破解AI高能耗瓶颈 中国探索算力电力共生发展新路径

Core Insights - The forum on "AI and Green Low-Carbon Development" highlighted the dual challenge of AI's rapid growth leading to significant computational energy consumption and the need for sustainable development models [1] - Experts emphasized that AI is both a critical tool for refined energy and carbon management and a substantial new load on energy systems [1] - The forum proposed systematic solutions to address the "AI energy paradox," focusing on energy efficiency in both software and hardware [1] Group 1: Technological Solutions - The software aspect aims to optimize algorithm training for more efficient specialized models, thereby saving computational power from the source [1] - The hardware aspect involves adopting advanced technologies like liquid cooling and AI self-optimization systems to significantly reduce data center energy consumption [1] Group 2: Energy Storage and Reliability - Development of long-duration and reliable energy storage technologies is essential to ensure stable operation of AI computing centers with green electricity, enabling participation in grid scheduling [2] Group 3: Market Mechanisms - Utilizing AI to quantitatively assess corporate green performance can guide capital markets in supporting green transitions [3] Group 4: System Integration - The release of the "Energy-Carbon Intelligent Computing Hub" aims to achieve integrated management and global optimization of energy flow, carbon flow, and data flow, marking a critical step from decentralized applications to systematic top-level design [4] - The consensus and proposed "Chinese wisdom" and systematic practical paths from the forum contribute important solutions to the global energy and environmental challenges posed by AI development, demonstrating a commitment to promoting intelligent and sustainable collaborative development [4]