研究早观点 2025 年 8 月 26 日 星期二 市场走势 资料来源：最闻 李明阳 执业登记编码：S0760525050002 邮箱：limingyang@sxzq.com 【今日要点】 【公司评论】应流股份(603308.SH)：应流股份 2025 中报点评-两机业 务高景气，在手订单充沛 【公司评论】天融信(002212.SZ)：天融信（002212.SZ）：二季度收入实现 正增长，一体机成为新增长点 【公司评论】深信服(300454.SZ)：深信服（300454.SZ）：收入端稳健增长， Q2 单季度实现扭亏 国内市场主要指数 | 指数 | | 收盘 | 涨跌幅% | | --- | --- | --- | --- | | 上证指数 | | 3,883.56 | 1.51 | | 深证成指 | | 12,441.07 | 2.26 | | 沪深 | 300 | 4,469.22 | 2.08 | | 中小板指 | | 7,580.95 | 1.79 | | 创业板指 | | 2,762.99 | 3.00 | | 科创 50 | | 1,287.73 | 3.20 | 资料来源：最闻 分析师： 【山证 ...

通用计算机设备 天融信（002212.SZ） 增持-A(维持) 二季度收入实现正增长，一体机成为新增长点 2025 年 8 月 25 日 公司研究/公司快报 公司近一年市场表现 事件描述 | 市场数据：2025 年 8 | 25 | | | 日 | 月 | | --- | --- | --- | --- | --- | --- | | 收盘价（元）： | | | 10.23 | | | | 年内最高/最低（元）： | | 11.17/3.99 | | | | | 流通 股/总股本（亿 A | | 11.67/11.79 | | | | | 股）： | | | | | | | 流通 A 股市值（亿）： | | | 119.37 | | | | 总市值（亿）： | | | 120.66 | | | | 年 基础数据：2025 | 月 6 | 日 30 | | | --- | --- | --- | --- | | 基本每股收益(元)： | | | -0.06 | | 摊薄每股收益(元)： | | | -0.06 | | 每股净资产(元)： | | | 7.91 | | 净资产收益率(%)： | | | -0.70 ...

（3）2025 年上半年，公司持续推进提质增效战略，收入质量持续提升，毛利率同比提升4.1 个百分 点；持续优化内部管理，提升运营效率，销售费用、研发费用及管理费用均实现同比下降，三项费用总 额同比减少14.04%。 聚焦网安和智算云业务，积极拥抱AI 产业趋势一方面，公司将人工智能技术与产品深度融合，实现威 胁检测精度提升、告警误报率下降及未知威胁对抗能力增强，带动产品附加值与市占率双升；同时以 AI替代重复性运营作业，显著降低交付成本，提升人效与盈利水平。另一方面，面向人工智能大模型 规模化应用，公司推出大模型安全防护网关、智算云平台、智算一体机等创新产品，打造"算力+安 全"融合方案，抢占增量市场，形成第二增长曲线。 国内领先的网络安全、大数据与云服务提供商 考虑网络安全下游景气度波动，我们下调公司2025-2026 年归母净利润预测为1.87、2.57 亿元（前值为 3.51、4.62 亿元），新增2027 年预测为3.01 亿元，EPS为0.16、0.22、0.26 元/股，当前股价对应2025- 2027 年PE 为64.0、46.6、39.8倍，考虑公司在云安全、AI、数据安全等领域的前瞻布局， ...

Group 1 - ZTE Corporation's A-shares surged to a limit up and H-shares rose over 14% due to a recent research report from Zheshang Securities highlighting the undervaluation of its AI computing and network business [1] - The report indicates that ZTE's progress in super nodes is exceeding expectations, positioning the company strategically in the super node complete machine, Switch Tray, and internal interconnection chips [1] - ZTE won significant contracts in a recent procurement project by China Mobile for AI general computing devices, securing over 70% of the largest package, amounting to approximately 8.85 billion RMB [1] Group 2 - In its quarterly report, ZTE capitalized on the increasing investment in computing power by major domestic internet companies and operators, leading to rapid growth in server and storage revenue [2] - The company reported that over 60% of its intelligent computing server orders are driven by the DeepSeek demand, with applications across various industries [2] - BlackRock has increased its stake in ZTE, indicating strong investment confidence, with holdings rising from 5.81% to 6.32% and then from 6.98% to 7.43% in August [2] Group 3 - Goldman Sachs predicts a 12% quarter-on-quarter increase in ZTE's net profit for the second quarter, estimating it to reach 2.7 billion RMB, driven by AI innovation [3] - The expected revenue for ZTE in the second quarter is projected to grow by 12% year-on-year to 35.6 billion RMB, supported by rising demand for AI infrastructure and devices [3] - The gross margin for the second quarter is anticipated to improve to 37.8%, up from 34.3% in the first quarter [3]

