Geopolitical Shock: Research on Algorithmic and Computing Power Optimization and China's Artificial Intelligence Development Path

Xie Jiaxin^*

Information Engineering College, Hangzhou Dianzi University

Abstract：

This paper analyzes how geopolitical shocks reshape the development path of artificial intelligence in China under tightening restrictions on advanced semiconductors and computing infrastructure. It argues that export controls influence artificial intelligence not only by limiting chip access, but also by raising training costs, slowing experimentation, and amplifying supply-chain uncertainty. At the same time, resource-efficient model design represented by DeepSeek shows that sparse activation, low-precision training, memory optimization, and software-hardware adaptation can partly offset hardware shortages. Based on recent literature on semiconductor competition, supply-chain resilience, and efficient large models, this paper builds a unified framework linking geopolitical pressure, compute constraints, algorithmic response, and industrial adjustment. The analysis suggests that China should avoid a single-path strategy. A more sustainable route combines domestic hardware reinforcement, algorithmic compensation for near-term compute scarcity, and application-driven ecosystem coordination. In this sense, the future competitiveness of China's artificial intelligence sector depends not only on the absolute quantity of computing power, but also on the efficiency with which limited computing resources are transformed into stable model capability, deployment feasibility, and industrial resilience.

Key Words：

geopolitical shock; algorithmic optimization; computing power constraint; semiconductor export control; Artificial Intelligence development path; Deepsee