2月24日，春节后首个工作日，广东再度召开"新春第一会"。今年大会聚焦"制造业与服务业协同发 展"，吹响"十五五"开局之年的冲锋号角。"今年，中国移动广东公司将着力发挥网络基础牢、产业链条 长、行业带动强、属地融入深等优势，注智赋能制造业和生产性服务业深度融合，为广东经济社会高质 量发展作出新贡献。"中国移动通信集团广东有限公司党委书记、董事长、总经理葛松海在会上发言表 示。他强调，要当好"供给者"，打造"网络+服务"支撑体系，推动2.7万家工业企业提质增效，持续打 磨"全光组网+大模型+工业AI应用"解决方案，加速韶关及珠三角算力机房征地建设，助力大湾区打造 全国领先的算力集群。推进AI质检、AI工业机器人等融合应用，支撑企业按需调度算力、精准适配场 景。（南方财经） ...

Core Viewpoint - The company,瑞风新能源, has signed an investment framework agreement with the 宣化经开管委会 to invest approximately RMB 24 billion in the construction of the largest AI computing power cluster in North China, marking a significant step in developing a national integrated computing network [1] Financial Performance - In 2024, the company's revenue is projected to be RMB 336 million, with a net loss of RMB 133 million [1] - For the first half of 2025, the company expects revenue of RMB 183 million and a net loss of RMB 35.53 million [1] Stock Performance - The stock price of 瑞风新能源 has increased by over 7%, currently trading at HKD 0.74, with a trading volume of HKD 19.0154 million [1] Legal Update - On January 22, the company announced that the Hong Kong High Court ordered the withdrawal of a petition and the cancellation of a hearing [1]

Core Viewpoint - The company announced a significant contract with Yixin Intelligent Manufacturing (Hubei) Electronic Information Technology Co., Ltd., which is expected to positively impact its future operating results [1] Group 1: Contract Details - The contract signed by the company's wholly-owned subsidiary, Dalian Xinlianwei, is valued at 132 million yuan (including tax) [1] - This contract accounts for over 50% of the company's most recent audited main business revenue [1] - The scope of the contract includes the provision of an integrated hardware and software solution for computing power cluster construction and services based on the needs of Yixin [1] Group 2: Future Implications - If the contract is effectively executed, it is anticipated to have a positive effect on the company's future business performance [1]

Core Viewpoint - The company is actively engaging in the development and supply of advanced circuit board solutions, particularly focusing on orthogonal backplanes and high-speed connectors to enhance server performance and meet increasing customer demands [2] Group 1: Product Development - The orthogonal backplane is recognized as a solution to optimize server space and heat dissipation, improving information exchange among computing units in large-scale computing clusters [2] - The company is closely monitoring the market for orthogonal backplanes and has made technical arrangements to support this development [2] Group 2: Supply Chain and Partnerships - The company has entered the NVIDIA and Google supply chain through OEM methods, providing specialized circuit board products necessary for the mass supply of high-speed connectors [2] - The company is actively collaborating with clients to meet rapid growth demands and is involved in the research, testing, and certification of next-generation products [2] Group 3: Business Strategy - The company aims to deepen its presence in its core areas while expanding into emerging sectors, actively promoting business development [2]

Core Viewpoint - The company is focusing on orthogonal backplane technology and high-speed connectors to enhance server space optimization and heat dissipation in large-scale computing clusters [1] Group 1: Technology Development - Orthogonal backplane is recognized as a solution for optimizing server space and heat dissipation, improving information exchange among computing units in large-scale clusters [1] - The company is closely monitoring this technology and making technical arrangements to capitalize on its potential [1] Group 2: Supply Chain and Partnerships - The company has entered the NVIDIA and Google supply chain through OEM methods, supplying specialized circuit board products necessary for high-speed connectors [1] - It is actively responding to customer demands for rapid increases in supply while participating in the research, testing, and certification of next-generation products [1] Group 3: Business Strategy - The company aims to deepen its advantages in existing fields while expanding into emerging sectors to promote business development [1]

Core Viewpoint - The company is focusing on orthogonal backplane technology and high-speed connectors to enhance server space optimization and heat dissipation in large-scale computing clusters [1] Group 1: Technology Development - Orthogonal backplane is identified as a key solution for improving information exchange among computing units in large-scale computing clusters [1] - The company is closely monitoring and making technical arrangements for the orthogonal backplane technology [1] Group 2: Supply Chain and Partnerships - The company has entered the supply chains of NVIDIA and Google through OEM methods, providing specialized circuit board products necessary for high-speed connectors [1] - The company is actively responding to the increasing demand from clients and participating in the research, testing, and certification of next-generation products [1] Group 3: Business Strategy - The company aims to deepen its advantages in existing fields while expanding into emerging sectors to promote business development [1]

