中科曙光总裁历军 图片来源：每经记者 于垚峰 摄 作为深耕超级计算领域近30年的领军企业，中科曙光正通过构建产业生态、突破核心技术等举措，破解算力领域创新难题；联合上下游合作伙伴，实现软硬 件核心技术突破，为VR等产业发展注入算力动能。 "超智融合"趋势下，算力产业的变革与挑战 每经记者｜于垚峰 每经编辑｜董兴生 10月19日，"2025世界VR产业大会"在江西南昌开幕。会上，中科曙光总裁历军围绕"VR让世界更精彩——VR+AI开启数智未来"的大会主题做主旨演讲。他 指出，随着VR（虚拟现实）产业与AI（人工智能）技术的深度融合，算力需求呈现井喷式增长，传统超算正向数据模型驱动的智能形态转型，"超智融 合"成为产业发展的核心方向。 随着人工智能技术飞速迭代，VR产业与AI技术加速融合，算力需求迎来井喷式增长。作为高端算力领域的核心企业，中科曙光深耕该领域已近30年。 历军在发言中强调，当前行业内看似独立的"智算"与"超算"，实则存在紧密关联：如今的智能计算，本质是传统高性能计算在人工智能应用场景下的分支与 延伸。而在虚拟现实、数字孪生、元宇宙等前沿领域，物理模型与数据模型正逐步耦合，天然呈现"超智融合"特征。 ...

Group 1 - The report highlights the strategic acquisition of Venator UK's titanium dioxide business assets by Longbai Group, aiming to enhance its European operations and global footprint [6][7][8] - Longbai Group plans to establish subsidiaries in Malaysia and the UK, investing $5 million and $50 million respectively, to further its globalization strategy [4][9] - The titanium dioxide market is experiencing price increases, with the current market price for sulfate titanium dioxide ranging from 12,700 to 13,800 RMB per ton, indicating a positive market trend for Longbai Group [10] Group 2 - Haiguang Information reported a significant revenue growth of 54.65% year-on-year for the first three quarters of 2025, with a notable acceleration in Q3 revenue growth reaching 69.60% [12][13] - The company is expanding its market presence through partnerships with key industry players, which is driving its revenue and profit growth [13][15] - Haiguang's CPU and DCU segments are expected to see substantial growth, particularly with increasing demand from government and internet sectors [16] Group 3 - The report indicates that the advertising business of Baidu is under short-term pressure, while AI SaaS is positively impacting cloud services [31][32] - Baidu's core revenue is projected to decline by 8% in Q3 2025, with online marketing revenue expected to drop by 22% [31][34] - The company is focusing on AI-driven search transformation to enhance user engagement and satisfaction, which may provide long-term growth potential [32] Group 4 - The report discusses the mechanical industry, specifically the Japanese motorcycle market, which saw a total production of 639,000 units in 2024, a decrease of 6.4% year-on-year [21] - The domestic sales of motorcycles in Japan totaled 368,000 units in 2024, reflecting a decline of 9.2% year-on-year [22] - The report emphasizes the export dynamics and overseas production capacity of Japanese motorcycle manufacturers, with a total export volume of 484,000 units in 2024 [24] Group 5 - Zhongke Shuguang reported a revenue of 8.804 billion RMB for the first three quarters of 2025, with a year-on-year growth of 9.49% [26][27] - The company's non-recurring net profit increased by 66.79% year-on-year, indicating a significant improvement in its core business profitability [27][28] - The launch of the Shuguang AI supercluster system is expected to enhance the company's competitive position in the AI computing infrastructure market [28] Group 6 - The report indicates that the coal price is expected to show a long-term upward trend due to rising production costs and increased taxation [43][46] - The coal industry has experienced a consistent upward price trend over the past 30 years, driven by factors such as labor costs and environmental investments [44][46] - The report suggests that the coal price will continue to face upward pressure in the long term, despite potential fluctuations [46] Group 7 - The report highlights the growth of the cloud computing and network security sectors, with a focus on AI-driven opportunities for companies like Deepin Technology [47] - Deepin Technology has seen a significant increase in its cloud computing revenue, which now constitutes 46.36% of its total revenue, reflecting its strategic shift towards cloud services [47]

