【AI+Major Bank Report】Daiwa Morgan: China’s AI Enters a New 2.0 Stage, with Focus on Value Capture; Recommends 18 Stocks

Morgan Stanley states that China’s artificial intelligence (AI) is moving toward a new phase 2.0. China is committed to advancing AI development simultaneously in various frontier fields, robotics, new energy vehicles, and large models, and its “AI+” plan is rapidly integrating newly acquired AI capabilities and AI agents into world-class intelligent products, services, and embodied intelligence.

Morgan points out that although empowering entities and foundational models remain the core investment logic at the current stage, the widespread application of AI presents potential investment opportunities for companies able to benefit from it.

In its risk-return analysis of AI application providers, Morgan highlights prominent performance from Beisen Holdings (09669), Meitu (01357), Stone Technology (Shanghai: 688169), Midea Group (Shenzhen: 000333), and Ecovacs Robotics (Shanghai: 603486). Morgan also notes that MiniMax (00100) and Zhipu (02513) are key foundational model providers for AI in China, while Alibaba (09988) is the best-positioned full-stack AI platform among the companies covered by the firm.

In the power sector, the firm mainly recommends CATL (03750), Yingliu Holdings (Shanghai: 603308), and Siyuan Electric (Shenzhen: 002028). Additionally, Morgan remains optimistic about Cambrian (Shanghai: 688256), TianShu Zhixin (09903), North Huachuang (Shenzhen: 002371), Semiconductor Manufacturing International Corporation (SMIC) (Shanghai: 688012), ACM Research (U.S.: ACMR), SMIC (00981), and Xinxing Electronics (Taiwan: 3037), as they benefit from China’s long-term semiconductor localization trend and AI empowerment.

China’s AI Focuses More on Value Capture, Shifting Attention to Three Aspects

Morgan states, “China’s AI is entering a new phase, no longer just about capability catching up, but more about value capture. The market focus has shifted from training to reasoning, from technology to application, and from potential to actual profitability.”

Rather than merely competing in frontier technologies, Morgan indicates that China is optimizing for speed, cost efficiency, and system-level integration, enabling AI to rapidly diffuse into the real economy. Meanwhile, the firm forecasts that China’s semiconductor self-sufficiency rate will rise from 41% in 2025 to 86% by 2030, supporting more resilient and cost-effective deployment.

Limited Short-term Impact of AI on China’s Growth, but AI Adopters Outperform Index and Market

On the macro level, Morgan expects AI to become a productivity lever in the medium term, estimating that over the next decade, AI could cumulatively boost China’s total factor productivity by about 3 percentage points, partially offsetting the effects of aging populations, and raising the potential GDP level in 2035 by 3.5 percentage points above a no-AI scenario. However, in the short term, the growth impact of AI is relatively moderate, as positive factors from AI-related capital expenditure cycles and initial productivity gains will be offset by labor market shocks during the transition period.

Morgan also notes that AI adoption is expanding, with the penetration rate of enablers or adopters rising from 43% to about 51% over two years. The economic benefits are becoming tangible, with AI adopters’ estimated next twelve months’ earnings per share (NTM EPS) growing 62% over the past two years, significantly outperforming the MSCI China Index’s 10%. Meanwhile, their EBIT margins are expected to expand by 12 to 13 percentage points to 16-17% by 2027, far surpassing the market.