Core Viewpoint - China Mobile has awarded a procurement project for a super node testing device to Huawei, with a bid of 135 million yuan, indicating a significant investment in advanced computing infrastructure [1] Group 1: Huawei's Innovations - Huawei's "Scale-up super large-scale super node computing platform" was recognized as its top invention, showcasing a new architecture that allows for flexible resource allocation among AI processors [1] - The Ascend 384 super node features 384 Ascend NPUs and 192 Kunpeng CPUs, achieving a total computing power of 300 Pflops, which is 1.7 times that of NVIDIA's NVL72 [2] - The total network bandwidth of the Ascend 384 super node reaches 269 TB/s, surpassing NVIDIA's NVL72 by 107% [2] Group 2: Industry Trends - The demand for computing power in large models is rapidly increasing, while traditional computing architectures face challenges such as low resource utilization and frequent failures [4] - Super nodes are emerging as a new norm in AI infrastructure, with several domestic manufacturers exploring super node systems [4] - Inspiring innovations include the launch of the world's first single-cabinet 640-card super node by Sugon, which enhances computing density by 20 times compared to traditional solutions [4][5] Group 3: Future Developments - Huawei plans to launch the Atlas 950 super node in Q4 2026, featuring 8192 Ascend 950DT chips and achieving FP8 computing power of 8 EFLOPS [12] - The Atlas 950 super node is projected to be significantly more powerful than NVIDIA's upcoming NVL144, with a total computing power 6.7 times greater and memory capacity 15 times larger [14] - Future plans include the Atlas 950 SuperCluster, which will consist of 64 Atlas 950 super nodes, achieving a total FP8 computing power of 524 EFLOPS, surpassing the current largest cluster, xAI Colossus [14]

Core Viewpoint - Ant Group has deployed a domestic computing power cluster with a capacity of 10,000 cards, which is designed to support self-developed and mainstream open-source models, achieving a training task stability of over 98% [1] Group 1 - The computing power cluster is capable of training and inference performance that can compete with international computing power clusters [1] - The cluster is fully applied in the training and inference services for large models in the field of security and risk control [1]

Core Insights - Huawei has launched the world's strongest computing supernodes and clusters composed of Ascend chips, addressing the challenges of insufficient supply, high costs, and the need for ecosystem development in China's computing infrastructure, potentially enabling the country to transition from "catching up" to "leading" in the AI industry [1] Group 1: Challenges in Computing Power - The lack of high-level domestic computing power is one of the most pressing challenges facing China's technology industry, exacerbated by U.S. export controls on key chip technologies [3] - The development of advanced AI chips requires significant upfront investment, including high costs for research and development and the need to build a competitive software ecosystem to attract developers [3][6] Group 2: Systemic Innovation - Huawei aims to overcome the limitations of domestic chip manufacturing by adopting a systemic approach that integrates chip performance, architectural design, and resource collaboration to create a new computing infrastructure [6] - The company leverages its telecommunications technology to enhance communication speed and reduce latency between chips, allowing thousands of chips to operate as a single machine [6] Group 3: Ecosystem Development - Huawei is opening its Ascend AI software platform to build an open and win-win industrial ecosystem, encouraging collaboration among various stakeholders to accelerate the growth of the domestic computing ecosystem [6] - The innovation in Huawei's supernode technology serves as a model for addressing similar challenges in other high-tech fields, emphasizing the need for a comprehensive ecosystem rather than isolated technological advancements [7] Group 4: Long-term Perspective - The newly launched supernode cluster is not expected to solve all problems immediately and will face challenges in application and promotion [8] - The development of AI technology is still in its exploratory phase, likened to the evolution of communication technology, indicating that the competition in AI will be a long-term endeavor [9] Group 5: Strategic Advantages - China's vast market, large pool of engineers, and robust energy infrastructure provide a solid foundation for overcoming technological gaps and external pressures, fostering a strong environment for independent innovation [9]

Core Insights - OpenAI is gradually reducing its dependency on Microsoft through new strategic partnerships with major companies like Nvidia, Oracle, and SoftBank [1][9] Group 1: Strategic Partnerships - OpenAI and Nvidia have announced a strategic partnership, where OpenAI will utilize millions of Nvidia GPUs to deploy at least 10GW of AI computing power [1] - Nvidia plans to invest $100 billion in OpenAI as the computing power clusters are deployed, with the first cluster expected to be operational in the second half of 2026 [1][2] - This partnership marks a significant step in providing the infrastructure needed for the next era of intelligence, as stated by both Nvidia's CEO Jensen Huang and OpenAI's CEO Sam Altman [1] Group 2: Financial Implications - The $100 billion investment from Nvidia is substantial, nearly equivalent to the total funding in the AI sector in the first half of 2025, which was approximately $104.4 billion [2] - Nvidia's net profit for 2024 is projected at $72.9 billion, with cash flow from investment activities at $20.4 billion, allowing for a feasible investment strategy to cover the $100 billion over five years [2] - If Nvidia's investment materializes, it would make Nvidia the largest shareholder of OpenAI, surpassing Microsoft's $13 billion investment by 7.7 times [3] Group 3: Computing Power Scale - The deployment of a 10GW computing power cluster will involve millions of Nvidia's next-generation Rubin series GPUs, with estimates suggesting a total of 5 to 10 million chips could be utilized [6][8] - This scale of investment and infrastructure will position Nvidia as a major supplier for OpenAI, locking in a significant customer for its chips [8] Group 4: Shift in Dependency - OpenAI's new partnerships indicate a shift away from its reliance on Microsoft, which has historically provided the computing power for OpenAI's models [9] - The "Star Gate" initiative, involving SoftBank, Oracle, and OpenAI, aims to provide $500 billion for computing infrastructure, further reducing Microsoft's role in OpenAI's operations [9][10] - A recent agreement between Oracle and OpenAI for a five-year, $300 billion computing power contract is part of this initiative, potentially impacting Microsoft's Azure growth [10]