协创数据在算力方面有哪些具体表现？ 在算力方面，协创数据表现出色。今年春节期间国产 AI 大模型井喷式发展，尤 其是在推理侧实现弯道超车，与海外大厂竞争。例如字节跳动 C 站引发现象级 传播，智谱和 Minimax 分别发布 GLM 5 和 M2.5，而阿里千问 3.5 Plus 性能 媲美 Gemini 3 Pro，总参数 3,970 亿但激活仅 170 亿，以小胜大部署显存占 较前代降低 60%，推理效率提升且 API 价格降至每百万 token 0.8 元，仅为谷 歌 1/18。这些发展推动云涨价逻辑，即 AI 推理所需 token 需求年化数十倍增 长，引发产业链全面通胀，从存储到 CPU、算力租赁再到云环环相扣。 近期 H100 等老卡价格不降反升，高算以及 B 卡价格也持续上涨，这主要由于折旧 期残值重估被全世界认可。此外，国产大模型迭代导致需求暴增，例如 GLM 5 上架仅 4 天就因服务排队、响应延迟及卡顿现象而多次扩容，但依然供不应求。 这种情况下，拥有稳定、高稳定性算力供给的厂商将站在 AI 推理时代战略制高 点。 协创数据 20260223 摘要 时租赁价格显著上涨，部分平台 A100 ...

截至2026年2月24日 11:13，创业板人工智能ETF南方(159382)上涨2.66%，盘中换手5.51%，成交1.26亿 元。跟踪指数创业板人工智能指数成分股润泽科技上涨15.23%，天孚通信上涨14.94%，北京君正上涨 14.34%，长芯博创，中际旭创等个股跟涨。 创业板人工智能ETF南方(159382)紧密跟踪创业板人工智能指数，创业板人工智能指数反映创业板人工 智能主题相关上市公司的股价变化情况。指数前十大权重股分别为中际旭创、新易盛、天孚通信、蓝色 光标、润泽科技、协创数据、北京君正、昆仑万维、同花顺、深信服。 创业板人工智能ETF南方(159382)，场外联接(A类：024725；C类：024726)。 消息面上，2026年春晚成为AI大厂竞逐超级入口的关键战场，字节跳动、阿里巴巴、腾讯、百度等头 部企业深度参与冠名、红包及技术支撑，将AI能力嵌入节目制作、实时互动与生活服务场景，旨在通 过国民级窗口教育用户、培养使用习惯。 开源证券认为，展望2026年，AI"虹吸效应"显著，全球AI或继续共振。海外方面，谷歌、Meta等巨头 不断上调AI资本开支指引，谷歌Gemini等大模型Tokens ...

Summary of Conference Call Records Industry Overview - The conference call discusses the AI infrastructure industry, highlighting the rapid advancements and product launches from major players such as Alibaba, ByteDance, Zhiyu, Kimi, Minimax, and Google during the Spring Festival period [1][1]. - The demand for AI infrastructure is increasing, particularly in the context of new storage architectures due to ongoing storage shortages [1][1]. Key Points and Arguments - **Token Usage Surge**: OpenClaw's token usage increased to approximately 13% of all tokens on OpenRouter within two weeks, indicating a growing interest in this platform [1][1]. - **Storage Shortages**: Hynix reported that all customer demands cannot be met, with DRAM and NAND inventory levels only sufficient for about four weeks [2][2]. - **New Product Launches**: SanDisk is expected to launch new HBF products, while there is a recommendation for new storage GPU-native database directions, particularly highlighting Xinghuan Technology's rapid progress [2][2]. - **Pricing Trends**: Following the release of new models by Zhiyu, there has been a simultaneous increase in GLMCoding Plan and API prices, reflecting the strong demand for computing power [2][2]. - **Rising Rental Prices**: The rental prices for computing power in overseas markets (H, A, B cards) continue to rise, indicating a robust demand for computational resources [2][2]. Additional Important Content - **Domestic Model Success**: Four Chinese models ranked among the top five globally in terms of usage from February 16 to February 22, showcasing China's competitive position in the global AI landscape [3][3]. - **Impact on Domestic Computing Demand**: The success of Chinese large models is expected to drive demand for domestic computing power, with relevant companies including Haiguang Information, Zhongke Shuguang, Cambrian, and others identified as key players [4][4]. - **Risks**: Potential risks include the underdevelopment of AI technology and geopolitical tensions between China and the United States [4][4].

