英伟达的好日子快到头了。
据报道，中芯国际将在上海建一座全新的5纳米芯片工厂。专门为华为代工麒麟芯片和AI 芯片。
AI 芯片是芯片里边最简单的一种。就是简单的并行重复累积。华为即将推出的AI 芯片将把英伟达的AI 芯片捻成碎片。
中国本土的EUV 光刻机应该已经在路上了。可能已经完成了客户验证。即将进入大规模的量产。否则以DUV多重曝光技术， 中心国际不可能专门建全新的5纳米芯片制造厂。
https://www.tomshardware.com/tech-industry/semiconductors/chinas-smic-foundry-on-track-to-produce-5nm-smartphone-chips-for-huawei-this-year-report
Semiconductor Manufacturing International Corp. (SMIC), China's national chip champion, is on track to mass produce processors on its 5nm-class process technology later this year, according to a Financial Times report that cites two sources with knowledge of the matter. Huawei's HiSilicon chip development arm will be SMIC's alpha customer for its 5nm node and is expected to make AI and smartphone processors.
SMIC has reportedly established new semiconductor production lines in Shanghai to mass-produce chips designed by technology giant Huawei on its 5nm fabrication technology. As expected, the company's 5nm-class manufacturing process will not rely on extreme ultraviolet (EUV) lithography as the company cannot get appropriate tools from ASML. Instead, SMIC will use 'stockpiled' deep ultraviolet (DUV) lithography tools, according to the Financial Times.
It remains to be seen what power, performance, and area advantages SMIC's 5nm-class process technology will offer. Meanwhile, some experts believe that this production node will enable Huawei to substantially upgrade its smartphone performance this year.
SMIC is expected to produce at least two chips for Huawei's HiSilicon: a highly integrated Kirin system-on-chips for smartphones and the rumored Ascend 920 processor for artificial intelligence (AI) applications.
But these chips will come at a cost. Three individuals connected to Chinese fabless chip designers reported that SMIC's pricing for its 5nm and 7nm products is 40% to 50% higher than what TSMC charges for similar technology nodes. Furthermore, SMIC's 7nm yield is claimed to be less 'than [1评]
快科技2月7日消息，今天，美国国家工程院（National Academy of Engineering, NAE）公布了新增院士名单，英伟达创始人兼CEO黄仁勋赫然在列。
NAE表示，黄仁勋入选美国工程院院士的理由是：“用高性能图形处理单元，推动了人工智能革命。”
美国工程院是工程学界的顶尖学术机构，是美国四大国家学术机构之一（其他三家分别为美国国家科学院 NAS、美国国家医学院 NAM 及美国国家科学基金会 NSF）。
当选美国国家工程院院士是工程领域专家的最高专业荣誉之一，主要是为了表彰入选者在工程研究、实践、教育等方面做出的重大贡献；
在新兴领域技术做出的开拓性工作和传统工程领域中的重大贡献，以及在开发/实施工程教育方面的创新贡献。
黄仁勋1963年2月17日出生于中国台湾省台南市，1983年毕业于俄勒冈州立大学，1990年取得斯坦福大学电子工程硕士学位，1993年黄仁勋创立英伟达公司。
2006年，推出并行计算平台和编程模型“CUDA”，通过GPU实现多个领域的高效计算，为人工智能发展提供了支撑。
作为为人工智能领域“无可争议的领导者”，最新的数据显示，英伟达在全球人工智能芯片市场的份额预计最高可能已经达到了90%，创下新高纪录。