Investment Rating - The report maintains an "Outperform" rating for the industry [1] Core Insights - The rapid iteration of large models is enhancing application capabilities, with global AI development leading to significant improvements in knowledge Q&A, mathematics, and programming, surpassing human-level performance in various tasks [2][4] - The increase in token usage is elevating the ranking of domestic models, with notable growth in API call volumes for Chinese models, indicating improved performance and cost-effectiveness [2][12] - AI applications are driving growth in the cloud market, leading to an expansion in IDC demand, as domestic internet and cloud companies lag behind their overseas counterparts in capital expenditure on AI infrastructure [2][3] Summary by Sections 1. Rapid Iteration of Large Models - The global large model industry has transitioned from annual to quarterly or even monthly iterations since 2025, with leading companies significantly reducing their model update cycles [11] - Domestic companies like Deepseek and ByteDance are also accelerating their model iterations, enhancing their capabilities and performance [11][12] 2. Increase in Token Usage and Domestic Model Ranking - The launch of viral AI applications like OpenClaw has spurred global AI application growth, leading to record-high token consumption [2] - By March 2026, over 50% of the top ten models on Openrouter were domestic, reflecting a significant rise in the performance and market acceptance of Chinese models [2] 3. AI Applications Driving Cloud Market Growth - The surge in domestic model usage is increasing the demand for local data centers, with a notable gap in capital expenditure on AI infrastructure compared to international firms [2] - As AI applications commercialize and grow rapidly, cloud services are becoming the primary platform for these applications, resulting in increased IaaS demand [2][3]

Group 1: Market Overview - The bond market is currently testing the 1.8 level, despite recent weakness, as inflation expectations rise domestically and internationally[1] - The current scale of wealth management products has returned to a high of 33 trillion yuan, with a shift towards multi-asset strategies due to insufficient bond coupon attractiveness[1] - The market is concerned about a potential slowdown in growth, exacerbated by input cost shocks and weak pricing power in the manufacturing sector[1] Group 2: Economic Indicators - Fixed asset investment has decreased by 3.80% year-on-year, while retail sales have increased by 0.90% year-on-year[5] - Exports have surged by 39.60% year-on-year, and M2 growth stands at 9.00%[5] Group 3: Asset Performance - From March 7 to March 14, the CSI 300 index rose by 0.19%, while the Hang Seng index fell by 1.14% and the S&P 500 dropped by 1.6%[13] - The 10-year government bond yield increased by 3.34 basis points, while the 10-year U.S. Treasury yield rose by 13 basis points[13] Group 4: Inventory and Fund Behavior - Crude oil inventory increased to 44,684 million tons, up by 46,789 million tons from the previous week[24] - The latest week saw a rise in long positions for the U.S. dollar to 16,384 contracts, an increase of 1,323 contracts[28]

Market Overview - The Hang Seng Index closed at 25,717, down 0.70% for the day and up 0.34% year-to-date[3] - The S&P 500 index closed at 6,776, down 1.52% for the day and down 2.53% year-to-date[3] - Brent crude oil prices surged to nearly $100 per barrel, raising inflation concerns[9] Sector Performance - The coal sector showed strength, with Feishang Anthracite rising over 18% and Nanshan Resources up over 12% due to rising oil and gas prices[9] - Renewable energy stocks also gained, with Datang New Energy up over 8% and Goldwind Technology up over 7% following the UK’s announcement to eliminate import tariffs on wind power components[9] - AI application stocks faced declines, with Zhizhu falling nearly 9% and MINIMAX down 5% due to regulatory concerns[9] Geopolitical Impact - Iran's new Supreme Leader, Mujtaba Khamenei, signaled a continuation of aggressive strategies, including the potential blockade of the Strait of Hormuz, impacting oil supply expectations[9] - The market's expectation for a quick resolution to Middle Eastern conflicts diminished, leading to widespread sell-offs in U.S. markets[9] Economic Indicators - The Federal Reserve is expected to maintain interest rates during the upcoming meeting, despite rising inflation pressures from geopolitical tensions[9] - The U.S. trade deficit narrowed more than expected in January, indicating some resilience in the economy[12]

