2025年算力调度平台行业 优化计算资源，支撑AI应用 ◼ 研究背景 随着人工智能技术的迅猛发展，全球 范围内对算力的需求呈现指数级增长， 因此需要算力调度来实现跨地域、跨 平台的算力资源整合与优化调度。 ◼ 研究目标 概览标签：算力、算力调度 Computing Power Scheduling Platform Industry 計算力スケジューリングプラットフォーム業界 报告提供的任何内容（包括但不限于数据、文本、图表、图像等）均系头豹研究院独有的高度机密性文件（在报告中另 行标明出处者除外）。 ，任何人不得以任何方式擅自复制、再造、传播、出版、引用、改 编、汇编本报告内容，若有违反上述约定的行为发生，头豹研究院保留采取法律措施、追究相关人员责任的权利。头豹 研究院开展的所有商业活动均使用"头豹研究院"或"头豹"的商号、商标，头豹研究院无任何前述名称之外的其他分支机构 ，也未授权或聘用其他任何第三方代表头豹研究院开展商业活动。 头豹研究院 1 01 异构算力调度面临的挑战： ◆ 异构算力调度面临多重核心挑战：资源异 构性与软件环境碎片化显著增加调度复杂 性；跨架构任务迁移成本高导致效率低下； 缺乏统一调 ...

Core Viewpoint - Linping District in Hangzhou is transforming into a hub for technological innovation and advanced manufacturing, aiming to become a model for common prosperity and a highland for digital transformation in the Yangtze River Delta region [1][14]. Group 1: Technological Innovation - Linping has seen significant growth in technological innovation, with 24 enterprises recognized as provincial-level R&D institutions, including 5 key enterprise research institutes, ranking first in the province [1]. - The district's R&D investment accounted for 3.79% of its GDP, with 1,498 new invention patents filed, marking a 60.6% increase [15]. - Linping has added 3 new "little giant" enterprises and 12 national-level specialized "little giant" enterprises [1][15]. Group 2: Industry Development - The China (Hangzhou) Computing Power Town, the first of its kind in the country, has attracted over 2,400 companies across various sectors, including AI algorithms, chip design, and software services [4]. - The Computing Power Town hosts 50 national high-tech enterprises and 127 provincial technology-based SMEs, benefiting from the surge in artificial intelligence [4]. - The establishment of the Hangzhou Data Element Operation Center in the Computing Power Town aims to create a collaborative innovation center for data elements [11]. Group 3: Policy Support and Ecosystem - Linping offers a comprehensive policy support system for enterprises, including rent reductions, R&D subsidies, and talent introduction, facilitating growth from startup to maturity [8]. - The district has implemented a one-stop service center that provides 100% coverage for enterprise needs, including talent recruitment and financial connections [11]. - Companies like Mude Micro-Nano and Zhizhen Technology have benefited from local policies, enabling them to innovate and reduce operational costs significantly [8][10]. Group 4: Strategic Location - Linping serves as a natural "bridgehead" for Hangzhou's integration with Shanghai and the Yangtze River Delta, supported by a well-developed transportation network [12]. - The district is strategically located at the intersection of major innovation corridors, enhancing its role in regional economic development [12][14]. Group 5: Future Outlook - Linping is expected to continue attracting innovative enterprises, with over 100 companies in the artificial intelligence sector already established [14]. - The district's focus on digital economy and intelligent manufacturing positions it as a key player in the future industrial landscape of Hangzhou and the Yangtze River Delta [16].

