电力AI系列访谈-算电协同专家交流
2026-03-13 04:46

Summary of Key Points from the Conference Call Industry Overview - The discussion revolves around the concept of "computing and electricity collaboration" (算电协同), which has become a national strategy in China. The policy mandates that by 2023, new data centers at hub nodes must have over 80% of their electricity sourced from green energy, with a Power Usage Effectiveness (PUE) target of 1.2-1.25 [1][2][3]. Core Concepts and Arguments - Electricity Cost: Electricity costs account for 60%-70% of the operating costs of computing centers. The direct connection to green electricity can reduce electricity prices by 10%-15%, with costs in regions like Xinjiang and Inner Mongolia potentially dropping to 0.3-0.35 yuan per kilowatt-hour [1][14]. - Green Electricity Connection Models: The mainstream model for green electricity connection is the "grid-connected model." The off-grid model is rarely adopted due to the high costs of energy storage and the volatility of renewable energy sources [1][2]. - Liquid Cooling Technology: Liquid cooling technologies (immersion and plate cooling) are identified as key methods to reduce PUE. Long-duration energy storage and grid-structured energy storage are crucial for replacing diesel generation and stabilizing heavy load impacts [1][3]. - Industry Collaboration: The trend is shifting towards a specialized division of labor where electricity providers supply resources and computing firms handle operations. This collaboration is essential for optimizing resource allocation and operational efficiency [1][2][10]. Policy Background and Future Outlook - The "computing and electricity collaboration" concept was driven by three main factors: the infrastructure of computing power, the explosive growth of artificial intelligence, and the green transformation of the energy system. It was officially included in national policy in December 2023, with pilot programs set for 2024-2025 [2][3]. - Future policy focus areas include: - Coordinated planning between computing and electricity sectors. - Development of key technologies for energy storage and collaborative scheduling. - Establishment of unified standards for computing and electricity collaboration. - Mechanism innovation to address challenges in power trading and transmission across different provinces [3][4]. Implementation Strategies - To meet the requirement of over 80% green electricity for new data centers, three main paths are identified: 1. Self-generated green energy for self-use. 2. Long-term power purchase agreements (PPAs) with green energy producers. 3. Purchasing green certificates, though this incurs additional operational costs [4][5]. - The PUE targets for data centers are challenging, requiring advancements in cooling technologies and energy management systems [5][6]. Economic Considerations - The economic attractiveness of green electricity direct connection projects is significant, as they can lead to substantial cost savings and additional revenue from selling surplus electricity back to the grid [9][14]. - Despite the lower electricity costs in western regions, factors such as network transmission costs can offset these savings, influencing the location decisions of data center operators [14][15]. Challenges for Energy Companies - Energy companies face challenges when entering the computing sector, including knowledge barriers, technical integration difficulties, and uncertain investment returns. Collaborations with specialized computing firms are becoming common to mitigate these challenges [10][11]. Technological Innovations - Key technologies and devices in the context of green electricity direct connection and computing collaboration include: - Integration of renewable energy sources with energy storage solutions. - Advanced power transmission networks capable of handling large-scale computing demands. - AI-driven platforms for optimizing energy management and computing task scheduling [17]. Conclusion - The collaboration between computing and electricity sectors is essential for achieving sustainable growth and efficiency in energy consumption. The ongoing developments in policy, technology, and industry practices will shape the future landscape of this collaboration in China.

电力AI系列访谈-算电协同专家交流 - Reportify