Group 1 - The core viewpoint of the articles highlights the strong performance of A-shares in the early trading session post-holiday, particularly in the optical module and CPO concepts, with significant gains in stocks like Huagong Technology and Tianfu Communication [1] - The communication ETF (515050) has seen a rise of over 3%, while the AI-focused ETF (159381) has increased by more than 2%, indicating a robust market sentiment towards these sectors [1] - The global token call volume has reached a new high, with a week-on-week growth rate exceeding 15%, driven by frequent iterations of global models including ClaudeOpus4.6, Gemini3DeepThink, and GPT-5.3 [1] Group 2 - Huatai Securities projects that the demand for cloud computing will surge with the rapid development of AI by 2025, significantly boosting the need for general computing power [2] - Tokens are described as the purest form of electricity derivatives, with the value of electricity being delivered across borders through tokens, despite the physical electricity remaining within China [2] - The cost structure of tokens is heavily influenced by electricity and computing power, which together account for over 70% of the total cost, leveraging China's low electricity prices to gain global pricing power in AI services [2] Group 3 - The communication ETF (515050) tracks the CSI 5G communication theme index, with a total scale of nearly 8 billion yuan, covering key areas such as optical modules and computing infrastructure, with CPO and CPB concept stocks making up over 76% of its weight [3] - The AI-focused ETF (159381) has a balanced allocation between optical module CPO stocks and AI software applications, with top holdings including Zhongji Xuchuang and Xinyi Sheng, and a total scale of nearly 2 billion yuan [3] - The cloud computing ETF (516630) focuses on domestic AI hardware and software, with a combined weight of 83.7% in computer software, cloud services, and computer equipment, also featuring a low fee structure [3]

Core Viewpoint - The artificial intelligence (AI) industry is experiencing a significant surge, driven by the release of new AI models and a shift in focus towards productivity and efficiency in AI applications [1][2]. Group 1: Market Performance - As of February 24, 2026, the CSI Artificial Intelligence Theme Index (930713) rose by 1.29%, with notable gains from stocks such as Runze Technology (up 20.00%), Beijing Junzheng (up 3.52%), and Xiechuang Data (up 2.88%) [1]. - The AI Artificial Intelligence ETF (512930) increased by 1.10%, with the latest price reported at 2.3 yuan [1]. Group 2: Industry Developments - During the Spring Festival, both domestic and international AI developments accelerated, with major companies releasing updated AI models, including ByteDance's Seedance 2.0, Alibaba's Qwen 3.5, and Google's Gemini 3.1 Pro [1]. - The focus of the current AI model updates has shifted from a "parameter competition" to a "productivity competition," emphasizing the ability of models to solve complex real-world tasks [1]. Group 3: Index Composition - The CSI Artificial Intelligence Theme Index comprises 50 listed companies that provide foundational resources, technology, and application support for AI, reflecting the overall performance of AI-related securities [1]. - As of January 30, 2026, the top ten weighted stocks in the index include Zhongji Xuchuang, Xinyi Sheng, and Hanwujing, accounting for a total of 57.27% of the index [2].

长江商报消息●长江商报记者 沈右荣 买服务器，协创数据（300857.SZ）再出重磅。 春节前夕，协创数据发布公告称，根据经营发展需要，公司拟向多家供应商采购服务器，采购总金额预 计不超过110亿元。购买服务器主要用于为客户提供云算力服务。 2025年12月，协创数据曾公告，公司拟不超90亿元采购服务器。 长江商报记者梳理统计发现，2025年3月以来的不到一年时间，协创数据已经六次官宣拟采购服务器， 合计金额为322亿元。 作为领先的数据存储与智能终端解决方案提供商，协创数据在算力布局方面持续深化，且取得了不错的 经营业绩。 根据业绩预告，2025年，协创数据预计实现的归母净利润、扣非净利润区间均为10亿元至12.5亿元。如 果选择预计数的中位数，则较2023年的归母净利润、扣非净利润增长3倍左右。 公司积极布局的算力业务，已经贡献了公司40%的营业收入，成为新的增长引擎。 豪掷110亿加码算力 购买服务器，协创数据不断祭出大手笔。 2月12日晚，协创数据发布公告，公司及/或子公司根据经营发展需要，拟向多家供应商采购服务器，并 签署相关采购合同，采购合同总金额预计不超过110亿元。 豪掷超百亿元采购服务器，这 ...

