本公司监事会及全体监事保证公告内容不存在任何虚假记载、误导性陈述或者 重大遗漏，并对其内容的真实性、准确性和完整性承担个别及连带责任。 根据《上市公司行业信息披露指引第二号——房地产、第四号——电力、第五号—— 零售》的要求，辽宁申华控股股份有限公司（简称"公司"）现将 2025 年一季度主要经 营数据披露如下： 证券代码：600653 证券简称：申华控股 编号：临 2025—19 号 辽宁申华控股股份有限公司 2025 年一季度主要经营数据公告 - 1 - 1、汽车销售：2025 年 1-3 月，公司实现合并口径宝马品牌销售 2,394 辆，同比下降 11%。 近日，宝马集团公布了 2025 年第一季度的销量，数据显示中国市场 2025 年第一季度累计交 付 155,195 辆，同比下滑 17.2%。公司凭借高效的管理和优质的服务已尽力减少市场下滑带来 的不利影响，面对竞争加剧和电动化转型的局面，公司和宝马厂商均已开始做出调整。 2025 年，宝马将正式发布全新的 Neue Klasse 平台架构，同时计划推出 10 款重磅新车型。 除已在售的全新 BMW X3 长轴距版，全新 BMW M235L 四门轿跑 ...

Core Insights - ST Jin Yi (002721) reported a revenue of 372.74 million yuan and a net profit of 14.54 million yuan for the year 2024, indicating a stable transition post-restructuring [1] - The company plans to enhance its operational efficiency through segmented management, focusing on AI and large model technology development, and optimizing its gold and jewelry business [1] - The acquisition of Beijing Kaike Weishi Technology Co., Ltd. marks a strategic shift towards digital economy assets and aligns with the company's goal of transitioning from the gold and jewelry sector to software and information services [1][2] Financial Performance - In 2024, ST Jin Yi achieved an operating income of 372.74 million yuan and a net profit attributable to shareholders of 14.54 million yuan, reflecting a stable financial performance [1] Strategic Development - The company aims to improve its software information service capabilities, integrate retail operations, and enhance corporate governance and investor relations [1] - The acquisition of Kaike Weishi is a significant step in the company's strategy to seek transformation towards technology and digital economy sectors [1][2] Industry Context - Kaike Weishi is positioned as a key player in the financial sector's digital transformation, leveraging its expertise in large models and big data technology to create intelligent solutions [2] - The collaboration between ST Jin Yi and Kaike Weishi is expected to utilize state-owned assets and resources from Haidian District to facilitate the strategic transition [2]

Core Insights - The article highlights the significant transformation in various industries driven by artificial intelligence and large model technologies, with a focus on the initiatives taken by Huangpi Tobacco Bureau to leverage digitalization for high-quality development [1] Group 1: Digital Transformation Strategy - Huangpi Tobacco views digital transformation as a key strategy to overcome management bottlenecks and enhance service efficiency [2] - The training session emphasized the application potential of large model tools like DeepSeek in data analysis, market forecasting, and customer service [2] Group 2: Practical Application of AI - The training adopted a problem-oriented approach, demonstrating the operational logic of AI-assisted decision-making systems through simulations [4] - AI systems can automatically identify abnormal orders and predict compliance risks, significantly improving the efficiency of clue screening in market regulation [4] - An intelligent response module based on natural language processing can analyze retail inquiries in real-time, reducing response delays [4] Group 3: Long-term Empowerment and Ecosystem Building - The training serves as a starting point for deeper reforms within Huangpi Tobacco, with plans to integrate specialized resources for digital applications in market regulation [5] - The bureau aims to enhance internal management efficiency and improve the precision of market regulation and customer service through a dual approach of "technology + business" [5] - Future collaborations with universities and technology companies are planned to position Huangpi Tobacco as a benchmark for intelligent transformation in the industry [5]

Core Viewpoint - Jiangbolong (301308) reported a stable revenue of 4.256 billion yuan in Q1 2025, with a net loss of 152 million yuan, indicating an improvement trend in profitability despite slow terminal recovery and inventory digestion [1] Group 1: Financial Performance - The company achieved a revenue of 4.256 billion yuan in Q1 2025, with a net loss of 152 million yuan and a non-recurring net loss of 202 million yuan [1] - The net loss margin significantly narrowed, showing signs of profitability improvement [1] - Revenue from enterprise-level storage (eSSD and RDIMM) reached 319 million yuan, growing over 200% year-on-year [2] Group 2: Product Development and Market Position - Jiangbolong has made significant progress in self-developed main control chips, successfully mass-producing UFS4.1 chips and applying over 30 million self-developed chips [3] - The company has launched new products to meet the urgent demand for high-capacity SSDs and DRAM, enhancing its offerings in mobile and smart wearable markets [3] - Jiangbolong has established deep cooperation with major automotive manufacturers, covering over 10 types of in-vehicle applications, leveraging its early entry into the automotive storage sector [2] Group 3: Brand and Market Expansion - Jiangbolong is the second-largest independent storage company globally and the largest in China, with significant brand presence in both B2B and B2C markets [4] - The Lexar brand achieved a 21% year-on-year revenue growth in Q1 2025, continuing its expansion in the global consumer market [4] - The acquisition of Zilia, a leading storage manufacturer in Brazil, has enhanced Jiangbolong's product offerings and supply chain efficiency [4] Group 4: Strategic Outlook - The company is well-positioned for future growth due to its strategic focus on high-end storage, technological breakthroughs, and effective market expansion [5] - The ongoing demand for AI hardware and high-performance computing is expected to create further opportunities for Jiangbolong's high-end business [1]

