Goldman Sachs: AI has already reshaped the landscape of China's stock market; bullish on China's AI export leaders and selected domestic AI leading enterprises.

Goldman Sachs Chief China Equity Strategist Kinger Lau and his team recently published a China strategy update. The research team reiterated that AI has already reshaped the landscape of China's stock market. Since the beginning of 2025, China's AI-related listed companies have created nearly US$5 trillion in market capitalization through share price gains and IPOs. Currently, AI stocks account for 42% of the total market capitalization of A-shares and Hong Kong stocks, and 57% of total turnover. Goldman Sachs' main views are as follows: China's two AI investment themes The narrative of a global AI duopoly is increasingly drawing investor attentionfor reasons including geopolitics, geostrategy, national security, and economic growth, two dominant and self-sustaining AI ecosystems are gradually taking shape globally: one led by the United States and its major trading partners, and the other led by China. Goldman Sachs Research previously launched a global AI equity investment universe covering more than 6,000 listed companies. Within this investment universe, Goldman Sachs' China macro strategy analysis team estimates that Chinese AI companies, weighted by AI business exposure, have a market capitalization of US$7.6 trillion across domestic and overseas markets, or US$10 trillion on a full market capitalization basis, accounting for roughly 10% of global relevant market capitalization and earnings, 14% of capital expenditure, and 24% of revenue. Leveraging strong manufacturing capabilities and cost competitiveness, China has been a key exporter in parts of the global AI supply chain, particularly in rare earth mining and refining, power equipment, PCBs, optical modules, and mature-node semiconductors, holding more than 50% share in each of these sub-markets globally. Meanwhile, with the development and evolution of domestic AI technology, China has gradually built an internally circulating and increasingly self-sufficient AI ecosystem. This system covers some of the "chokepoint" segments of the value chain historically dominated by overseas players, particularly advanced semiconductors and memory chips, semiconductor packaging technology, open-weight/open-source large language models, token generation and consumption, and physical AI application scenarios. The performance level of China's large language models is already well known, but China has also made breakthroughs in other foundational technology areas, including advanced semiconductor packaging and computing, bioengineering, and aerospace. Importantly, the technology gap between China and overseas is narrowing. Although, based on public information, China still has a generational gap in cutting-edge semiconductor technology, in some power, infrastructure, model, and (physical) AI application areas, China may be on par with or even ahead of global competitors. At the same time, the research team also noted that achieving full self-sufficiency in AI technology remains a long journey for China, for reasons including: (1) significant differences between Chinese and US AI-listed companies in market capitalization, profitability, and product pricing, and therefore a gap in financing and investment capacity to support future growth; (2) specific bottlenecks in the local supply chain that may hinder the expansion of the Chinese ecosystem's technology frontier; (3) AI technology inherently has multiplier effectsmultiple components in the supply chain each lagging by one generation could result in total computing output/efficiency lagging by more than one generation. China's two AI investment themes The China AI stocks within the team's research coverage include more than 1,700 listed companies, of which more than 360 generate over 40% of revenue from overseas and are typically viewed by investors as proxies for the overseas AI capital expenditure boom. Thematically, these companies are mainly concentrated in AI infrastructure and power. On the other hand, more than 880 companies primarily target the domestic market, with a combined total market capitalization of US$3.2 trillion. In the domestic market, AI capital expenditure by China's leading cloud service providers/CSPs has also increased substantially, demand and scenarios for AI applications are flourishing, and policy support for advancing AI self-reliance and self-strengthening is relatively strong. Goldman Sachs Research's top-down return analysis shows that Chinese AI exporters have a higher return beta to the overall performance of global AI stocks than their domestic peers, and are more correlated with geopolitics. However, domestic AI companies are more sensitive to China's local AI capital expenditure trends and policy momentum, which may offer international investors diversification benefits. Regarding specific investment opportunities, the research team stated that although geopolitical and trade policy risks still warrant attention, they remain bullish on China's AI export leaders. As a group, Chinese AI exporters are expected to further expand global market share by leveraging strong cost competitiveness, manufacturing capability and efficiency, R&D intensity, and global vision and execution capability. They are also bullish on selected Chinese domestic AI leading enterprises. As the local AI ecosystem continues to develop, these companies have scarcity value and can provide diversification benefits. Meanwhile, in the research report, the analysts once again emphasized a diversification strategynamely, that AI is far from a homogeneous investment theme. In addition to distinguishing AI companies by demand and the geographic source of revenue, they reiterated several investment ideas centered on AI and spanning other dimensional elements of the AI investment landscape, including at the market level A-shares vs. Hong Kong stocks, the power and nuclear energy supply chain across the five key AI layers, STAR 50 and ChiNext vs. Hang Seng Tech Index in the context of hard tech versus soft tech, and long/short strategies around AI disruption risk.