本篇报告梳理了春节假期科技相关热点。全球多个 AI 大模型在春节期间发布重 磅更新，国产 AI 模型算力持续紧张。春晚带动机器人搜索量大增，宇树科技三 登春晚，我国机器人技术持续突破，产业商业化应用渐行渐近。 推荐（维持） 行业规模 | | | 占比% | | --- | --- | --- | | 股票家数（只） | 285 | 5.5 | | 总市值（十亿元） | 4481.3 | 4.0 | | 流通市值（十亿元） | 3977.0 | 3.9 | 行业指数 证券研究报告 | 行业定期报告 2026 年 02 月 23 日 春节假期科技热点梳理 计算机周观察 20260223 TMT 及中小盘/计算机 % 1m 6m 12m 绝对表现 -6.6 13.6 16.7 相对表现 -4.5 2.1 -2.6 资料来源：公司数据、招商证券 -30 -20 -10 0 10 20 30 Feb/25 Jun/25 Sep/25 Jan/26 (%) 计算机 沪深300 相关报告 1、《AI 军备竞赛持续升级，关注高 壁垒软件及云服务——计算机周观察 20260208》2026-02-08 2、《AI 入口大战开启 ...

Investment Rating - Industry rating: Recommended [5] Core Insights - The recent surge in China's AI models during the "Spring Festival" has shifted the focus from general chat tools to vertical productivity tools and real agent implementations, particularly in industrial video generation and engineering-level programming [1][13] - Major players like ByteDance, Alibaba, and Zhipu AI are leading the market, with competition now centered on efficiency, cost control, and scene adaptation rather than just parameter scale [1][13] - The demand for computing power is expected to boost hardware and computing rental markets, benefiting companies involved in computing leasing and third-party data centers [2][14] Summary by Sections Section 1: AI Model Developments - Multiple flagship models were released during the Spring Festival, enhancing capabilities in video and programming applications, with significant improvements in quality and efficiency [7][13] - ByteDance's Seedance 2.0 model supports multi-modal input and has shown substantial advancements in generating complex interactions and maintaining physical accuracy [8][9] - Alibaba's Qwen3.5-Plus model has achieved breakthroughs in multi-modal capabilities and cost efficiency, making it competitive with leading models globally [10][11] Section 2: Market Outlook - The current market is experiencing liquidity concerns and changing capital expenditure expectations from overseas players like OpenAI, leading to a cautious outlook for the AI sector [3][15] - The AI development phase is seen as a critical acceleration period, with increasing demand for token usage and ongoing expansion of foundational computing infrastructure [3][15] Section 3: Recommended Investment Targets - Key investment opportunities include companies in computing and communication infrastructure, such as China Mobile, China Telecom, and third-party data center firms like Huohuan New Network and Data Port [15][17] - The report highlights the potential of low-orbit satellite components and chip suppliers in the commercial aerospace sector [3][15] - Recommendations also extend to edge computing firms and those involved in satellite internet and low-altitude economy sectors [17][19]

