RTX 3060
Search documents
摩尔线程上市临近“中国版英伟达”之争硝烟渐浓
Jing Ji Guan Cha Bao· 2025-11-22 05:13
经济观察报记者 郑晨烨 11月21日下午两点,摩尔线程创始人、董事长兼总经理张建中带着高管团队,出现在了上证路演中心的直播间里,摩尔线程首次公开发行股票并在科创板上 市网上投资者交流会准时开始了。 就在前一天晚间,这家公司敲定了科创板的发行价,114.28元/股,按此计算,其发行后总市值约为537亿元。 这是2025年以来科创板融资规模最大的IPO项目,从受理到过会,摩尔线程只用了88天。 在上市路演长达三个小时的问答里,投资者的问题淹没了交流平台,大家抛出的问题都很直白:公司什么时候能盈利?"全功能GPU"路线,到底成色几何? 其招股书里的数据很直观,营收增速确实快,2025年上半年收入就达到了7.02亿元;但代价是巨大的研发投入,过去三年半累计亏损了54.86亿元,因此摩尔 线程此次挂牌科创板募集的80亿元,不仅要用来补充紧缺的流动资金,更是支撑它继续进行高强度技术迭代的本钱。 毕竟,顶着"中国版英伟达"的名号,市场给了它高溢价,未来,就需要它用真金白银的业绩,来证明自己配得上这个身价。 光速过会 摩尔线程这次上市,确实赶上了一个好时候。 从6月30日受理到9月26日过会,中间只隔了88天,这个速度,放在 ...
摩尔线程上市临近 “中国版英伟达”之争硝烟渐浓
Jing Ji Guan Cha Wang· 2025-11-21 15:11
经济观察报记者 郑晨烨 11月21日下午两点,摩尔线程创始人、董事长兼总经理张建中带着高管团队,出现在了上证路演中心的直播间里,摩尔线程首次公开发行股票并在科创板上 市网上投资者交流会准时开始了。 就在前一天晚间,这家公司敲定了科创板的发行价,114.28元/股,按此计算,其发行后总市值约为537亿元。 在上市路演长达三个小时的问答里,投资者的问题淹没了交流平台,大家抛出的问题都很直白:公司什么时候能盈利?"全功能GPU"路线,到底成色几何? 光速过会 摩尔线程这次上市,确实赶上了一个好时候。 从6月30日受理到9月26日过会,中间只隔了88天,这个速度,放在过去两年的A股市场里,几乎是不可想象的,这背后当然有"科创板八条"的政策推力,也 有监管层对未盈利硬科技企业重新打开大门的信号。 从招股书上看,摩尔线程最显眼的变化在于营收规模的跃迁。 这是2025年以来科创板融资规模最大的IPO项目,从受理到过会,摩尔线程只用了88天。 2022年,摩尔线程还是个年入4000多万元的"小"公司;到了2024年,营收已经干到了4.38亿元;再到今年上半年,仅仅6个月,它就入账了7.02亿元。 这种成倍的增长,说明摩尔线程 ...
CPU主频,越高越好吗?
半导体行业观察· 2025-08-01 01:12
Core Viewpoint - The article emphasizes that clock speed, once the primary metric for measuring component performance, has become less significant due to advancements in hardware design and software optimization. Modern applications are increasingly designed to leverage multi-threading, making clock speed a less reliable indicator of performance [2][3]. Group 1: Importance of Specifications - Clock speed is a deceptive metric that is often misused in marketing, as it only indicates the number of instructions a CPU can execute per second, not the overall performance [2][3]. - The evolution of multi-core CPUs has diminished the importance of clock speed, with modern applications designed to utilize multiple threads for improved performance [3][4]. - Cache size has emerged as a critical specification, with larger caches significantly impacting performance, especially in complex applications like modern open-world games [4][5]. Group 2: GPU and Memory Considerations - For GPUs, memory capacity (VRAM) has become a more crucial specification than clock speed, as modern games require more VRAM for optimal performance [6]. - The relationship between VRAM and cache size is important; for instance, a GPU with larger VRAM may not outperform one with a smaller VRAM but larger cache [6]. - Memory speed is measured in MT/s, and understanding the actual latency in nanoseconds is essential for evaluating performance, as higher speed can sometimes lead to increased latency [7][8]. Group 3: Performance Testing - Specifications alone cannot determine actual performance, as hardware interacts with other components and software, leading to various performance variables [9]. - Real-world testing is necessary to assess component performance, as different applications prioritize different specifications, such as CPU threads for video editing versus gaming [9].