Event Overview:
On the afternoon of December 16, Shanghai Steel Federation officially released the first major vertical language model for the commodity industry: the “Steel Union Master” big language model and the application of the first big industry model: “Xiaogang” digital intelligent assistant.
The first large-scale vertical commodity model was released. “Little Steel” empowers cost reduction and efficiency. From December 14 to 17, 2023, the “2024 Major Changes · New Momentum, China Steel Market Outlook and 'My Steel' Annual Meeting” was held in Shanghai. On the afternoon of December 16, the Shanghai Steel Federation officially released the “Steel Link Master” commodity industry vertical language model and “Xiaogang” digital intelligent assistant, which aims to solve many pain points in commodity information services, such as traditional human-computer interaction methods and low information acquisition efficiency; too much industry information, low knowledge acquisition efficiency; many text and knowledge work, low knowledge production efficiency; large general model costs and high thresholds. It is difficult to provide high-value services in vertical industries. What the industry needs is low cost and replicable big model technology. Since December 2019, Shanghai Steel Union has been researching data and information services. Since December 2019, it has formulated a “three-step” industry model strategic development plan: the first step is to solve the problems of internal cost reduction, efficiency, and increment; the second step is to complete the commercialization and product modeling of the big model, so that the industry can use big model technology at low cost; the third step is to go global, empower industry users, and help users customize the big model or small model needed for enterprise development.
The company completed the development of the 1.0 version of the “Steel Link Master” commodity industry vertical language model in a year by pre-training a large amount of high-quality commodity industry materials, command fine-tuning, and enhanced alignment based on the basic model, and successfully developed the first application of the commodity industry's first vertical language model in the commodity industry based on trade, analysis, consultation, and content generation in the commodity industry, based on the ability of “Steel Link Master” -- the “Xiaogang” digital intelligence assistant. “Xiaogang” Assistant has various capabilities such as price and data queries, industry encyclopedia questions and answers, information collation and content aggregation, knowledge content generation, report interpretation, analysis, and prediction. It interacts with users through human-computer dialogue, streamlines traditional complicated and inefficient steps of information acquisition, content analysis, and knowledge generation, and assists industry users to complete their daily tasks more efficiently.
Data asset entry is expected to benefit from optimized financial statements. The “Interim Provisions on Accounting Treatment Related to Enterprise Data Resources” now apply to data resources that enterprises recognize as intangible assets or asset classes such as inventory in accordance with the relevant regulations of enterprise accounting standards, as well as data resources legally owned or controlled by enterprises that are expected to bring economic benefits to the enterprise, but are not recognized as assets because they do not meet the asset confirmation conditions related to the enterprise accounting standards. The interim regulation came into effect on January 1, 2024. Before the regulation was implemented, the company's data resource-related expenses had already been included in profit and loss; after implementation, expenses related to the production and processing of the company's data products will be included in relevant assets, which is expected to optimize the operating results shown in the company's financial reports and more fully disclose data asset information.
Under the framework of existing accounting theory and accounting standards, data resources used within an enterprise and data resources for foreign transactions are all expensed when related costs are incurred. When data assets are listed, the costs associated with data assets are capitalized to form data-related assets such as intangible assets and inventory.
According to the understanding of relevant policies, there are two prerequisites for data asset listing: one is that the enterprise has the right to use and control the data resources, and the other is that it is likely to earn profits in the future.
According to the investor relations activity record sheet, the company has formed diversified and commercialized data products. In the first half of 2023, data subscription revenue was 220 million yuan, and research consulting revenue was 33.02 million yuan. In the future, when data assets are listed, the company sees no substantial barriers from the methodological and practical perspective of using cost measurement for initial measurement and subsequent measurement.
Investment advice
Shanghai Steel Union is one of the world's leading data service providers for commodities and related industries, and has a domestic e-commerce platform for smart B2B services for B2B steel transactions. On the one hand, it continues to expand the influence of the steel and silver e-commerce platform and increase platform settlement volume, and on the other hand, it continues to promote the industrial data business replication strategy, further opening up market space through industrial chain fission. We expect the company's operating income for 2023-2025 to be 879.63/1011.83/116.431 billion yuan, respectively, and net profit of 220/3.25/421 million yuan respectively, maintaining a buy-A investment rating. The target price for 6 months is 35.37 yuan, which is equivalent to a dynamic price-earnings ratio of 35 times that of 2024.
Risk warning:
The risk of fluctuations in the steel industry; the risk of credit management in e-commerce platforms.