Quantitative Models and Construction Methods 1. Model Name: 250-Day New High Distance Model - **Model Construction Idea**: This model tracks the distance of stock prices or indices from their 250-day high to identify market trends and hotspots. It is based on the momentum and trend-following strategy, which has been proven effective in various studies[11][18]. - **Model Construction Process**: The 250-day new high distance is calculated as follows: $ 250\text{-day new high distance} = 1 - \frac{Close_t}{ts\_max(Close, 250)} $ Where: - $ Close_t $ is the latest closing price - $ ts\_max(Close, 250) $ is the maximum closing price over the past 250 trading days If the latest closing price reaches a new high, the distance is 0. If the price falls from the high, the distance is a positive value, representing the degree of decline[11]. - **Model Evaluation**: The model effectively identifies stocks and indices with strong momentum and highlights market leaders, aligning with the principles of momentum investing[11][18]. 2. Model Name: Stable New High Stock Selection Model - **Model Construction Idea**: This model focuses on selecting stocks that exhibit stable price paths and consistent momentum, as smoother price trajectories tend to yield stronger momentum effects[24]. - **Model Construction Process**: Stocks are filtered from the pool of those that hit a 250-day high in the past 20 trading days. The selection criteria include: - Analyst attention: At least 5 buy or overweight ratings in the past 3 months - Relative strength: Top 20% in 250-day price performance - Price stability: Evaluated using the following metrics: - Absolute value of price changes over the past 120 days - Sum of absolute daily price changes over the past 120 days - Momentum continuity: - Average 250-day new high distance over the past 120 days - Average 250-day new high distance over the past 5 days The top 50 stocks based on these criteria are selected[24][26]. - **Model Evaluation**: The model emphasizes smooth price trajectories and consistent momentum, which are less likely to attract excessive attention, thereby enhancing the momentum effect[24]. --- Model Backtesting Results 1. 250-Day New High Distance Model - **Indices' 250-Day New High Distance (as of February 13, 2026)**: - Shanghai Composite Index: 2.00% - Shenzhen Component Index: 2.35% - CSI 300: 2.72% - CSI 500: 3.51% - CSI 1000: 3.14% - CSI 2000: 2.54% - ChiNext Index: 3.32% - STAR 50 Index: 5.50%[12][13]. 2. Stable New High Stock Selection Model - **Selected Stocks**: 50 stocks were identified, including Zhongtung High-tech, Dike Co., and Xiechuang Data. - **Sector Distribution**: - Technology: 21 stocks (most from the electronics industry) - Manufacturing: 16 stocks (most from the machinery industry)[27][32]. --- Quantitative Factors and Construction Methods 1. Factor Name: 250-Day New High Distance - **Factor Construction Idea**: Measures the relative position of a stock's price to its 250-day high, capturing momentum and trend-following characteristics[11]. - **Factor Construction Process**: $ 250\text{-day new high distance} = 1 - \frac{Close_t}{ts\_max(Close, 250)} $ Where: - $ Close_t $ is the latest closing price - $ ts\_max(Close, 250) $ is the maximum closing price over the past 250 trading days[11]. - **Factor Evaluation**: The factor effectively identifies stocks with strong momentum and aligns with established momentum strategies[11][18]. 2. Factor Name: Price Stability - **Factor Construction Idea**: Evaluates the smoothness of a stock's price trajectory, as smoother paths are associated with stronger momentum effects[24]. - **Factor Construction Process**: - Absolute value of price changes over the past 120 days - Sum of absolute daily price changes over the past 120 days[24]. - **Factor Evaluation**: The factor highlights stocks with stable price movements, which are less likely to attract excessive attention, enhancing the momentum effect[24]. --- Factor Backtesting Results 1. 250-Day New High Distance Factor - **Indices' 250-Day New High Distance (as of February 13, 2026)**: - Shanghai Composite Index: 2.00% - Shenzhen Component Index: 2.35% - CSI 300: 2.72% - CSI 500: 3.51% - CSI 1000: 3.14% - CSI 2000: 2.54% - ChiNext Index: 3.32% - STAR 50 Index: 5.50%[12][13]. 2. Price Stability Factor - **Selected Stocks**: 50 stocks were identified, including Zhongtung High-tech, Dike Co., and Xiechuang Data. - **Sector Distribution**: - Technology: 21 stocks (most from the electronics industry) - Manufacturing: 16 stocks (most from the machinery industry)[27][32].

- The report introduces a quantitative model called "250-day new high distance" to track market trends and identify market hotspots. The model is based on momentum and trend-following strategies, emphasizing the effectiveness of stocks reaching new highs as market indicators. The formula for calculating the 250-day new high distance is: $ 250\text{-day new high distance} = 1 - \frac{Close_t}{ts\_max(Close, 250)} $ where $ Close_t $ represents the latest closing price, and $ ts\_max(Close, 250) $ is the maximum closing price over the past 250 trading days. If the latest closing price reaches a new high, the distance is 0; otherwise, it is a positive value indicating the degree of fallback from the new high[11][12][13] - The report evaluates the model positively, citing its ability to capture market trends and identify leading stocks that perform well during market uptrends. It references studies by [George@2004], William O'Neil, and Mark Minervini, which support the idea that stocks near their 52-week highs tend to outperform those far from their highs[11][18][21] - The report provides backtesting results for the 250-day new high distance model. As of February 13, 2026, major indices such as the Shanghai Composite Index, Shenzhen Component Index, CSI 300, CSI 500, CSI 1000, CSI 2000, ChiNext Index, and STAR 50 Index have respective 250-day new high distances of 2.00%, 2.35%, 2.72%, 3.51%, 3.14%, 2.54%, 3.32%, and 5.50%[12][31] - The report introduces a quantitative factor called "Stable New High Stocks" to identify stocks with smooth price paths and consistent momentum. The factor construction involves screening stocks that have reached a 250-day new high in the past 20 trading days and applying criteria such as analyst attention (at least five buy or overweight ratings in the past three months), relative price strength (top 20% in 250-day returns), price path smoothness (measured by price displacement ratio), and sustained new high performance (average 250-day new high distance over the past 120 days and past 5 days)[24][26][27] - The report evaluates the factor positively, citing research by [Turan G Bali, Nusret et al@2011] and [Da, Gurun et al@2012], which highlight the superior performance of stocks with smooth price paths and strong momentum. The factor is designed to capture these characteristics effectively[24][26][27] - Backtesting results for the "Stable New High Stocks" factor show that 50 stocks were selected based on the criteria, with the highest representation in the technology and manufacturing sectors. Specifically, 21 stocks from the technology sector (dominated by the electronics industry) and 16 stocks from the manufacturing sector (dominated by the machinery industry) were included[27][32][30]