金融工程2026年度策略：拥抱AI投研巨浪，迎接量化投资新篇章

Group 1: Large Model Ecosystem and Applications - The iteration speed of large models remains high, with a stable ecosystem and trends expected in the short term, including the dominance of closed-source models and the increasing importance of multimodal capabilities [11][12][18] - The application of Agentic AI is accelerating, with a well-established infrastructure supporting rapid deployment in investment research, indicating a shift towards expert agents in the field [21][25] Group 2: 2026 Asset Allocation Strategy Outlook - The macroeconomic environment is currently in a weak recovery phase, with manufacturing PMI and PPI showing gradual improvement, suggesting a potential for inflation to rise in 2026 due to external factors like interest rate cuts and AI-driven capital expenditures [53][56][62] - The report anticipates a dual-line market trend of cyclical and technological growth, with a shift in style allocation from small-cap growth to large-cap balance, and a focus on fundamental factors in industry allocation [2][56] Group 3: Factor Stock Selection Outlook - The trend of using AI models for stock selection has increased, but the strategies have become crowded, leading to potential collective drawdowns; optimization methods are being explored to enhance model performance [2][3][21] - The introduction of advanced techniques such as Huber Loss and memory modules aims to reduce excess drawdowns and improve the models' adaptability to market fluctuations [2][3][21] Group 4: 2025 Equity Fund Investment Outlook - Active equity funds are expected to see a return of alpha, particularly in the context of a dual-line market of technology and cyclical sectors, with recommendations for both broad-based and thematic funds [3][4][30] - The new regulations on performance benchmarks are likely to shift the focus towards stock selection alpha as a primary source of excess returns [3][4][30]