Core Insights - The demand for AI computing power is accelerating from single-point breakthroughs to system-level integration, with "super nodes" emerging as a new product form to overcome traditional computing bottlenecks [1][2] - Major Chinese tech companies are leading the development of super nodes, with Huawei and Alibaba launching advanced AI server products that significantly enhance computing capabilities [1][4] Group 1: Super Node Development - The super node architecture is defined as an AI system composed of AI computing nodes interconnected through high-speed protocols, supporting 32 or more AI chips with bandwidth of at least 400GB/s [2] - Huawei's CloudMatrix384 super node integrates 384 Ascend NPUs and 192 Kunpeng CPUs, achieving a single-card inference throughput of 2300 Tokens/s [1][3] - Alibaba's new generation Panjiu 128 super node AI server features self-developed CIPU 2.0 chips and supports 128 AI computing chips in a single cabinet [1][4] Group 2: Global AI Infrastructure Trends - Global tech giants like NVIDIA, OpenAI, and Meta are accelerating AI infrastructure development, with significant investments planned for the coming years [1][7] - OpenAI has partnered with AMD to deploy 6 gigawatts of AMD GPU computing power, and plans to use NVIDIA systems for its next-generation AI infrastructure [7][8] - NVIDIA executives project that AI infrastructure spending will reach $3 trillion to $4 trillion by 2030, indicating a robust growth trajectory for the sector [9][10] Group 3: Industry Challenges and Opportunities - The AI computing landscape faces challenges such as communication walls, power consumption, and complexity, necessitating the development of super nodes [2] - The core challenge for the domestic computing industry lies in the maturity of the ecosystem, despite opportunities arising from advancements in chip manufacturing and other related fields [5][6]

Core Insights - The report from Guohai Securities highlights the successful integration of the Anhui Province computing power scheduling platform with four national-level computing power platforms, adding six data centers including the Yangtze River Core (Zhongke Shuguang Intelligent Computing Center) [1] Group 1: Industry Developments - The "Star Gate" project in China is emerging, indicating a gradual realization of a domestic computing power industry closed loop [1] - The Anhui computing power public service platform has successfully connected with four major platforms and six data centers, enhancing its infrastructure [1] Group 2: Company Performance - Zhongke Shuguang has launched the Shuguang AI Supercluster system, and its subsidiary Haiguang Information is showing promising business prospects [1] - The company is building a comprehensive competitive advantage in the domestic computing power sector through national project orders, regional penetration, and the integration of self-developed technologies and business scenarios [1] Group 3: Market Outlook - With the continuous growth of national intelligent computing projects and the increasing demand for cloud service providers (CSP), along with the ongoing push for domestic chip substitution, the company is expected to achieve rapid growth [1] - The report maintains a "buy" rating for the company, reflecting confidence in its future performance [1]

Group 1: Company Insights - The report highlights the emergence of China's "Stargate" project, which aims to solidify the core position of domestic computing power, with the successful integration of four national-level computing platforms and six data centers into the Yangtze River Delta hub in Wuhu [4][5] - The computing power public service platform in Wuhu has connected to 34 data centers, aggregating nearly 640P of general computing power, 26,000P of intelligent computing power, 33.3P of supercomputing power, and 2,070 qubits of quantum computing power [4][5] - The report indicates that the share of intelligent computing power in China's total computing power is expected to rise from 3% in 2016 to 35% by 2025, with over 250 intelligent computing centers already established or under construction [5][6] Group 2: Industry Trends - The report discusses the strategic restructuring of coal assets between Pingmei Group and Henan Energy Group, which is seen as a timely move for regional coal asset reorganization [14] - The coal industry is experiencing a slight decrease in port coal prices, while pithead coal prices continue to rise, indicating a mixed supply-demand scenario [15] - The report notes that the petrochemical industry is expected to see an average annual growth of over 5% in added value from 2025 to 2026, driven by a new growth plan released by seven government departments [19][20]

Investment Rating - The report maintains a "Buy" rating for the company [1] Core Insights - The "Star Gate" project in China is emerging, indicating a gradual realization of a closed-loop industry for domestic computing power. The Yangtze River Delta hub in Wuhu has successfully integrated four national-level computing platforms and six data centers, including the Jiangsu Zhixin (Zhongke Shuguang Intelligent Computing Center) [9][10] - The domestic intelligent computing power is steadily increasing, projected to rise from 3% in 2016 to 35% by 2025. Over 250 intelligent computing centers are currently built or under construction in China [10] - The demand for computing power from internet clients is continuously rising, with an accelerated adoption of domestic AI chips. Major companies like Alibaba and Tencent are significantly increasing their investments in AI infrastructure [11][13] Financial Performance and Projections - The company is expected to achieve revenues of 156.63 billion, 218.04 billion, and 285.72 billion yuan for the years 2025, 2026, and 2027 respectively, with net profits of 28.43 billion, 40.98 billion, and 58.02 billion yuan for the same years [14][15] - The earnings per share (EPS) are projected to be 1.94, 2.80, and 3.97 yuan for 2025, 2026, and 2027 respectively, with corresponding price-to-earnings (P/E) ratios of 61.98, 43.01, and 30.37 [14][15] Market Position and Competitive Advantage - The company is building a comprehensive competitive advantage in the domestic computing power sector through national-level project orders, regional penetration, and the implementation of proprietary technologies [14] - The launch of the Shuguang AI supercluster system marks a significant advancement, capable of supporting extensive AI computing needs with high efficiency [13][14]