Investment Rating - The report indicates a positive investment outlook for the computing power scheduling platform industry, driven by the rapid growth in AI model applications and the increasing demand for high-performance computing resources [10][14]. Core Insights - The demand for intelligent computing power in China is experiencing unprecedented growth, particularly in large model applications, which account for nearly 60% of the demand, highlighting the significant potential of the future computing power market [14][30]. - The profitability of computing power scheduling centers heavily relies on government subsidies, which are designed to ensure local resource utilization and risk control. Successful engagement with government decision-making is crucial for companies to secure these subsidies [22][30]. - The core value of computing power scheduling platforms lies in their ability to efficiently integrate and schedule heterogeneous computing resources, significantly improving resource utilization and reducing costs for users [17][20]. Summary by Sections Computing Power Scheduling Platform's Scaling Path - The rapid growth of intelligent computing demand in China is driven by AI large models, leading to accelerated development of large-scale, high-performance computing centers [11][10]. - The profitability of computing power scheduling centers is highly dependent on government subsidies, which aim to ensure local resource utilization and risk management [22][30]. - Platforms must possess efficient integration and scheduling capabilities for heterogeneous computing resources to achieve low-cost, scalable, and marketable monetization [31][39]. Value of Computing Power Scheduling Platforms - The platforms enhance resource utilization, lower user costs, and simplify management processes, providing efficient and convenient computing power services [17][20]. - The core technologies required include resource virtualization, fine-grained slicing, real-time monitoring, and tidal scheduling, which enable low-cost and efficient utilization of resources [31][38]. Government Subsidies as Core Source - Government subsidies are essential for the profitability of computing power scheduling centers, with a structured mechanism to ensure local resource utilization and risk control [22][30]. - Companies must strategically engage early in government decision-making to influence standards and secure contracts [23][30]. Technical Features of Quality Computing Power Scheduling Platforms - Platforms need to have capabilities for fine-grained resource slicing, heterogeneous compatibility, cross-regional scheduling, real-time monitoring, and dynamic scheduling to achieve efficient resource reuse and low-cost monetization [31][38]. Core of Scaling Computing Power Centers - The core of monetizing the value of computing power platforms lies in a large and diverse customer base, which determines profitability speed and pricing potential [39][42]. - A diverse customer base allows for tiered pricing strategies to maximize revenue, while partnerships can help focus on high-margin computing power sales [42][39].

Investment Rating - The report does not explicitly state an investment rating for the industry Core Insights - The demand for computing power in China is rapidly increasing, driven by AI large models, leading to accelerated development of large-scale, high-performance computing centers [10][11] - The profitability of computing power scheduling centers heavily relies on government subsidies, which are designed to ensure local resource utilization and risk control [10][22] - The core value of computing power scheduling platforms lies in their ability to efficiently integrate and schedule heterogeneous computing resources, significantly improving resource utilization and reducing costs for users [17][20] Summary by Sections 1. Rapid Growth of Computing Power Demand - The demand for intelligent computing power is experiencing unprecedented growth, particularly in the application of large models, which account for nearly 60% of the demand [14][16] - By 2028, the overall demand for intelligent computing power in China is expected to exceed 20,000 MW, with a compound annual growth rate (CAGR) of 68.97% [16] 2. Value of Computing Power Scheduling Platforms - These platforms enhance resource utilization, lower user costs, and simplify management processes, providing efficient and economical computing power services [17][20] - The platforms must possess advanced technical capabilities, including resource virtualization, fine-grained slicing, real-time monitoring, and tidal scheduling to achieve low-cost and efficient utilization of resources [31][38] 3. Government Subsidies as Core Source - The success of computing power scheduling centers is closely tied to government subsidies, which aim to ensure local resource utilization and mitigate risks [22][30] - Companies must strategically engage early in government decision-making processes to influence outcomes and align their proposals with government economic and social goals [23][24] 4. Technical Features of Quality Scheduling Platforms - Essential technical capabilities include fine-grained resource slicing, heterogeneous compatibility, cross-regional scheduling, real-time monitoring, and dynamic adjustment [31][38] - These features enable efficient reuse of computing resources and low-cost monetization [38] 5. Core of Scale for Computing Power Centers - The scale of value realization for computing power platforms is fundamentally based on a large and diverse customer base, which influences profitability and pricing potential [39][42] - High-value, stable demand from large customers is a primary target for platforms, facilitating efficient service delivery [42]

