Goldman Sachs is bullish on China's AI: there is no overall bubble in the industry, and it is worth investing in the power supply chain, hardware infrastructure, and embodied AI.

Goldman Sachs' Chief China Equity Strategist Kinger Lau stated that there is currently no bubble in China's artificial intelligence (AI) industry as a whole. The recent market adjustments have returned related valuations to healthier levels, and he identified three sub-industries with significant investment opportunities. In an interview, Lau mentioned that the overall market capitalization of AI-related stocks in China has not yet fully reflected the potential economic benefits that this technology will bring to broader economic sectors. He acknowledged that there was a phase of localized overheating in the market in June, when the valuations of some AI hardware companies on the STAR Market and ChiNext reached their highest levels in five years. However, the month-long market correction that followed has restored a "reasonable and healthy" balance between stock prices and future earnings growth expectations. Lau pointed out that AI companies in China currently account for approximately 11% of the total market capitalization of global AI-related stocks, yet the proportion of overseas funds allocated to the Chinese market within their AI investment portfolios is only about 1%. He believes that there are three specific sub-industries in the Chinese AI sector with the strongest investment prospects: the power supply chain, hardware infrastructure, and embodied AI. Specifically, the power sector presents long-term structural opportunities, as domestic equipment manufacturers in China are continuously expanding their global market share, further reinforced by the ongoing large-scale construction of computing infrastructure in China. Meanwhile, the hardware infrastructure sector, which includes printed circuit boards (PCBs), optical modules, and data centers, has a high degree of earnings certainty in the next two to three years. Furthermore, the embodied AI fieldcovering industrial intelligence and humanoid Siasun Robot & Automationbenefits from China's robust manufacturing ecosystem and possesses strong global competitiveness. Lau also stated that besides physical infrastructure, China's AI software applications also have enormous growth potential. Among them, AI tokenswhich represent digital measurement units for AI computing servicesare expected to become a significant new driver of China's export growth. He noted that the cost of a single AI token from China's domestic large language models is significantly lower than that of international competitors, and this cost advantage is accelerating the commercialization of AI agents and industry-specific cloud solutions in the Chinese market. He added that this structural cost efficiency advantage could soon redefine China's trade landscape.