电子行业智能驾驶芯片：概览-220831（36页）.pdf

1、Overseas Observation Series:Reference to the competitive landscape of intelligent driving chips from Tesla,Nvidia and MobileyeShirui Ouyang(852)3892 .hk31 August 2022Soochow Securities International Brokerage Limited would like to acknowledge the contribution and support provided by Soochow Research

2、 Institute,and in particular its employees Liangwei Zhang(张良卫)and Ruizhe Liu(刘睿哲).Equity Research Industry Research SemiconductorIntelligent driving chips(also known as autonomous driving chips,ADAS chips,etc.)are mainly computing units that enable vehicles to achieve autonomous driving and are a ty

3、pe of AI chips.Starting from Computer Vision(CV),ADAS chips have gradually evolved algorithms for situations that people might encounter when driving.There are two types of algorithms:top-down(Google,Baidu)and bottom-up(Tesla,XPeng).Based on these two algorithms,GPU(NVIDIA)and ASIC chips(Tesla,Qualc

4、omm,Mobileye,Horizon Robotics)are derived.With the accelerated development of automobile intelligence,intelligent driving chips will usher in a stage of rapid expansion,and there is a lot of room for growth.We estimate that the market size of the field will increase from US$1.9 bn to US$5.4 bn from

5、2021 to 2025,with a CAGR of 30%.Since autonomous driving is a new field and there is very little gap between domestic and foreign manufacturers,there are opportunities for domestic manufacturers.We recommend Nvidia(NVDA.O)and put the following three into watchlist:Tesla(TSLA.O),Mobileye&Intel(INTC.O

6、),Horizon Robotics(unlisted),etc.Risks:The governments fade out in policy support leads to a slowdown of autonomous driving market growth;regulations on autonomous driving and artificial intelligence have become stricter,resulting in the delay of commercial projects;Autonomous driving technology can