China's leading AI model training data provider. Founded in 2005, the company specializes in specialized processing of natural data sources into engineering data sets for AI training. It can be described as the "gas station" of AI algorithm, which has covered Microsoft Corp, Tencent, Baidu, Inc., iFLYTEK and other head AI enterprise customers. The company's core team has academic and industry background, laying a solid foundation for the company's long-term development. Despite the short-term disturbance caused by the epidemic, the company's overall revenue and profit continued to grow rapidly. In 2016-2020, the operating income CAGR was 29.02%, and the parent net profit CAGR was 68.06%. The proportion of income from highly standardized database products continued to rise to 56% of 21Q1, and the gross profit margin remained stable at about 70% in recent years.

The commercialization of AI opens up tens of billions of space in the data service market, and algorithm iteration and scene innovation drive the growth of the industry, which is expected to improve market concentration. 1) AI has entered a stage of omni-directional commercialization, with a global industry of more than US $50 billion and a domestic market of more than 40 billion yuan. China's AI basic data service market is in a period of rapid growth and is expected to exceed 10 billion in 2025. 2) the AI basic data service is connected with the data producer and the AI application side, which is an indispensable part of the landing process of AI. We believe that the natural demand for iteration of machine learning algorithms and the innovation of application scenarios are the main driving forces for the expansion of the AI basic data service industry. On the one hand, specialized training data is the rigid requirement of the current AI algorithm development iteration; on the other hand, the deep learning model puts forward higher requirements for the data volume, diversity and update speed of the training data, which requires professional data providers to provide services. In addition, new scenarios such as intelligent Internet of things and industrial Internet give rise to new incremental demand. 3) at present, the industry concentration of the AI data service market is low, and the CR5 is only 26.2% in 2019. In the future, as the market demand changes to refinement, we expect the market concentration to further increase.

The foundation of the integrated platform is algorithms, resources and experience, data accumulation brings economies of scale and stickiness, and multi-domain and multilingual ability opens the space for growth. 1) based on the core algorithm, the company creates an integrated platform "from design to quality inspection", empowers the whole process of data production, efficiently realizes the quality inspection of large-scale data processing, and forms the core technical barrier. We believe that behind the competitive advantage is technology, resources and understanding of vertical industries. 2) the closed loop of "data accumulation-technology iteration-improving data quality" brings obvious scale advantages, the understanding of vertical industry Know-how broadens the company's competitive moat, and the high repurchase rate of head customers also shows the high stickiness of products, which further drives the company's understanding of industry Know-how. 3) the product capability covers three major areas: intelligent voice, computer vision and natural language, and the number of data products with its own training set ranks first in the industry. Although it is disturbed by the epidemic in the short term, we believe that the ability of multi-domain and multilingual coverage will open up the company's global growth space.

Covered for the first time, given a "buy" rating, with a target market capitalization of 7.2 billion yuan in 2022. We forecast that the company's income from 2021 to 2023 will be 282x39999,547 million, corresponding to the parent net profit of 0.84 billion; the horizontal comparison of AI's average PE in 2021 is about 61 times, taking into account the scarcity of the company in the AI data training industry, give the 2022 target market capitalization of about 7.2 billion yuan, given a "buy" rating.

Risk tips: the decline of the prosperity of the AI industry; the intensification of market competition; the risk of overseas trade friction; the risk of the deterioration of the epidemic situation of COVID-19.