Core Viewpoint - The emergence of the fifth-generation industrial park, also known as the empowering industrial park, signifies a profound paradigm shift in global industrial parks, moving beyond merely providing physical space to offering comprehensive, customized empowering services through advanced technologies and innovative concepts [1][2]. Group 1: Evolution of Industrial Parks - The rise of the fifth-generation industrial park is a natural outcome of the intertwining of multiple socio-economic and technological transformations in today's world [2]. - Traditional industrial parks are increasingly unable to meet the diverse and high-quality demands of modern industry due to their limitations in functionality and management [2]. - Breakthroughs in new information technologies, such as artificial intelligence, IoT, and big data, have fundamentally changed the operational management models of industrial parks [2][3]. Group 2: Technological Integration - IoT technology enables real-time remote monitoring of equipment within the park, significantly improving operational efficiency and reliability [3]. - Big data analytics allows for in-depth analysis of operational and market data, providing personalized and efficient services to enterprises [3]. - The deep integration of digital technology with the real economy has led to the emergence of new industrial forms, such as smart manufacturing and blockchain applications [3]. Group 3: Focus on Emerging Industries - The fifth-generation industrial park emphasizes the development of high-tech industries like semiconductor equipment and biomedicine, creating a robust ecosystem through the integration of innovation, industry, and finance [4]. - Traditional industries are compelled to transition towards high-end manufacturing and strategic emerging industries due to long-standing challenges of high resource consumption and low added value [4]. Group 4: Urban Integration - The acceleration of urbanization has made "integration of production and city" a mainstream trend, with the fifth-generation industrial park addressing issues of employee living conditions by enhancing supporting facilities [5]. - These parks aim to create a comprehensive environment that integrates work, life, and ecology, significantly improving the quality of life and attractiveness for talent [5]. Group 5: Comprehensive Innovation - Compared to previous generations, the fifth-generation industrial park has achieved a comprehensive innovation in concepts, models, and functions, transitioning from "space providers" to "ecosystem enablers" [6]. - The parks offer diversified services such as equity investment and technology sharing, supporting the full lifecycle growth of enterprises [6]. - The integration of six elements (industry, education, research, finance, users, and government) fosters an open and collaborative innovation network [6]. Group 6: International Collaboration - The fifth-generation industrial park breaks regional limitations and promotes international resource integration, exemplified by Sino-German cooperation parks [7]. - These parks leverage international technology and local market resources to enhance production efficiency and innovation [7]. Group 7: Future Prospects - The fifth-generation industrial park is positioned as a key carrier of new productive forces, emphasizing the importance of service-oriented production capabilities [8]. - The parks are expected to lead a new round of industrial revolution, injecting sustainable momentum into regional economic development [11]. - Balancing innovation with regulation and efficiency with equity will be crucial for the success of these parks [11].

Group 1: AI Development in Shanghai - The Shanghai "Mosu Space" innovation ecosystem community aims to incubate the AI industry, focusing on a large model industry ecosystem [1] - The community includes five public service platforms: open data, testing and evaluation, computing power scheduling, financing services, and comprehensive services [1] - As of February 2023, the computing power scheduling platform has established partnerships with nine computing power partners and planned 30 EFLOPS of computing resources [1] Group 2: Investment Trends in AI - Tencent has restructured its mixed Yuan model R&D system, focusing on computing power, algorithms, and data [2] - The global AI IT investment and generative AI market are expected to grow rapidly, with China's investment in the AI market projected to exceed $100 billion by 2028 [2] - The global data center market is forecasted to grow from $242.72 billion in 2024 to $584.86 billion by 2032, with a CAGR of 11.6% [2] Group 3: Company Developments - Yunsai Zhili is a leading computing power company in Shanghai, owning multiple data centers [3] - Tuolisi provides training dataset sales services to major model manufacturers such as Xiaomi and Ant Group [3] - Xingchen Technology has released SoC chips suitable for AI glasses, featuring advanced ISP visual effects [4]

中新网福州4月29日电 (记者 刘育英)4月28日，第八届数字中国建设峰会期间，政、企、研多方代表， 聚焦人工智能、大模型、算力、数据要素等前沿技术，在分论坛中掀起了一场关于首都智能化跃迁 的"头脑风暴"。 北京市经济和信息化局党组成员、副局长苏国斌指出，当前北京正加快数字经济与人工智能深度融合， 夯实数字基础设施和智能算力体系，未来将聚焦前沿技术，深化与北京电信等企业的合作，推动首都数 字经济生态优化与智能化升级。 中国信息通信研究院云计算与大数据研究所副所长栗蔚指出，AI云时代下，大模型产业正在经历从"百 模大战"的预训练竞赛向云端推理服务效能优化的战略迁移，工程化作为大模型技术实现产业化的必要 手段，在训练和推理阶段分别侧重软件工程和应用工程，需实现"硬件、软件、模型、应用+专业服 务"五位一体协同，打通全链条价值体系。 会上，北京电信与北京邮电大学共建的"行业数据智能标注联合实验室"正式揭牌。论坛期间，北京电信 展示了在AI、算力、数据要素等领域的相关创新和探索。(完) 目前，中国电信在北京正加速布局算力基础设施，体系化搭建算力调度平台、大模型训练平台、智能体 开发平台和标准化的数据标注体系。北京电信 ...