Core Insights - The article discusses the launch of the "Digital Government Efficiency Micro-Record" column by the Southern Metropolis Big Data Research Institute, aimed at addressing gaps in government services for businesses and citizens, and enhancing communication with service departments [2] - Following a report on discrepancies between two intelligent Q&A systems on the Huangpu District government website, the Huangpu District Data Management Bureau responded by taking down one of the systems and implementing optimization measures [3][5] Group 1: System Discrepancies - The Huangpu District government website previously featured two intelligent Q&A systems, "Huangpu Government Service Model Intelligent Body" and "Kao Pu Answer," which provided conflicting answers to the same questions, causing confusion among users [3] - The "Intelligent Body," launched on April 8, 2025, is based on a large model and integrates government services, while "Kao Pu Answer," which was based on knowledge graph technology, was launched in November 2023 and has since been discontinued [3] Group 2: Policy Updates and User Experience - The Huangpu District Data Management Bureau acknowledged issues with the "Intelligent Body" referencing outdated policies and has committed to adding clear labels indicating when policies are in a "buffer period" despite being expired [6] - Enhancements to the user experience are planned, including allowing continuous input in the Q&A process, as current limitations were attributed to security considerations [7] Group 3: Future Directions - The Huangpu District Data Management Bureau will maintain human consultation services to complement the "Intelligent Body," ensuring that complex inquiries can be addressed effectively [8] - The bureau aims to continuously improve the large model application and update the policy knowledge base dynamically, responding to public feedback and media scrutiny [8]

大模型作为产业变革的核心引擎。通过RAG、Agent与多模态技术正在重塑AI与现实的交互边界。三者协同演进，不仅攻克了数据时效性、专业适配等核 心挑战，更推动行业从效率革新迈向业务重构。本文将解析技术演进脉络、实战经验与未来图景，为读者提供前沿趋势的全局视角与产业升级的实践指 引。 作者 | 蒋进 出品丨腾讯云开发者 大模型技术正加速渗透至产业核心场景，成为驱动数字化转型的智能引擎。全球机器学习大会（ML-Summit）聚焦大模型技术的创新突破与产业实 践，深入探讨其前沿方向与落地路径。作为AI发展的核心驱动力， 检索增强生成（RAG） 通过动态知识融合技术突破大模型的静态知识边界； 智能体 （Agent） 借助自主决策与多任务协同能力重构人机协作范式； 多模态大模型 则依托跨模态语义理解技术解锁复杂场景的落地潜力。三者协同演进， 不仅攻克了数据时效性、隐私安全与专业适配等关键难题，更在医疗诊断、金融风控、智能制造等领域催生从效率革新到业务重构的行业级变革。 ML-Summit会议大模型内容分布 RAG： 大模型的动态知识引擎，解决模型静态知识边界、时效性与可信度问题。 大模型在很多领域表现出色，但依然存在局 ...