1. Report Industry Investment Rating No relevant content provided. 2. Core Viewpoints of the Report - Last week (from March 9, 2026, to March 13, 2026), the median weekly return of equity ETFs was -0.28%. Among broad-based ETFs, the median return of ChiNext ETFs was 2.49%, the highest. By sector, the median return of cyclical ETFs was -0.12%, the smallest decline. By theme, the median return of new energy vehicle ETFs was 5.01%, the highest [1][12]. - Last week, equity ETFs had a net redemption of 8.649 billion yuan, and the overall scale decreased by 32.72 billion yuan. Among broad-based ETFs, the Shanghai 50 ETF had the least net redemption, at 598 million yuan. By sector, technology ETFs had the most net subscriptions, at 5.178 billion yuan. By theme, chip ETFs had the most net subscriptions, at 4.572 billion yuan [2][28][33]. - As of last Friday, the top three fund companies in terms of the total scale of listed, non-monetary ETFs were Huaxia, E Fund, and Huatai-PineBridge. This week, 7 ETFs will be issued, including Huabao CSI All-Share Household Appliances ETF, Fullgoal CSI Hong Kong Stock Connect Information Technology Comprehensive ETF, etc. [5][56] 3. Summary by Relevant Catalogs ETF Performance - Last week, the median weekly return of equity ETFs was -0.28%. The median returns of ChiNext, CSI 300, A500, CSI 1000, Shanghai 50, CSI 500, and STAR Market ETFs were 2.49%, 0.17%, -0.26%, -0.42%, -1.21%, -1.44%, and -1.89% respectively. The median returns of money market, bond, cross-border, and commodity ETFs were 0.02%, -0.00%, -0.73%, and -0.73% respectively [12]. - By sector, the median returns of cyclical, consumer, large financial, and technology sector ETFs were -0.12%, -0.13%, -1.41%, and -1.89% respectively. By theme, the median returns of new energy vehicle, power, and photovoltaic ETFs were 5.01%, 4.02%, and 3.55% respectively, showing relatively strong performance, while the median returns of military, resource, and AI ETFs were -6.86%, -4.00%, and -3.85% respectively, showing relatively weak performance [16]. ETF Scale Changes and Net Subscriptions/Redeemptions - As of last Friday, the scales of equity, cross-border, and bond ETFs were 3.0086 trillion yuan, 981.2 billion yuan, and 726.1 billion yuan respectively. The scales of commodity and money market ETFs were relatively small, at 360.1 billion yuan and 170.9 billion yuan respectively. Among broad-based ETFs, the CSI 300 and A500 ETFs had relatively large scales, at 571.4 billion yuan and 242.7 billion yuan respectively [19]. - By sector, the scale of technology sector ETFs was 511.3 billion yuan as of last Friday, followed by cyclical sector ETFs at 401.5 billion yuan. The scales of consumer and large financial ETFs were relatively small, at 198.7 billion yuan and 192.5 billion yuan respectively. By theme, the scales of chip, resource, and securities ETFs were the highest, at 184.2 billion yuan, 141.7 billion yuan, and 136.7 billion yuan respectively [26]. - Last week, equity ETFs had a net redemption of 8.649 billion yuan, and the overall scale decreased by 32.72 billion yuan. Money market ETFs had a net redemption of 2.369 billion yuan, and the overall scale decreased by 2.354 billion yuan. Among broad-based ETFs, the Shanghai 50 ETF had the least net redemption, at 598 million yuan, and its scale decreased by 1.533 billion yuan. ChiNext ETFs had the most net redemptions, at 6.204 billion yuan, and their scale decreased by 3.205 billion yuan [28]. - By sector, technology ETFs had the most net subscriptions last week, at 5.178 billion yuan, and their scale decreased by 7.419 billion yuan. Large financial ETFs had the most net redemptions, at 83 million yuan, and