【英文】摩根士丹利报告：全球半导体竞赛：中国将如何布局（35页）

英文研究报告 2020年04月20日 06:38 管理员

In the past, CPUs have been the biggest factor affecting the performance and speed of datacenters. However, if there are bandwidth limitations, even the most powerful CPU cannot deal with increasing M data traffic. To achieve faster networking speeds, then, a more diverse range of companies needs to work together, including semiconductor manufacturers, hardware players, and telecom operators. Semiconductor companies, for example, need to come out with more switch ICs to deal with internet traffic. However, CPUs are approaching certain limits as their computing power cannot fully match the requirements of big data analysis. As a result, GPUs have emerged as an alternative to be used in parallel computing for AI training. Networking speeds continue to rise but data transmission within the system has reached a limit. This implies that interfaces in the system also need to be upgraded to speed up the whole datacenter. Both I/O and communications interfaces need to be upgraded.

Localizing datacenter semiconductors is a tough task because of the high technology entry barriers. For example, Hisilicon has introduced its own CPU for Huawei servers, and government also wants to fully support it and increase the penetration. But given the performance gap with Intel, the pace of increasing penetration will be gradual. The power management IC sector is also dominated by foreign companies like Infineon and TI. However, high-speed interfaces are now presenting great opportunities for Greater China semi companies. Right now most high-speed interface chips are still dominated by US companies. However, we believe the performance gap between Greater China semis and their US peers is small. Some companies in Asia have even better design capabilities than US companies. Huawei, Inspur, and Dawning account for 20-30% of global server demand and offer significant opportunities for Greater China semi companies.