Core Insights - JD Health released its 2024 Environmental, Social, and Governance (ESG) report, showcasing its sustainable development practices in the healthcare sector, emphasizing safety, efficiency, convenience, and warmth in health consumption experiences [1] Compliance and Governance - In 2024, JD Health established a comprehensive compliance system with a focus on user health, implementing a "three-line defense" and "five dimensions" approach to ensure compliance across all operations, including drug management and internet hospital services [3] - The company achieved ISO13485 certification for medical devices, expanding its compliance certifications to cover all business areas, including drug management and internet hospitals [3] - JD Health published standardized treatment paths for 43 diseases and established 4 core medical quality management systems, totaling 265 standardized treatment paths and 16 core systems over two years, enhancing the quality management of online medical services [3] Service Innovation - JD Health innovated its healthcare service model to meet diverse user needs, achieving an average of over 490,000 online consultations per day by the end of 2024 [4] - The company created a closed-loop service model integrating "doctor consultations," "home tests," "home delivery of medications," and "home nursing," allowing patients to access comprehensive healthcare services from home [4] - JD Health expanded its "JD Home Fast Testing" services to 12 cities, offering 149 home testing products, and launched nearly 40 home nursing projects through the "JD Nurse at Home" service [4] Technological Empowerment - JD Health accelerated the application of large model technology across various service scenarios, aiming to provide a seamless online and offline healthcare experience [7] - The company introduced several intelligent health services, including "KangKang" health assistant and "ChatGPT-like" AI psychological companion, enhancing the efficiency and accuracy of healthcare services [7] - JD Health's "Smart Doctor Assistant" became the most widely used intelligent diagnostic tool among doctors, improving clinical decision-making and patient management [7] Social Responsibility - JD Health deepened its corporate social responsibility initiatives, aiding 637 patients through its rare disease care program, with a total value of assistance reaching 6.24 million yuan [8] - The "Medical Heart Station" project, in collaboration with top medical institutions, provided support to over 23,000 patients, improving the healthcare experience for rare disease patients [8] - The company conducted 22 donation activities in 2024, donating over 1.53 million health items to rural communities and disaster relief efforts [10] - JD Health launched a "Expired Medicine Recycling" initiative, successfully collecting and professionally disposing of 11 tons of expired medications, engaging 57,000 users across 315 cities [10]

Core Viewpoint - The third "Wenxin Cup" entrepreneurship competition organized by Baidu will officially start on April 25, offering nearly 100 million RMB in cash and resources to support AI application innovation and implementation [1][3]. Group 1: Competition Overview - The competition aims to lower the barriers for AI application entrepreneurship by providing a combination of cash and resources, with a tiered incentive system for participants [3][4]. - Participants will receive a reward of tokens valued at thousands of RMB upon registration, facilitating rapid validation of product prototypes and business models [3][4]. - The previous two competitions attracted over 2,500 entrepreneurial teams, with awarded projects spanning various sectors including office, entertainment, e-commerce, marketing, finance, and healthcare, receiving more than 200 million RMB in funding support [4]. Group 2: Investment and Support - The prize structure includes 20 million RMB for first place, 10 million RMB for second place, 5 million RMB for third place, and a special award of up to 70 million RMB [6]. - Baidu emphasizes its commitment to supporting developers and entrepreneurs by providing models, development tools, and financial resources, aiming to foster a prosperous AI ecosystem [4][5]. Group 3: Future Directions - Baidu plans to leverage the "Wenxin Cup" as a starting point to enhance AI infrastructure and application implementation, promoting an open innovation approach and collaboration with developers to establish a Chinese paradigm for AI-native applications [4].

| | 本报告期 | 上年同期 | 本报告期比上年同期增减 | | | --- | --- | --- | --- | --- | | | | | (%) | | | 营业收入(元) | 748.187.189.04 | 618,822,422.15 | | 20.90% | | 归属于上市公司股东的净利 | 120,393,590.22 | 103.866.228.77 | | 15.91% | | 润(元) | | | | | | 归属于上市公司股东的扣除 非经常性损益的净利润 | 112.534.527.21 | 98.960.278.13 | | 13.72% | | (元) | | | | | | 经营活动产生的现金流量净 | 313.291.816.68 | -50.498.687.45 | | 720.40% | | 额(元) | | | | | | 基本每股收益(元/股) | 0.22 | 0.19 | | 15.79% | | 稀释每股收益(元/股) | 0.22 | 0.19 | | 15.79% | | 加权平均净资产收益率 | 1.50% | 1.41% | | 0.09% | | | ...

快速发展的AI大模型技术红利向机器人行业外溢，使机器人成为今年的热门赛道。 4月15日，第137届中国进出口商品交易会（广交会）在广州开幕。今年，广交会首次增设了服务机器人 专区，46家企业携明星产品亮相，机器狗、仿生猫"花活"不断，长出"四肢"和"大脑"的人形机器人更是 受到关注。 在广东，机器人的技术进度条正快速刷新：众擎机器人实现了全球首例平地起身前空翻，越疆科技实现 了全球首款灵巧操作+直膝行走，优必选人形机器人进入全球最多工厂……同时，新政策也层出不穷。 日前，《广东省推动人工智能与机器人产业创新发展若干政策措施》也提出，开展"机器人+"行动，深 入挖掘开放应用场景。 科技企业的集聚，体现了广东发展人形机器人的什么优势？这又将如何加速落地转化？ 围绕这些问题，南方财经全媒体记者专访了广东省机器人协会执行会长、中国机器人产业联盟理事任玉 桐。在他看来，广东是全球最大的机器人应用市场，重点是开放应用场景，让机器人企业与需求方建立 对接机制，以场景牵引技术落地。 关于未来，任玉桐很乐观，他预测，大约3-5年，人形机器人有望真正规模化落地，"且这一时间很可能 会提前，未来十年，机器人产业要超过汽车产业，广东 ...