Investment Rating - The report does not explicitly provide an investment rating for the industry, but it discusses various quantitative strategies that have shown potential for generating excess returns. Core Insights - The report highlights the challenges and breakthroughs in quantitative industry rotation, emphasizing the limitations of traditional fundamental factors and the effectiveness of technical indicators like residual momentum and crowding indicators in generating excess returns [1][11]. - It proposes three key strategies for enhancing industry rotation: style timing to assist industry rotation, combining industry rotation with CTA signals, and applying large language models to industry rotation [1]. Summary by Sections Quantitative Industry Rotation Challenges - Traditional fundamental factors have faced significant challenges, particularly in 2024, where they nearly failed entirely due to unstable temporal logic and poor comparability across sectors [1][11]. - Technical indicators, such as residual momentum and crowding metrics, have proven resilient and continue to generate excess returns for investors [1][11]. Residual Momentum Factor - The residual momentum factor captures price-driving factors like industrial policy and technological advancements by excluding market and style influences. An improved version incorporates volatility reversal effects, achieving an annualized excess return of 12.90% from January 1, 2017, to February 28, 2026 [2][22]. - Despite a significant drawdown in excess returns from Q3 2024 to Q2 2025, the strategy quickly recovered and reached new highs in early 2026 [2][22]. Crowding Indicator - The crowding indicator model, based on threshold testing, successfully identified trading risks in sectors like defense, industrial metals, and precious metals at the beginning of 2026. A score of 3 or 4 on the crowding scale often precedes sector index peaks [3][29]. - The model's backtesting indicates that avoiding high-crowding sectors positively contributes to long-term performance [3][29]. Machine Learning Empowerment - Genetic programming, a classic factor mining method, has been enhanced through GPU acceleration and multi-objective frameworks, resulting in improved performance. From October 10, 2022, to February 28, 2026, the strategy achieved an annualized excess return of 25.39% [4][51]. - The report suggests that large language models may provide a breakthrough in understanding non-structured information for fundamental industry rotation [4][51]. Future Outlook - The report anticipates a "second half" for quantitative industry rotation, focusing on integrating style timing, CTA signals, and advanced machine learning techniques to adapt to market changes [1][4].

Market Performance - The Shanghai Composite Index rose by 1.19%, the Shenzhen Component Index increased by 2.4%, and the ChiNext Index surged by 3.39% [1] - The total market turnover reached 1.814 trillion yuan, an increase of 65.9 billion yuan compared to the previous day [1] Sector Highlights - Energy metals showed strong performance, with lithium mining leading the gains [1] - Computing hardware stocks saw a resurgence in the afternoon, particularly in the CPO and server sectors [1] - Brokerage, insurance, and aviation sectors had notable gains [1] - Commercial aerospace and Hainan Free Trade Zone themes experienced significant pullbacks [1] ETF Performance - The AI hardware sector rebounded strongly, with the Cathay Communications ETF, E Fund Growth ETF, and Southern ChiNext AI ETF all rising over 5% [1] - The rare metals sector also saw gains, with GF Fund Rare Metals ETF and Harvest Fund Rare Metals ETF increasing by 4.38% and 4.3%, respectively [1] - Two Brazilian ETFs fell by over 3% [1] - The satellite sector declined, with both the Aerospace ETF and General Aviation ETF dropping by more than 1% [1] - Crude oil futures decreased, leading to a 1.27% drop in the S&P Oil & Gas ETF [1]