华为杨超斌：液冷技术将有效破解数据中心散热难题

Group 1 - The rapid development of artificial intelligence is driving an increase in chip computing power, which in turn presents significant heat dissipation challenges [1] - Current chip heat density has reached high levels and is expected to continue rising, making efficient heat dissipation crucial for stable chip operation [1] - Traditional cooling methods, such as air cooling, are only effective for small-scale scenarios, while liquid cooling technology is essential for large-scale data centers with thousands to tens of thousands of processors [1] Group 2 - Liquid cooling offers significant advantages over air cooling in terms of energy profit margins and data center utilization rates, supporting efficient AI computing and data center operations [1] - There are notable shortcomings in the current state of industry construction, particularly in the standardization and productization of AI-related data centers and equipment [1] - The construction and standardization of AI data centers are lagging, leading to delays in equipment deployment and resulting in wasted computing resources [1] Group 3 - The conference aims to gather industry forces to discuss the establishment of a standard system for China's AIDC business data centers and the entire industry chain [2] - The Global Computing Alliance is leading the development of future AI data center standards, which will facilitate efficient innovation across the industry chain and support the sustainable development of China's AI industry [2]