关注GPT 类技术落地场景
3 月22 日，百度文心一言专家与上海钢联进行了AIGC 领域深度合作的洽谈。我们认为公司依托数据积累、场景卡位，有望持续受益于GPT 类技术的落地。预计2022-2024 年EPS 分别为0.76/0.93/1.18 元。分部估值，预计2023 年产业数据服务收入9.32 亿元，可比公司平均23E 10.6xPS(Wind)，给予23E 10.6xPS；钢银电商对上市公司贡献净利23E 1.71 亿元，可比公司平均23E 27.4xPE(Wind)，给予23E 27.4xPE，目标价54.20 元，维持“买入”评级。
多环节应用AI 技术，具备丰富AI 技术落地经验公司积极使用AI 的方式赋能资讯生产流程及用户产品体验，具备良好的AI技术落地经验。在AI 技术提升数据收集质量方面，通过卫星遥感影像AI算法训练，完成对港口堆料的识别、分类、演算，实现全球铁矿石港口库存的监测，采集调研与卫星调研相互印证，通过大量样本数据标注训练算法，可实现近乎全样本的高频调研；在“钢联数据”终端产品中通过知识图谱、NLP 和深度学习等方法，完善终端内的搜索功能。公司在AI 领域具备良好的技术落地经验，有助于加速AI 新技术的落地。
GPT 类技术具备良好场景基础，新技术或推动收入加速增长公司业务覆盖多个知识梳理、数据分析、营销推广类场景，在产业数据服务板块中，有数据订阅、商务推广、会务培训、研究咨询等业务，涉及AIGC、智能搜索、智能营销、智能客服多个应用场景。公司在资讯、搜索、报告制作工具（Mreport）、大数据预测系统和客服等功能模块及场景上尝试用GPT类的AI 技术，提升客户体验。公司具备良好的GPT 类技术落地场景，我们认为随着GPT 类技术的逐步应用，有望进一步为公司相关产品服务收入提供增长动力。
具备数据积累+场景卡位优势，关注GPT 类技术应用落地我们认为嵌入产品、场景是GPT 类技术重要的落地方式，GPT 类模型在基于内容理解的数据、知识梳理以及内容生成领域展现出较强的能力，具备数据、知识积累及丰富交互场景的厂商有望首先受益。上海钢联作为大宗商品行业的数据资讯服务商，形成了围绕价格波动的多维度数据积累，此外公司也具备丰富的内容、知识生成、营销推广场景，我们认为公司有望持续受益于GPT 类技术应用的落地。
风险提示：产业数据服务拓展低于预期风险；钢银电商交易量下降的风险。
Focus on GPT technology implementation scenarios
On March 22, experts from Wenxin Baidu conducted in-depth cooperation negotiations with the Shanghai Steel Federation in the field of AIGC. We believe that the company relies on data accumulation and scenario cards, and is expected to continue to benefit from the implementation of GPT technology. EPS for 2022-2024 is expected to be 0.76/0.93/1.18 yuan, respectively. According to the segmental valuation, industry data service revenue is estimated at 932 million yuan in 2023, with an average of 23E 10.6xps (Wind) for comparable companies, giving 23E 10.6xps; steel and silver e-commerce contributed 23E 171 million yuan to listed companies, and an average of 23E 27.4xPE (Wind) for comparable companies, giving 23E 27.4xPE (Wind), with a target price of 54.20 yuan, maintaining the “buy” rating.
Applying AI technology in multiple links, companies with rich experience in implementing AI technology actively use AI methods to empower information production processes and user product experience, and have good experience in implementing AI technology. In terms of improving the quality of data collection with AI technology, AI algorithm training on satellite remote sensing images completes identification, classification and calculation of port material, enables monitoring of global iron ore port inventories, and mutual confirmation between collection and research. Through the training algorithm for labeling a large number of sample data, high-frequency research of nearly all samples can be realized; in the “Steel Link Data” terminal product, the search function within the terminal is improved through knowledge maps, NLP and deep learning methods. The company has good technology implementation experience in the field of AI, which helps accelerate the implementation of new AI technology.
GPT technology has a good scenario foundation. New technology may accelerate revenue growth. The company's business covers multiple knowledge sorting, data analysis, and marketing promotion scenarios. In the industrial data service sector, there are services such as data subscription, business promotion, conference training, research and consulting, etc., involving multiple application scenarios for AIGC, intelligent search, intelligent marketing, and intelligent customer service. The company is experimenting with GPT AI technology in functional modules and scenarios such as information, search, report production tools (Mreport), big data prediction systems, and customer service to enhance the customer experience. The company has a good implementation scenario for GPT technology. We believe that with the gradual application of GPT technology, it is expected that it will further provide growth impetus for the company's related product and service revenue.
It has the advantages of data accumulation+scenario cards. Focus on the application implementation of GPT technology. We believe that embedded products and scenarios are an important implementation pattern for GPT technology. GPT models have shown strong capabilities in the fields of data, knowledge sorting and content generation based on content understanding. Vendors with data and knowledge accumulation and rich interactive scenarios are expected to benefit first. As a data and information service provider for the commodity industry, Shanghai Steel Federation has formed multi-dimensional data accumulation surrounding price fluctuations. Furthermore, the company also has rich content, knowledge generation, and marketing promotion scenarios. We believe that the company is expected to continue to benefit from the implementation of GPT technology applications.
Risk warning: The expansion of industrial data services is lower than expected; the risk of a decline in e-commerce transaction volume for steel and silver.
