英伟达-从硬件GPU设计到软件CUDA+ Omniverse开发建立人工智能和元宇宙生态系统-220527（52页）.pdf

1、NVIDIA（NVDA.O） Equity ResearchCompany ResearchSemiconductor 1 / 52 东吴证券（香港）东吴证券（香港） 请务必阅读正文之后的免责声明部分请务必阅读正文之后的免责声明部分 NVIDIA is building an ecosystem in the field of AI and metaverse, through designing GPU hardware and developing CUDA + Omniverse software Outperform（Initiation） Forecast & Valuation 2

2、021A 2022E 2023E 2024E Revenue (USD mn) 26,914 33,393 45,612 63,256 Growth (YoY %) 61% 24% 37% 39% Net profit (USD mn) 9,752 10,209 16,177 24,245 Growth (YoY %) 125% 5% 58% 50% Adjusted EPS (USD) 3.89 4.07 6.44 9.66 P/E (X) 71.27 68.08 42.97 28.67 Investment Thesis NVIDIA is far-sighted and actively

3、 developing software tool chains, trying to build an ecosystem in the field of AI and metaverse where have a vast market space in the future, in order to strengthen industry barriers by the binding of hardware chips. Compared to companies that mainly focus on chip design such as Intel and AMD, we be

4、lieve that NVIDIA has a higher valuation since its no longer just a chip design company but forming an ecosystem in both software and hardware field. NVIDIA attaches great importance to expanding the use of GPU. In 2006, NVIDIA began to invest heavily in the development of CUDA, a software tool chai

5、n, after realizing the advantages of GPU parallel computing. NVIDIA allowed AI researchers to use CUDA to invoke the computing resources of GPU for free, making NVIDIA an important facilitator of deep learning training and inference in AI. Therefore, AI industry personnel have been using NVIDIAs har

6、dware and software suite since the AI industry started to develop. NVIDIA has also applied a software and hardware combination model to new areas such as the Omniverse. We believe that in the field of AI and the metaverse, NVIDIA has the potential to become a supplier of general hardware platforms +