their scale decreased by 2.533 billion yuan. By theme, chip ETFs had the most net subscriptions, at 4.572 billion yuan, and their scale decreased by 2.133 billion yuan. Resource ETFs had the most net redemptions, at 7.247 billion yuan, and their scale decreased by 12.551 billion yuan [31]. ETF Benchmark Index Valuation - As of last Friday, the price-to-earnings ratios of Shanghai 50, CSI 300, CSI 500, CSI 1000, ChiNext, and A500 ETFs were at the 80.68%, 89.93%, 97.27%, 97.69%, 63.17%, and 91.69% quantile levels respectively, and the price-to-book ratios were at the 54.95%, 77.29%, 97.52%, 81.50%, 65.24%, and 93.59% quantile levels respectively. Since December 31, 2019, the price-to-earnings and price-to-book ratios of STAR Market ETFs are currently at the 74.57% and 77.37% quantile levels respectively [34]. - As of last Friday, the price-to-earnings ratios of cyclical, large financial, consumer, and technology sector ETFs were at the 88.93%, 23.04%, 20.48%, and 91.82% quantile levels respectively, and their price-to-book ratios were at the 94.63%, 35.14%, 23.78%, and 83.65% quantile levels respectively [37]. - As of last Friday, the price-to-earnings ratio quantiles of photovoltaic, dividend, and military ETFs were relatively high, at 99.92%, 98.51%, and 96.45% respectively. The price-to-book ratio quantiles of dividend, AI, and robot ETFs were relatively high, at 99.83%, 95.71%, and 90.67% respectively. Compared with the previous week, the valuation quantile of power ETFs increased significantly. Overall, among broad-based ETFs, the valuation quantile of ChiNext ETFs was relatively low. By sector, the valuation quantiles of consumer and large financial ETFs were relatively moderate. By theme, the valuation quantiles of liquor and securities ETFs were relatively low [42][46]. ETF Margin Trading and Short Selling - Overall, the margin balance and short selling volume of equity ETFs have both increased in the past year. As of last Thursday, the margin balance of equity ETFs decreased from 47.824 billion yuan in the previous week to 47.386 billion yuan, and the short selling volume increased from 2.385 billion shares in the previous week to 2.5 billion shares [47]. - Among the top 10 ETFs with the highest average daily margin purchases and short selling volumes from last Monday to Thursday, securities ETFs and STAR Market ETFs had relatively high average daily margin purchases, and CSI 1000 ETFs and CSI 500 ETFs had relatively high average daily short selling volumes [50][52]. ETF Managers - As of last Friday, Huaxia Fund ranked first in the total scale of listed non-monetary ETFs and had a relatively high management scale in multiple sub - fields such as scale index ETFs, theme, style, and strategy index ETFs, and cross - border ETFs. E Fund ranked second, with a relatively high management scale in scale index ETFs and cross - border ETFs. Huatai - PineBridge Fund ranked third, with a relatively high management scale in scale index ETFs and theme, style, and strategy index ETFs [53]. - Last week, 10 new ETFs were established. This week, 7 ETFs will be issued, including Huabao CSI All - Share Household Appliances ETF, Fullgoal CSI Hong Kong Stock Connect Information Technology Comprehensive ETF, etc. [56].

- The CSI Livestock Breeding Industry Index (931946.CSI) was launched on July 28, 2023, selecting securities from industries such as livestock products, animal health and breeding, feed, meat products, and dairy products to reflect the overall performance of listed companies in the livestock breeding industry[31][50] - The index is highly concentrated, with the top 10 weighted stocks accounting for 66.47% as of February 28, 2026. Major contributors include leading companies in pig farming such as Muyuan Foods and Wen's Foodstuff, which together account for nearly 30% of the index weight. Livestock farming represents over 54% of the index weight, while feed processing accounts for 15.41%, showcasing the index's pure representation of the industry chain[32][33][50] - The index exhibits small-cap characteristics, with an average market capitalization of 239.67 billion yuan across its 50 constituent