Group 1 - The core viewpoint emphasizes that "large models + large computing power + large data" is a primary route to achieving general intelligence, with the Scaling Law still holding true as foundational models rapidly iterate [1] - Recent releases of foundational models like Gemini 3.0 and DeepSeek 3.2 indicate a fast-paced evolution in language, visual, and multimodal capabilities, with anticipation for DeepSeek V4/R2 [1] - The demand for AI applications is accelerating, as evidenced by Alibaba's Qianwen APP reaching over 30 million monthly active users within 23 days of public testing, and Tencent's AI investments driving growth [1] Group 2 - The "pan-sinicization" movement is penetrating from government and party sectors into industries, aiming to build a self-sufficient ecosystem, with state-owned enterprises required to replace various information systems with domestic alternatives [2] - The demand side is supported by policies promoting "domestic goods for domestic use," with a focus on enhancing the resilience and security of key industrial supply chains [2] - On the supply side, the China Information Security Evaluation Center updated its guidelines to include new categories for CPU product evaluations, reflecting a push for domestic technology development [2] Group 3 - The impact of tariffs is noted, with the proportion of overseas income in the computer sector showing a fluctuating upward trend from 11.20% in 2019 to 14.75% in 2024, indicating a degree of desensitization to external changes [3] - The overall sentiment towards the computer sector remains positive, with expectations for future market performance despite potential risks related to technology development and policy implementation [4]

Investment Rating - The report maintains a rating of "Outperform" for the AI industry [6] Core Insights - The AI sector is experiencing a surge in interest due to recent positive developments, including the launch of Byte's Doubao mobile assistant and the DeepSeek-V3.2 model. The global AI industry is projected to reach a scale of $1.3 trillion by 2032, with a clearer transition from technological breakthroughs to value realization expected in 2026, making it a core investment theme for the year [2][14] - As of November 30, there are 24 ETFs tracking AI indices, with 6 corresponding to the AI index. The largest is the CS Artificial Intelligence Index ETF with a scale of 35.414 billion, followed by the Sci-Tech AI Index ETF at 16.719 billion. The best-performing ETF this year is the Sci-Tech Entrepreneurship AI Index, which has seen a year-to-date increase of 85.82% [16][14] Summary by Sections AI Industry Overview - The AI sector is witnessing a convergence of favorable market conditions and heightened interest, with significant events driving attention. The DeepSeek index, AI computing index, and AI application index have shown respective increases of 58.72%, 36.13%, and 50.76% year-to-date, outperforming major indices like the Shanghai Composite and CSI 300 [14][17] ETF Products and Performance - The report identifies 24 ETFs tracking AI indices, with the largest being the CS Artificial Intelligence Index ETF at 35.414 billion and the Sci-Tech AI Index ETF at 16.719 billion. The Sci-Tech Entrepreneurship AI Index has the highest year-to-date return of 85.82% [16][18] Stock Selection for AI ETFs - The report emphasizes the importance of selecting ETFs with high purity in AI-related stocks across the industry chain. It categorizes stocks into upstream (hardware), midstream (network and platform), and downstream (application scenarios) to identify ETFs with significant exposure to AI [3][25] Key Stocks in AI Hardware - The report lists key stocks in the AI hardware sector, including companies like Cambricon (寒武纪-U) and Inspur (工业富联), highlighting their market capitalizations and projected earnings per share (EPS) for 2025, 2026, and 2027 [28]