大行看好！中国科技资产仍存在超预期空间

Core Viewpoint - The capital expenditure demand in the AI investment sector is shifting from US suppliers to Chinese suppliers, indicating significant growth potential for Chinese technology assets, particularly in the domestic substitution direction [1][2]. Group 1: Chinese Technology Assets - Chinese technology assets, especially in the domestic substitution sector, are expected to have unexpected growth potential despite short-term market volatility [2]. - The US's entry into a rate-cutting cycle will lead to increased market liquidity, prompting funds to pursue assets with higher potential returns [2]. - Chinese assets are currently underrepresented in global allocations, indicating significant room for increased investment [2]. - The recognition of China's model capabilities by global tech companies, particularly in the open-source field, is a positive sign for the future [2]. Group 2: Hardware Breakthroughs - By 2025, capital expenditure demand in the AI investment sector is expected to gradually shift from overseas suppliers to domestic suppliers [3]. - The current trend among Chinese tech companies is moving from hoarding imported hardware to actively embracing domestic solutions, which is optimistic for the AI industry [3]. - As leading companies begin large-scale procurement of domestic servers equipped with local chips, profits and capital will flow back to local suppliers, creating a virtuous cycle for technological breakthroughs [3]. Group 3: Global AI Market Dynamics - The global model market has transitioned from a "hundred schools of thought" to a commercialization phase, with a focus on vertical companies [4]. - The funding focus in the AI market is expected to shift towards hardware, with anticipation for the emergence of application-level breakthrough products [4]. - The integration phase of the global model market is nearly complete, with only a few institutions remaining in model development [4]. - Vertical industry data will become key to creating differentiated advantages as model capabilities become more homogeneous [4]. Group 4: AI Commercialization - The path to AI commercialization is clearer for B-end applications compared to C-end applications, making implementation easier [6][7]. - In the e-commerce sector, AI can replace traditional models, reducing operational costs significantly [6]. - The logic behind B-end commercialization is clear and reasonable, focusing on cost savings rather than creating entirely new AI revenue streams [7]. - C-end commercialization faces challenges due to unclear directions and intense competition, with user willingness to pay being low in the Chinese market [7].