新华财经北京3月28日电 中国建设银行首席信息官金磐石28日在该行2024年年度业绩发布会上表示， DeepSeek系列大语言模型发布后，该行第一时间用高质量文本数据进行微调，形成基于DeepSeek-R1的 推理类金融大模型，并于今年2月在生产环境完成私有化部署，赋能全集团的应用场景。 "截至目前，我行的金融大模型应用已经覆盖全集团一半以上的员工，46个业务领域，200多个场 景。"金磐石表示。 金磐石进一步举例说，建行的金融大模型已经应用于客户经营管理领域的工单生成，信用风险管理领域 的客户调查报告自动生成，支付结算领域的报文智能翻译，托管领域的基金分红信息抽取，还有IT研发 领域的代码检查等，大幅提高了员工的工作效率和工作质量，并有效控制了部分领域风险。 建行28日发布2024年年报显示，截至2024年末，该行已完成金融大模型的迭代更新16次，金融大模型通 用能力评分和业务场景能力评分明显提升。2024年上线168个金融大模型应用场景，覆盖集团约一半员 工。 此外，截至2024年末，建行算力规模507.72PFlops，较上年增长9.58%，其中图形处理器（GPU）等新 型算力占比超23.39%，整体算 ...

Investment Rating - The report does not explicitly state an investment rating for the financial large model industry Core Insights - The financial large model industry is characterized by the integration of AI technologies to enhance decision-making accuracy and operational efficiency in financial institutions [3][11] - Key players in the industry include Alibaba Cloud, Baidu Intelligent Cloud, Huawei Cloud, iFlytek, and Volcano Engine, each offering unique strengths and solutions tailored to financial applications [13][18][23][26][27] Summary by Sections Financial Large Models - Financial large models are large language models applied in the financial sector, designed to analyze financial data and predict market trends, thereby improving decision-making precision and efficiency [3] Competitive Analysis of Companies - **Alibaba Cloud**: Offers a robust technology platform and comprehensive solutions, focusing on data security and compliance, catering to various financial institution sizes [15][16] - **Baidu Intelligent Cloud**: Provides a customizable model-building capability through its Qianfan platform, significantly reducing technical costs and enhancing business innovation [19][20] - **Huawei Cloud**: Utilizes self-developed Ascend AI processors and Kunpeng servers to deliver efficient computing power, meeting the demands of complex model training and data processing [23][24] - **iFlytek**: Emphasizes self-controlled technology and deep integration with industry applications, providing efficient and secure solutions while promoting AI technology in finance [26] - **Volcano Engine**: Implements a model-application-data flywheel mechanism, ensuring tight integration of technology with business scenarios, and offers flexible service systems to meet diverse financial institution needs [27]

Investment Rating - The report maintains an "Accumulate" rating for the company with a target price of 30.80 CNY per share, up from the previous forecast of 28.61 CNY [2][15]. Core Views - The report highlights that active trading in the fourth quarter has driven the company's performance growth, with a continued optimistic outlook for the company's earnings due to high market trading volumes and the integration of AI capabilities to enhance competitiveness [3][15]. Financial Summary - For 2023, the company reported a revenue of 11,081 million CNY, which is expected to increase to 11,604 million CNY in 2024, reflecting a growth of 4.7%. By 2025, revenue is projected to reach 13,743 million CNY, marking an 18.4% increase [6][16]. - The net profit attributable to the parent company for 2023 was 8,193 million CNY, anticipated to rise to 9,610 million CNY in 2024, representing a 17.3% growth. By 2025, net profit is expected to reach 12,213 million CNY, a 27.1% increase [6][16]. - The earnings per share (EPS) is projected to grow from 0.52 CNY in 2023 to 0.61 CNY in 2024, and further to 0.77 CNY in 2025 [6][16]. Market Data - The company's total market capitalization is reported at 385,641 million CNY, with a current share price of 24.43 CNY [7]. - The stock has traded within a range of 9.96 CNY to 30.00 CNY over the past 52 weeks [7]. Investment Drivers - The report identifies active trading in the capital markets and a sustained demand for wealth management among residents as key catalysts for growth [5][15]. - The company’s commission income is expected to increase by 23.07% to 61.13 million CNY in 2024, contributing significantly to revenue growth [15]. Future Outlook - The report anticipates that the active trading environment will continue to support the company's performance, with daily average stock trading volumes reaching 1.56 trillion CNY, a year-on-year increase of 80% [15]. - The integration of the "Miaoxiang" financial AI model is expected to enhance the company's competitive edge and open up new growth opportunities [15].

Core Viewpoint - The financial industry is systematically embracing AI, with financial large models evolving from efficiency tools to strategic engines, poised to reshape the entire lifecycle of financial services and drive future business growth [1][9]. Group 1: AI and Financial Industry Transformation - The global financial sector is undergoing a significant transformation driven by AI and big data, marking the onset of the "fourth industrial revolution" [2]. - The application of large models in finance is expanding, transitioning from cautious observation to widespread adoption, particularly with the emergence of high-quality, low-cost open-source models like DeepSeek [2][3]. - Financial institutions are increasingly focusing on how to deepen the application of large models, moving from basic scenarios to more complex decision-making contexts such as investment advisory and trading [3]. Group 2: Challenges in AI Implementation - Financial institutions face three main challenges in implementing large models: improving model accuracy, ensuring sustainable technology investment, and addressing data and privacy protection issues [1][10][11]. - The accuracy of large models is critical in finance, where precision and interpretability are paramount, especially in complex scenarios [10]. - The exponential growth in computational demands for model training and inference poses a significant financial burden on institutions, necessitating a balance between cost and value creation [11]. Group 3: Innovations by Dongfang Wealth - Dongfang Wealth has introduced the next-generation intelligent financial terminal, "Miaoxiang Investment Research Assistant," aimed at transforming the investment research process from a multi-tool operation to a single assistant model [6]. - The company is enhancing model capabilities to ensure more reliable and transparent financial AI, including the introduction of a tiered information source mechanism to improve the credibility and interpretability of responses [7]. - Dongfang Wealth is actively collaborating with various stakeholders to establish industry standards for the application of financial large models, contributing to the safe and robust development of AI in finance [8]. Group 4: Future Directions in Financial Services - The integration of AI in financial services is expected to enhance customer service by shifting from standardized responses to personalized interactions, thereby expanding the boundaries of service offerings [9]. - The proactive risk management capabilities of large models allow for a shift from reactive analysis to real-time monitoring and early warning systems, improving the overall risk control framework [9]. - The financial sector is encouraged to adopt a customer-centric approach, leveraging AI and big data to create differentiated services and enhance investor engagement [12].