Core Viewpoint - The article discusses the collaboration between 找钢集团 (Zhaogang Group) and 托克集团 (Trafigura Group) to establish a joint venture, 宗金汇 (Zongjin Hui), aimed at creating a comprehensive e-commerce platform for non-ferrous metals, leveraging AI technology and extensive industry resources [2][3][12]. Group 1: Joint Venture and Market Potential - 找钢集团 has signed a joint venture agreement with Lykos Holding Hong Kong Limited, investing 45 million yuan to establish 宗金汇, where it will hold an 8.57% stake [2]. - The collaboration aims to build a full-category e-commerce platform for non-ferrous metals, addressing the growing demand from customers who also have steel procurement needs [3][6]. - The non-ferrous metal market in China has surpassed 2 trillion yuan in 2023, with a projected production of 7.919 million tons for ten types of non-ferrous metals in 2024, reflecting a year-on-year growth of 3.4% [10]. Group 2: Technology and AI Integration - 找钢集团's AI infrastructure, developed from its steel trading platform, is expected to enhance cross-category capabilities, facilitating efficient transactions in the non-ferrous metals sector [7][8]. - The company has launched a B2B trading SaaS tool, "腾采通," which utilizes AI technology to analyze over 6.73 million business opportunities daily, covering more than 11 million SKUs and facilitating over 100 million transaction matches [7]. - The AI-driven solutions are designed to improve operational efficiency, with AI agents achieving over 95% accuracy in matching inventory and responding to quotes, significantly enhancing transaction efficiency [7]. Group 3: Industry Challenges and Solutions - The non-ferrous metals market faces challenges such as low online penetration (5% compared to over 20% for black metals), inefficiencies in traditional trading models, and high financing costs for small and medium enterprises [11]. - The joint venture aims to address these pain points by providing comprehensive services, including intelligent pricing, B2B payment solutions, and supply chain management, thereby enhancing operational efficiency for businesses in the sector [14][15]. - 宗金汇's development strategy includes leveraging托克's self-operated resources and integrating 找钢's infrastructure to enhance service capabilities for small clients and create a complete industry chain [16].

Core Viewpoint - Jingbeifang is a leading financial technology service provider in China, expected to experience business growth beyond expectations due to the dual drivers of the accelerated implementation of the digital RMB national strategy and the deep integration of AI technology in the financial industry [1] Group 1: Over-Expectation Logic - Market expectations suggest that the advancement of digital currency is slow, and AI's application in the financial sector is also lagging, leading to a belief that these factors will not contribute to short-term performance [2] - The company believes that the pace of policy implementation is accelerating, with the digital currency rollout speeding up, as evidenced by the recent introduction of the stablecoin legislation in Hong Kong [2] - The company has established deep collaborations with banking clients and possesses full-stack service capabilities for digital currency, positioning it to benefit from the promotion of stablecoins [2] Group 2: Driving Factors - The rollout of stablecoins presents business opportunities, with the Hong Kong Legislative Council passing the "Stablecoin Ordinance" on May 21, 2025, indicating a shift in stablecoins from being mere connectors of crypto assets to becoming a new infrastructure for global financial payments [3] - Domestic large models are showing significant breakthroughs in cost and capability, leading to accelerated adoption in the banking sector, which directly benefits the company through private deployment and AI application demands [2][3] Group 3: AI Application and Product Launch - Domestic large models like Deepseek are rapidly being adopted in the financial sector due to their efficient performance, cost-effectiveness, and open-source nature, with banks prioritizing AI for various applications such as intelligent customer service and risk control [3] - The company has officially launched an AI large model service platform and four major products in 2024, catering to the private deployment and AI application needs of major state-owned and joint-stock commercial banks [3] Group 4: Performance Metrics and Catalysts - Key performance indicators include the implementation of stablecoin policy details, AI business orders, and the proportion of AI revenue [4] - Catalysts for growth include the implementation of stablecoin regulations, approval of licenses for tech giants, and the emergence of popular AI products [5] Group 5: Profit Forecast and Valuation - The company forecasts revenues of 5.147 billion, 5.685 billion, and 6.291 billion yuan for 2025-2027, with year-on-year growth rates of 11.01%, 10.46%, and 10.67% respectively, and net profits of 363 million, 416 million, and 472 million yuan, with growth rates of 16.34%, 14.63%, and 13.54% [6] - The company is compared with peers in financial technology and banking IT, with a projected PE of 44 times for 2025, while the average PE of comparable companies is 67 times, indicating a potential target stock price of 28.11 yuan, representing an upside of 54.39% [6]

Core Insights - The report from Northeast Securities highlights that banks are expected to become pioneers in AI implementation in China due to ample IT budget, market-oriented systems, and high integration of internal data [1][3] - DeepSeek-R1's inference cost is only 1/30 of comparable products, marking a new phase of "AI popularization" in the industry [1] - The year 2025 is anticipated to be the starting point for AI Agents, with significant competition among major companies in this area [2] Group 1: AI Technology and Applications - DeepSeek has launched several well-known open-source models since its establishment in July 2023, with the DeepSeek-R1 model achieving performance comparable to OpenAI's o1 series at a significantly lower cost [1] - Major banks have actively integrated AI technology into various applications such as investment research, customer service, credit approval, and more, enhancing the intelligence of financial services [3] - IDC predicts that the banking sector will account for over 20% of global AI solution spending from 2024 to 2028 [3] Group 2: Specific Companies and Their AI Initiatives - Yuxin Technology has fully integrated DeepSeek models into its product system, focusing on applications in credit, data, and marketing channels [4] - Jingbeifang has launched an AI large model service platform and several intelligent assistants, achieving breakthroughs in smart fraud prevention and investment advisory across multiple industries [4] - Gaoweida has deeply integrated DeepSeek with its credit business, enhancing credit efficiency and financial report analysis through AI applications [4] - Tianyang Technology has released intelligent testing analysis systems and compliance models, providing banks with intelligent data analysis solutions [4] - Shenzhou Information has upgraded its financial knowledge Q&A and coding assistants, improving development efficiency by 20% and automating 30% of code generation [5]

随着新一代信息技术、人工智能（AI）、智能驾驶、影视渲染、生物医药、智慧城市等领域的创新， 高性能算力需求不断增长。发放训力券，支持创新主体用于租用智能算力开展AI大模型训练、推理。 这是深圳有效链接算力供需端，破解智算瓶颈，推动新兴产业发展的务实举措。 近年来，深圳密集出台政策支持AI产业发展，《深圳市加快推进人工智能终端产业发展行动计划 （2025-2026年）》提到，发放"训力券""语料券""模型券"，降低企业成本；《深圳市打造人工智能先锋 城市的若干措施》在丰富生态要素供给方面，提出每年发放最高5亿元训力券，降低AI模型研发和训练 成本。 今年3月，深圳启动了训力券服务机构入库申请和训力券申请工作。其中，服务机构的申请者要求是在 深圳市（含深汕特别合作区）依法经营、具有独立法人资格的智能算力服务机构，具备较强服务能力的 相关机构在深圳市设立的分支机构。 【深圳商报讯】（首席记者王海荣）2025年度深圳市训力券兑现申请已正式启动，5月8日，记者从深圳 市科技创新局获悉，符合条件的申请单位可从即日起至5月28日，登录深圳市科技业务管理系统在线申 请。 此次启动的训力券兑现工作中，申请单位应当是按照相关规定 ...