Core Insights - The report from Guotai Haitong highlights the opening of the CPU interconnect bus by Haiguang Information, which is expected to address the current fragmentation in the domestic computing chip ecosystem and accelerate the unification of the intelligent computing ecosystem [1][2]. Group 1: Haiguang Information - Haiguang Information has opened its CPU interconnect bus to full-stack industry partners, including direct connection IP, communication protocols, and customized instruction sets [2]. - This initiative aims to resolve the lack of uniformity in the technology routes of domestic computing chips and the fragmentation of the system ecosystem, enhancing the scheduling capabilities between CPUs and accelerator cards [2]. Group 2: Shuguang AI Supercluster System - Shuguang has launched the first AI computing open architecture-based supernode, the Shuguang AI Supercluster System, which features high performance with the capability to house 96 GPU cards in a single cabinet, achieving a computing scale of hundreds of PFlops [3]. - The system supports the expansion of super-large clusters with a million cards and demonstrates high computational efficiency, achieving 2.3 times the performance of mainstream models for large model training and inference [3]. - The supercluster system has passed over 30 days of reliability testing and can automatically analyze and isolate faults in millions of components within seconds [3]. - The open architecture allows for hardware compatibility with multiple brands of AI acceleration chips and software compatibility with mainstream AI computing ecosystems, promoting technology sharing and collaboration [3]. Group 3: Investment Recommendations - Recommended stocks include Cambrian-U (688256.SH), Haiguang Information (688041.SH), SMIC (688981.SH), Zhaoyi Innovation (603986.SH), and Shengke Communication-U (688702.SH) [1]. - Related stocks mentioned are Chipone (688521.SH) [1].

Core Viewpoint - The article discusses the launch of China's first AI computing open architecture by Zhongke Shuguang, aimed at addressing the increasing demand for AI computing power and fostering collaboration among industry players to overcome existing challenges in the AI computing sector [1][6]. Group 1: AI Computing Open Architecture - Zhongke Shuguang, in collaboration with over 20 industry chain enterprises, has introduced an AI computing open architecture to promote open collaboration and innovation in the AI computing field [1][6]. - The AI computing open architecture focuses on large-scale intelligent computing scenarios, utilizing GPU-centric designs to enhance system efficiency and collaboration across the AI industry chain [1][3]. Group 2: AI Super Cluster System - The newly launched Shuguang AI Super Cluster System supports multiple brands of accelerator cards and is designed to provide a robust computing foundation for trillion-parameter model training and various AI applications [3][5]. - The system features capabilities such as supporting 96 accelerator cards per cabinet, achieving over 100 petaflops of AI computing power, and allowing for large-scale expansion up to one million cards [5]. Group 3: Technical Innovations and Collaborations - The Shuguang AI Super Cluster has achieved significant performance improvements, including a 2.3 times increase in model training efficiency and a 55% increase in GPU computing efficiency compared to industry standards [5]. - Zhongke Shuguang has announced the opening of key technical capabilities, including AI storage optimization and liquid cooling infrastructure design, to reduce development barriers for small and medium enterprises [5][6].

Core Viewpoint - The launch of China's first AI computing open architecture and AI supercluster system by Zhongke Shuguang and over 20 industry chain enterprises aims to enhance collaborative innovation in AI computing, addressing computing power bottlenecks and promoting accessible computing power [1] Group 1: AI Computing Open Architecture - The AI computing open architecture is designed for large-scale intelligent computing scenarios, focusing on efficient tightly-coupled system design centered around GPUs [1] - The initiative aims to connect AI industry chain enterprises, transitioning from single-point breakthroughs in computing, storage, networking, power, cooling, management, and software to cluster innovation [1] - Zhongke Shuguang has built over 20 large-scale computing clusters in the past decade, deploying more than 500,000 heterogeneous accelerator cards [1] Group 2: AI Supercluster System - The newly launched Shuguang AI supercluster system features four main characteristics: ultra-high performance, ultra-high efficiency, ultra-high reliability, and comprehensive openness [1] - This system is designed to provide a robust computing foundation for trillion-parameter large model training and inference, industry large model fine-tuning, high-throughput inference, and multi-modal large model development [1]