Core Viewpoint - ZTE Corporation's stock has experienced significant fluctuations, with a notable surge attributed to a recent report from Zheshang Securities highlighting the undervaluation of its AI computing and network business progress, suggesting that AI could rejuvenate the company [1][5]. Group 1: Stock Performance - On August 21, ZTE's A-shares hit the daily limit, while H-shares rose over 14% before retracting [1][3]. - The stock has seen a substantial increase, with A-shares rising over 20% this week and more than 4% the previous week [3][7]. Group 2: Research Report Impact - The surge in ZTE's stock is closely linked to a report from Zheshang Securities, which is the only report on ZTE in the last three months, emphasizing the company's strategic position in AI and network capabilities [5][6]. - The report indicates that ZTE's core capabilities in super nodes, Switch Tray, and internal interconnection chips position it favorably in the market [5]. Group 3: Recent Contracts and Market Opportunities - ZTE recently won significant contracts in a procurement project by China Mobile for AI general computing devices, securing approximately 70% of the largest package, amounting to around 8.85 billion yuan [5][6]. - The company has capitalized on the growing demand for computing power, with rapid growth in server and storage revenue, particularly in intelligent computing servers [6]. Group 4: Foreign Investment Trends - Foreign investment in ZTE shows mixed signals, with BlackRock increasing its stake from 5.81% to 6.32% and then to 7.43%, indicating confidence in the company [9]. - Conversely, UBS reduced its stake from 7.05% to 6.94%, reflecting a potential shift in investment strategy [9]. - Goldman Sachs forecasts a 12% quarter-on-quarter increase in ZTE's net profit, projecting revenues to rise by 12% year-on-year, driven by demand for AI infrastructure and devices [9].

Core Insights - The emergence of integrated AI computing machines, driven by DeepSeek, is making AI large models more accessible to the mass market, but challenges remain in application deployment [1][2] - The evolution of AI large models continues, with a focus on reducing model hallucinations and adapting to ongoing technological advancements [1][4] Group 1: Integrated AI Computing Machines - Integrated AI computing machines are pre-integrated solutions combining hardware, software platforms, models, and applications, lowering the barriers to AI adoption [2][3] - The current market features a diverse range of integrated AI computing machines, necessitating differentiation through performance optimization, customization, and business innovation to avoid price wars [2][3] - Key focus areas for integrated AI computing machines include computing power, model accessibility, and application development to meet diverse enterprise needs [2][3] Group 2: Application and Industry Impact - The deployment of integrated AI computing machines is not a one-time process; it requires continuous adjustment of computing power and model capabilities based on business needs [3][4] - The introduction of AIS one-stop intelligent platforms aids enterprises in quickly developing industry-specific models and exploring potential value scenarios [4][5] - Current applications of integrated AI computing machines are primarily in knowledge Q&A, customer assistance, and code assistance, but face challenges such as data quality and integration into existing workflows [3][4] Group 3: Addressing Model Hallucinations - Model hallucinations arise from the interplay of models, data, and tasks, and are an inherent risk in generative models [5][6] - Short-term solutions to mitigate hallucinations include using RAG, supervised fine-tuning, and exploring verification mechanisms [5][6] - The relationship between integrated AI computing machines and agents is synergistic, with each enhancing the capabilities of the other, leading to more efficient solutions [6]

Market Overview - On July 29, the A-share market experienced a fluctuating upward trend, with the ChiNext Index leading the gains. The Shanghai Composite Index rose by 0.33%, the Shenzhen Component Index increased by 0.64%, and the ChiNext Index surged by 1.86% [1][2] - The total trading volume in the Shanghai and Shenzhen markets reached 1.8 trillion yuan, an increase of 609 billion yuan compared to the previous trading day [1] Sector Performance - The market showed a mixed performance with over 3,000 stocks declining. Notable sectors included a significant rise in pharmaceutical stocks, particularly in the CRO and innovative drug sectors, while CPO and computing hardware stocks also showed strength [1] - The top-performing sectors included CRO, steel, CPO, and advanced packaging, while insurance, pork, banking, and agriculture sectors faced declines [1] International Market - The U.S. stock market saw declines on July 29, with the Dow Jones Industrial Average falling by 204.57 points (0.46%) to close at 44,632.99 points, the S&P 500 dropping by 18.91 points (0.30%) to 6,370.86 points, and the Nasdaq Composite decreasing by 80.29 points (0.38%) to 21,098.29 points [3][4] - In contrast, European markets closed higher, with the UK FTSE 100 rising by 54.88 points (0.60%), the French CAC40 increasing by 56.48 points (0.72%), and the German DAX gaining 247.01 points (1.03%) [3][4] - International oil prices increased, with WTI crude oil rising by $2.50 (3.75%) to $69.21 per barrel and Brent crude oil increasing by $2.47 (3.53%) to $72.51 per barrel [3][4] Economic Insights - The Chinese government reported that from January to June, state-owned enterprises generated total operating revenue of 40.75 trillion yuan, a year-on-year decrease of 0.2%, and total profits fell by 3.1% [7] - The Hong Kong Monetary Authority announced that institutions wishing to apply for stablecoin licenses can submit applications by September 30, 2025, as part of a new regulatory framework [8][9] Industry Developments - A seminar on the performance standards framework for integrated large models will be held on August 1, aimed at establishing standardized performance indicators and technical capabilities for the industry [10] - The China Photovoltaic Industry Association clarified rumors regarding the multi-crystalline silicon sector, emphasizing the need for accurate information and cautioning against misinformation [11]