stocks. Compared to broader indices like CSI 300 and CSI 800, the CSI Livestock Breeding Industry Index has a significantly lower average market cap, positioning it between CSI 500 and CSI 1000[39][41][50] - The index is valued at historical lows, with a price-to-earnings (P/E) ratio of 20.48 and a price-to-book (P/B) ratio of 2.75 as of March 11, 2026. The P/E ratio is at the 36.55% percentile, while the P/B ratio is at the 76.90% percentile, indicating a relatively low valuation and offering a safety margin for investors[42][43][50] - Performance-wise, the index has demonstrated stronger resilience compared to similar indices during the livestock industry's cyclical downturn. Since 2022, the CSI Livestock Breeding Industry Index has achieved an annualized return of -5.64%, outperforming the CSI Livestock Index, which recorded an annualized return of -7.61%[45][47][50]

Group 1 - The report highlights that the four active quantitative strategies have performed well, ranking in the top 25% of active equity funds this year [1][12][13] - The "Excellent Fund Performance Enhancement Portfolio" achieved an absolute return of 9.06% this year, outperforming the mixed equity fund index by 4.53% [1][25] - The "Super Expected Selection Portfolio" recorded an absolute return of 11.91% this year, with a relative outperformance of 7.38% against the mixed equity fund index [1][33] Group 2 - The "Brokerage Golden Stock Performance Enhancement Portfolio" has an absolute return of 11.08% this year, outperforming the mixed equity fund index by 6.55% [1][38] - The "Growth Stability Portfolio" achieved an absolute return of 15.54% this year, with a relative outperformance of 11.01% against the mixed equity fund index [1][45] - The report indicates that the median return for stocks this year is 3.94%, with 62% of stocks rising and 38% falling [2][51] Group 3 - The "Excellent Fund Performance Enhancement Portfolio" is constructed by benchmarking against active equity funds, utilizing quantitative methods to enhance performance [3][18] - The "Super Expected Selection Portfolio" selects stocks based on positive earnings surprises and analyst upgrades, focusing on both fundamental and technical criteria [4][58] - The "Brokerage Golden Stock Performance Enhancement Portfolio" is based on a selection of stocks from the brokerage's top picks, aiming to optimize performance against the mixed equity fund index [5][63] Group 4 - The "Growth Stability Portfolio" employs a two-dimensional evaluation system for growth stocks, prioritizing those closer to their earnings report dates to capture potential excess returns [6][68] - The report emphasizes that the active quantitative strategies aim to outperform the median of active equity funds, with a focus on stability and consistent performance [12][54]

Quantitative Models and Construction Methods Model Name: Maximized Factor Exposure Portfolio (MFE) - **Model Construction Idea**: The MFE portfolio is designed to test the effectiveness of single factors under real-world constraints, such as industry exposure, style exposure, stock weight limits, and turnover rate. This approach ensures that factors deemed "effective" can genuinely contribute to return prediction in practical portfolio construction [42][43][44]. - **Model Construction Process**: - The optimization model aims to maximize single-factor exposure, with the objective function defined as: $$ \begin{array}{ll} \max & f^{T}w \\ \text{s.t.} & s_{l} \leq X(w-w_{b}) \leq s_{h} \\ & h_{l} \leq H(w-w_{b}) \leq h_{h} \\ & w_{l} \leq w-w_{b} \leq w_{h} \\ & b_{l} \leq B_{b}w \leq b_{h} \\ & \mathbf{0} \leq w \leq l \\ & \mathbf{1}^{T}w = 1 \end{array} $$ - **Explanation of Parameters**: - \( f \): Factor values - \( w \): Stock weight vector - \( X \): Factor exposure matrix for style factors - \( H \): Industry exposure matrix - \( w_b \): Benchmark index weight vector - \( s_l, s_h \): Lower and upper bounds for style factor exposure - \( h_l, h_h \): Lower and upper bounds for industry exposure - \( w_l, w_h \): Lower and upper bounds for individual stock weight deviation - \( b_l, b_h \): Lower and upper bounds for benchmark