weekly Review and Focus - The telecommunications sector experienced a slight decline of 0.47% this week, underperforming compared to major indices [18] - Key stocks to watch include those in the AIDC and related supply chains, as well as long-term focus on major telecom operators and infrastructure providers [22]

Group 1 - The report highlights the significant progress made during the recent China-US trade talks in Geneva, with both sides agreeing to lower tariffs and establish a negotiation mechanism for future discussions [3][25][26] - The Chinese government emphasizes the importance of maintaining high-quality development and expanding international cooperation to navigate external uncertainties [3][30][31] - The report suggests that the trade agreement is a rational choice based on the current economic realities of both countries, indicating a shift towards pragmatic dialogue [3][27][28] Group 2 - The communication equipment industry is experiencing a recovery, with revenue growth of 2.9% and 12.9% year-on-year in Q4 2024 and Q1 2025, respectively, driven by AI computing power [5] - The net profit for the communication equipment sector showed a decline of 14.1% in Q4 2024, followed by a 24% increase in Q1 2025, indicating a positive trend in profitability [5] - The report recommends focusing on investment opportunities in "AI + overseas + satellite" sectors, particularly in areas like optical modules and liquid cooling technologies [5] Group 3 - The electronics sector is witnessing a recovery, with a total fund holding value of approximately 234.4 billion, representing 18.09% of the market, with a notable increase in semiconductor holdings [6] - Consumer electronics companies reported a significant revenue growth rate exceeding double digits in Q1 2025, supported by government subsidies and tariff preparations [6] - The AI performance dividends are becoming evident in passive components and PCB sectors, with automotive and AI computing emerging as new growth drivers [6] Group 4 - The automotive parts industry is focusing on the growth of electric vehicle components, with sales in the new energy sector reaching 7.249 billion yuan in 2024, a year-on-year increase of 21.3% [10] - The report indicates that the company is strategically positioning itself in the robotics sector, leveraging its technological advantages to expand into this emerging market [10] - The projected revenue for the company from 2025 to 2027 is expected to grow significantly, with net profits forecasted to reach 1.58 billion yuan by 2027 [10] Group 5 - The report on the company 盛业 (06069) indicates a strategic partnership with investors to enhance its supply chain technology business, focusing on AI and e-commerce sectors [38][39] - The company aims to leverage its collaboration with strategic investors to expand its market presence and improve its AI research capabilities [38][39] - The expected revenue growth for the company from 2025 to 2027 is projected to be substantial, with net profits anticipated to reach 836 million yuan by 2027 [38]

行业报告 | 行业研究周报 消费电子 证券研究报告 海外科技股 25Q1：Capex 保持较高强度，看好 AI 算力链修复 端侧 AI：AI 红利开始在业绩端兑现。1）苹果营收和利润同比增长 5%，服务与 iPad 业务表现突出，尽管大中华区 营收下滑，但公司积极通过全球生产布局应对关税带来的成本上升；2）高通创下营收历史新高，车用与物联网芯片 业务强劲增长，推动整体业绩上扬；3） 联发科则受益于 AI 旗舰芯片升级和市场需求增长，实现营收同比增长近 15%，虽毛利率因基期效应略有下滑，但预期 AI 产品组合将继续提升市场份额与产品均价。 云侧 AI：资本开支强度保持，产业链信心有望迎来修复。Meta、微软和亚马逊均实现了稳步增长，并在资本开支上 大幅加码，重点投入 AI 数据中心、云服务及相关技术基础设施建设。1） Meta 的 25Q1 收入增长 16%，主要由应用 家族驱动，并大幅提升资本支出以支持 AI 基础设施建设；2） 微软则通过 Azure 服务的强劲增长以及自研 AI 芯片 的投入，进一步推动 AI 技术的发展，并提高了资本支出的预期；3） 亚马逊25Q1 净销售额增长 9%，AWS 云服务 ...

分析师：方竞 分析师：李萌 邮箱：fangjing@mszq.com 邮箱：limeng@mszq.com 电子行业点评 384 架构引领，910C 蓄势待发 2025 年 04 月 29 日 [Table_Author] 执业证号：S0100521120004 执业证号：S0100522080001 ➢ 事件：据韶关发布 4 月 26 日消息，全球首个商用智算昇腾超节点在中国电 信粤港澳大湾区（韶关）算力集群正式上线。本次昇腾超节点的商用落地，标志 着国产 AI 算力在实际应用场景中的渗透步伐进一步加快，国产算力迎来破局元 年。 ➢ CloudMatrix 384 商用上线，国产算力破局元年。从发布历程来看， CloudMatrix 384 于今年 4 月 10 日在华为云生态大会上正式发布，在面向 AI 时代海量算力需求背景下，华为云实现从服务器级到矩阵级的资源供给模式转 变。目前，CloudMatrix 384 超节点已在芜湖地区规模部署，并在实际应用中展 现出强劲性能。在 70 亿参数规模的 DeepSeek-R1 模型训练中，相较于 H100 集群，迭代速度提升 1.8 倍，高横向带宽架构有效突破了 ...