Group 1 - The China Artificial Intelligence Industry Development Alliance is launching the development of performance standards for AI integrated machines to standardize performance indicators and technical capabilities [1] - The AI integrated machine includes high-performance computing hardware, pre-trained large AI models, and supporting software systems, designed for "plug-and-play" use, reducing the barriers to AI application [1] - The market for AI integrated machines is rapidly expanding, with over 100 manufacturers in China launching AI integrated machines, indicating a competitive landscape [1] Group 2 - Tianrongxin is advancing its intelligent computing cloud business by releasing a series of products including intelligent computing cloud platforms, integrated machines, and computing power servers, integrating multiple capabilities such as computing, storage, networking, security, and intelligence [2]

Core Viewpoint - Tianrongxin Technology Group Co., Ltd. (hereinafter referred to as "the company") expects significant growth in its financial performance for the first half of 2025, with projected revenue between 800 million to 850 million yuan and a net profit increase of 65.98% to 70.84% year-on-year [2] Financial Performance - The company anticipates a revenue increase of over 8% year-on-year for the second quarter [2] - The gross profit margin has improved by over 4 percentage points compared to the previous year [2] - Total expenses for sales, research and development, and management have decreased by over 13% year-on-year [2] Strategic Initiatives - The company is deepening its "AI + Security" strategy to capture the market for large model application security [2][3] - A significant upgrade to the Tianwen large model has been implemented, enhancing its intelligent defense capabilities [2] - The company has launched a comprehensive security protection framework for large models, covering the entire lifecycle of large model applications [3] Market Positioning - The company aims to become a leading provider of cybersecurity and intelligent computing cloud solutions in China over the next decade [3] - A partnership with Huawei has been established to develop an intelligent computing integrated machine based on the Kunpeng architecture [3] - The company has signed and is expected to sign orders exceeding 20 million yuan for the intelligent computing integrated machine, with annual sales projected to exceed 100 million yuan [4]

Core Viewpoint - Tianrongxin, a leading cybersecurity company, is expected to significantly reduce its losses in the first half of 2025, with projected revenues of 800 million to 850 million yuan and a net loss of 60 million to 70 million yuan, indicating a strong improvement in profitability [1] Group 1: Financial Performance - The company anticipates a year-on-year revenue growth of over 8% in the second quarter, suggesting it may have achieved profitability during this period [1] - Tianrongxin's gross margin has increased by over 4 percentage points year-on-year, while total sales, R&D, and management expenses have decreased by 13% compared to the previous year [1] Group 2: Strategic Initiatives - The company is advancing its "AI + Security" strategy, with significant upgrades to its Tianwen large model, transitioning from "rule-driven" to "intelligent decision-making" for cybersecurity defense [1] - Tianrongxin has successfully passed tests for its large model security gateway, receiving the first "Large Model Security Protection Fence Product Certification (Enhanced Level)" from the Ministry of Public Security [1] Group 3: Future Goals and Collaborations - The company's strategic goal for the next decade is to become a leading provider of cybersecurity and intelligent computing cloud solutions in China [2] - In March, Tianrongxin began a comprehensive partnership with Huawei to develop an intelligent computing integrated machine based on the Kunpeng architecture, aiming to support the digital transformation of critical industries in China [2] - The intelligent computing integrated machine is expected to generate over 100 million yuan in sales this year, with future cloud computing business goals targeting at least half the output of the security segment over the next 3-5 years [2] Group 4: Technological Developments - Tianrongxin is involved in research and application of technologies related to stablecoins, blockchain, zero trust, privacy computing, and commercial cryptography [2] - The company is participating in key infrastructure security projects for digital currency and cross-border payment in the financial industry [2]