component weights - Constraints include: - Style factor exposure limits (\( s_l, s_h \)) [43] - Industry exposure limits (\( h_l, h_h \)) [43] - Individual stock weight deviation limits (\( w_l, w_h \)) [43] - Benchmark component weight limits (\( b_l, b_h \)) [43] - No short selling and full investment (\( \mathbf{1}^{T}w = 1 \)) [44] - The portfolio is rebalanced monthly, and historical returns are calculated after accounting for transaction costs (0.3% on both sides) [46]. - **Model Evaluation**: The MFE portfolio effectively identifies factors that can predict returns under practical constraints, making it a robust tool for factor validation in real-world scenarios [42][43]. --- Quantitative Factors and Construction Methods Factor Name: EPTTM (Earnings-to-Price Trailing Twelve Months) - **Factor Construction Idea**: Measures valuation by comparing trailing twelve-month earnings to market capitalization [17]. - **Factor Construction Process**: - Formula: \( \text{EPTTM} = \frac{\text{Net Income (TTM)}}{\text{Market Capitalization}} \) [17]. - **Factor Evaluation**: Demonstrates strong performance across multiple sample spaces, particularly in valuation-based strategies [19][20][22]. Factor Name: Pre-Expected EPTTM - **Factor Construction Idea**: Uses consensus analyst expectations for earnings to calculate a forward-looking valuation metric [17]. - **Factor Construction Process**: - Formula: \( \text{Pre-Expected EPTTM} = \frac{\text{Consensus Expected Earnings (TTM)}}{\text{Market Capitalization}} \) [17]. - **Factor Evaluation**: Consistently ranks among the top-performing factors in various sample spaces, indicating its predictive power for returns [19][20][22]. Factor Name: BP (Book-to-Price) - **Factor Construction Idea**: Measures valuation by comparing book value to market capitalization [17]. - **Factor Construction Process**: - Formula: \( \text{BP} = \frac{\text{Book Value}}{\text{Market Capitalization}} \) [17]. - **Factor Evaluation**: Shows strong historical performance, particularly in mid-cap and small-cap sample spaces [20][22]. Factor Name: Three-Month Reversal - **Factor Construction Idea**: Captures short-term mean reversion by measuring the price change over the past three months [17]. - **Factor Construction Process**: - Formula: \( \text{Three-Month Reversal} = \text{Cumulative Return over Last 60 Trading Days} \) [17]. - **Factor Evaluation**: Effective in identifying short-term price corrections, though performance varies across sample spaces [19][20][22]. --- Factor Backtesting Results Performance in CSI 300 Sample Space - **Top-Performing Factors (Recent Week)**: Pre-Expected EPTTM (1.64%), EPTTM (1.34%), EPTTM Year Percentile (1.32%) [19]. - **Underperforming Factors (Recent Week)**: Single-Quarter Operating Profit YoY (-0.87%), One-Year Momentum (-0.83%), Pre-Expected Net Profit QoQ (-0.79%) [19]. Performance in CSI 500 Sample Space - **Top-Performing Factors (Recent Week)**: Pre-Expected EPTTM (2.90%), Single-Quarter EP (2.40%), Standardized Unexpected Earnings (2.25%) [20]. - **Underperforming Factors (Recent Week)**: Pre-Expected Net Profit QoQ (0.06%), Standardized Unexpected Revenue (0.13%), Three-Month Institutional Coverage (0.22%) [20]. Performance in CSI 1000 Sample Space - **Top-Performing Factors (Recent Week)**: Pre-Expected PEG (1.87%), Three-Month Reversal (1.33%), Pre-Expected BP (1.19%) [22]. - **Underperforming Factors (Recent Week)**: One-Year Momentum (-1.91%), Single-Quarter ROA (-0.62%), Three-Month Institutional Coverage (-0.58%) [22]. Performance in CSI A500 Sample Space - **Top-Performing Factors (Recent Week)**: Pre-Expected EPTTM (2.51%), EPTTM (1.99%), Three-Month Reversal (1.91%) [25]. - **Underperforming Factors (Recent Week)**: One-Year Momentum (-1.07%), Three-Month Institutional Coverage (-0.41%), Pre-Expected Net Profit QoQ (-0.18%) [25]. Performance in Public Fund Heavyweight Index Sample Space - **Top-Performing Factors (Recent Week)**: Pre-Expected EPTTM, Pre-Expected PEG, Three-Month Reversal [27]. - **Underperforming Factors (Recent Week)**: One-Year Momentum, Three-Month Institutional Coverage, Standardized Unexpected Revenue [27].

