Workflow
液态金属TIM材料
icon
Search documents
A股三大指数开盘涨跌不一,国产芯片概念集体高开
Market Overview - The Shanghai Composite Index rose by 0.03%, while the Shenzhen Component Index fell by 0.15%, and the ChiNext Index decreased by 0.31% [1] - Domestic chip concepts opened higher, with Cambrian Technology rising nearly 4%, while sectors like financial technology and rare earths experienced slight declines [1] Securities Insights - Huaxi Securities noted a market sentiment recovery after a brief fluctuation, with a focus on tracking fund movements, indicating a positive participation willingness from investors [2] - Huatai Securities highlighted a significant increase in new account openings and a shift towards wealth management among brokerages, suggesting a recovery opportunity for firms leading in this transformation [2] - The A-share brokerage index PBLF is at 1.67x, while the H-share brokerage index PBLF is at 1.00x, indicating potential for value reassessment in the sector [2] Commercial Aerospace - Galaxy Securities projected sustained high growth in the commercial aerospace industry, with advancements in reusable rockets and low-cost satellites [3] - The demand for low-altitude economy and space tourism is expected to expand, supported by recent policy measures in Guangdong province [3] Consumer and Commodity Trends - CITIC Construction pointed out the approaching consumption peak season, with rising demand leading to increased prices for rare earths and lithium [4] - The price of lithium carbonate surged due to supply disruptions, while the prices of praseodymium-neodymium and cobalt are also on the rise due to supply constraints [4] AI Server Cooling Technology - Guojin Securities reported that the liquid cooling technology for AI servers is gaining traction, with cold plate liquid cooling being favored for its cost-effectiveness and compatibility with various cooling liquids [5] - The use of high thermal conductivity materials, such as liquid metal TIM, is expected to enhance the cooling performance of cold plate systems, making it a preferred choice for high-end chips [5][6]
AI服务器散热进入液冷时代,导热散热材料需求突显 | 投研报告
英伟达GB300NVL72服务器算力进一步提升,液冷散热需求迅速提升。英伟达今年即将 发布其专为AI推理时代而打造的AI服务器NVIDIA DGX GB300NVL72,AI算力进一步提 升,推论效能较DGX GB200提升了1.5倍。超高的芯片运算能力使得热设计功耗(TDP)进 一步提升,传统风冷散热技术热导率较低,难以满足较高TDP的散热需求,GB300服务器采 用了完全液冷式机架设计,引领了AI服务器液冷的新潮流。AI服务器液冷散热需求或将伴 随AI服务器出货量的提升而迅速提升。 现有液冷方案众多,主要包括冷板式、浸没式及喷淋式,高效液冷方案对氟化液材料依 赖较高。根据冷却液是否与电子器件直接接触,液冷可分为间接液冷与直接液冷;根据冷却 液工作过程中是否会发生液相变为气相的相变过程,可分为单相液冷与两相液冷。其中间接 液冷当前主流的形式是单相/两相冷板式液冷,直接液冷主要是单相/两相浸没式液冷与单相/ 两相喷淋式液冷。在这些液冷方案中,冷板式技术较为成熟,应用较为广泛,但是冷却效果 上限较低;浸没式和喷淋式液冷的冷却效果优异,但是经济成本较高,同时对使用材料要求 较高:由于冷却液与电子器件直接接触,因此 ...