Morgan Stanley: China's wafer foundry capacity and chip supply will meet core sovereignty needs around 2028.

Morgan Stanley released a research report stating that China has made substantial progress in alleviating equipment and wafer foundry bottlenecks in the past 12 months. With expected policy support, China's wafer foundry capacity and chip supply volume will meet core sovereignty needs around 2028. The report points out that policy support can accelerate early development, but long-term value depends on commercial competitiveness. Chinese AI GPU suppliers need to demonstrate convincing economic benefits in order to sustain continued growth after 2028. The analysis by the bank shows that with lower prices, cheaper electricity costs, and increasingly sophisticated infrastructure, China's AI data centers have competitive total cost of ownership (TCO). In the field of inference workloads, the cost per token is more critical than high performance, further enhancing China's competitive advantage in solutions. The bank believes that China's localization strategy of expanding chip, wafer factory, and equipment scales to compensate for process technology disadvantages is continuing to be effective. In an optimistic scenario, domestic GPUs will expand to training workloads and may be adopted overseas; in a pessimistic scenario, the fading of differentiation advantage leads to commoditization and industrial integration. The bank also predicts that the size of the Chinese AI chip market will reach $67 billion by 2030, implying a compound annual growth rate of 23% from 2024 to 2030. The bank expects China's self-sufficiency in AI chips to reach 76% by 2030. The bank continues to be optimistic about the Chinese AI semiconductor supply chain, including SMIC (00981), NAURA Technology Group (002371.SZ), and ASMPT (00522), as well as AI chip investments that help strengthen the strategic layout of Chinese internet platforms.