- The "Hong Kong Stock Selection Portfolio" strategy aims to construct a portfolio by dual-layer screening based on fundamental and technical aspects of stocks recommended by analysts. The portfolio includes stocks with fundamental support and technical resonance, focusing on analyst recommendation events such as upward earnings revisions, initial coverage, and unexpected research report titles. The backtesting period is from January 1, 2010, to December 31, 2025, with an annualized return of 19.08% and an excess return of 18.06% relative to the Hang Seng Index after considering transaction costs in a fully invested state[13][15][19] - The "Stable New High Stock Screening Method" identifies stocks that have reached a 250-day high in the past 20 trading days. The screening criteria include analyst attention (at least 5 buy or overweight ratings in the past 6 months), relative stock strength (top 20% in 250-day returns), and stock price stability. Stability is assessed using metrics like price path smoothness and the average 250-day high distance over the past 120 days and the last 5 days. Stocks are ranked, and the top 50% are selected, with a minimum of 50 stocks[20][22][23] - The formula for calculating the "250-day high distance" is: $ 250\text{-day high distance} = 1 - \frac{\text{Close}_{t}}{\text{ts\_max}(\text{Close}, 250)} $ where $\text{Close}_{t}$ represents the latest closing price, and $\text{ts\_max}(\text{Close}, 250)$ is the maximum closing price over the past 250 trading days. A value of 0 indicates a new high, while positive values represent the percentage drop from the high[22] - Backtesting results for the "Hong Kong Stock Selection Portfolio" show annualized returns of 19.08%, excess returns of 18.06%, and various performance metrics such as IR (1.19), tracking error (14.60%), and maximum drawdown (23.73%). The portfolio consistently outperformed the Hang Seng Index across multiple years, with notable excess returns in years like 2020 (70.00%) and 2019 (33.78%)[15][19][17] - The "Stable New High Stock Screening Method" identified stocks like Yancoal Australia, China Shenhua, and China Ocean Oil as stable new high performers. The sector distribution includes 8 stocks from cyclical industries, followed by technology (4 stocks), manufacturing (3 stocks), pharmaceuticals (1 stock), and financials (1 stock)[22][23][27]

Investment Rating - The investment rating for the company is "Outperform the Market" [5][42]. Core Views - The company is expected to achieve a net profit of 77.201 billion yuan in 2025, representing a year-on-year increase of 42%. Revenue is projected to reach 423.702 billion yuan, up 17% year-on-year. The gross margin is forecasted at 26.27%, an increase of 1.83 percentage points, while the net margin is expected to be 18.12%, up 3.20 percentage points [1][8]. - The company is accelerating its capacity expansion to meet strong demand, with a capacity utilization rate of 96.9% expected by the end of 2025. The total planned capacity is projected to reach 1,093 GWh, an increase of 198 GWh from the end of 2024 [3][38]. - The company maintains a leading position in the power battery market, with a global market share of 39.2% in 2025, marking a year-on-year increase of 1.2 percentage points. The company has held the top position in the global market for nine consecutive years [21][26]. Summary by Sections Financial Performance - In Q4 2025, the company is expected to achieve revenue of 1,406.30 billion yuan, a year-on-year increase of 37% and a quarter-on-quarter increase of 35%. The net profit for the same period is projected to be 231.67 billion yuan, up 57% year-on-year and 25% quarter-on-quarter [1][8]. - The company’s operating income for 2025 is forecasted at 423.702 billion yuan, with a net profit of 72.201 billion yuan, and a gross margin of 26.27% [4][8]. Product and Market Position - The company’s power battery revenue is expected to reach 316.51 billion yuan in 2025, with a stable gross margin of 23.8%. The sales volume is projected to be 541 GWh, reflecting a year-on-year increase of 42% [21][26]. - The company has introduced several new products in 2025, including the second-generation supercharging battery and sodium-ion batteries, reinforcing its competitive advantage [21][33]. Capacity Expansion and Globalization - By the end of 2025, the company will have established a production capacity of 772 GWh, with an additional 321 GWh under construction. The company is also expanding its global footprint with ongoing projects in Germany, Hungary, Spain, and Indonesia [3][38]. - The company’s strategic partnerships with various firms aim to enhance its service capabilities and achieve mutual benefits [2][38].

- The report does not contain any quantitative models or factors for analysis[1][2][3]