Adhere to the “digital intelligence+circuit” development strategy and accelerate the empowerment of digital intelligence in all walks of life. The company was founded in February 1993. It is the founder of Chinese full-text search technology and a leading provider of artificial intelligence, big data and data security products and services. Since its establishment, the company has always insisted on independent research and development of core technology, and has been leading the original innovation and application scenario implementation of full-text search, search engine databases, and natural language processing (NLP) technology. According to different industry applications, it can be divided into five sections: digital government, integrated media, financial technology, digital enterprises, and public safety. The main products and services mainly include big data software products and services, artificial intelligence software products and services, security products, system integration and others. The company always adheres to the “digital intelligence+circuit” development strategy, creates rich industry solutions with excellent technological capabilities, and empowers enterprises and governments to achieve digital upgrading.

Overall 2023 results were under pressure, and performance declined due to reduced customer investment in the industry and increased expenses during the period.

In terms of performance, in 2023, the company achieved revenue of 782 million yuan/-13.84%, mainly affected by factors such as delays in the progress of the original projects and delays in procurement of new projects for downstream government-related industry users of the company's big data software products and services, which led to a certain decrease in revenue contributed by related users compared to the previous year; net profit to mother of 36 million yuan/-71.45% year on year, net profit not returned to mother of 0.2 billion yuan/ -96.97%. On a quarterly basis, Q1-Q4 companies achieved revenue of 2.11/2.34/1.57/ 180 million yuan, respectively, +10.62%/-0.31%/-25.94/ -33.32%, and net profit to mother of 0.3/0.39/ -0.47/ 0.14 billion yuan, respectively, +43.76%/-26.44%/-260.66%/-43.2%. By industry, the financial business achieved revenue of 313 million yuan/-9.79%; the public security and security business achieved revenue of 199 million yuan/YoY -22.40%; the government and public institution business achieved revenue of 171 million yuan/YoY -12.85%; and the integrated media business achieved revenue of 0.98 million yuan/-8.23% YoY. With the continuous introduction of new products related to artificial intelligence, such as the company's big models, and the horizontal expansion of the industry, the company's performance is expected to gradually recover and rebound. By the end of 2023, the company's current order amount was 543 million yuan, an increase of 14.485 million yuan over the end of 2022, and the amount of orders in hand at the end of the period began to rise year-on-year.

In terms of profitability, ① in terms of gross margin, the company's overall gross profit margin in 2023 was 68.37% /year on year +4.66 pct, mainly due to an increase of +6.61 pct in the gross margin of the main business excluding finance media, government and public institutions: 64.51% /year on year +4.19pct, finance media: 64.74% /year on year -2.65pct); ② in terms of rate, the company's overall expense ratio in 2023 was 64.51% /year over year +14.04 pct, sales expenses ratio 19.89% /year on year +4.4pct, mainly the company strengthened Sales and management layout; management expense ratio 25.91% /year over year +8.87pct, the main reason is that the company strengthened and improved management, added employee remuneration, increased depreciation of new office buildings, etc.; financial expenses ratio was -0.1% /year over year -4 pct; R&D expenses rate was 18.81% /year over year +4.77pct, mainly due to the company's comprehensive efforts in model development and industry implementation, increasing R&D investment in artificial intelligence and data platforms; ③ In terms of net interest rate, the company's net interest rate in 2023 was 4.27% /y-9.34pct.

Large models build new productivity and accelerate the implementation of AI+ vertical applications. On the one hand, since the release of the “Piotanda Model” at the end of June 2023, the company has successively launched major models applicable to the finance, media, government affairs, public opinion and other industries, gradually exploring and implementing application scenarios in the finance, media, government affairs, public safety and other industries, promoting the industry's transformation from “+AI” to “AI+”, creating new quality productivity, and activating greater “smart potential”. On the other hand, the company has accumulated a deep “financial+media+government” customer resource base, covering 72% of central media, 80% of central and State Council institutions, 5 major state-owned banks, 3 major policy banks, the Ministry of Public Security, etc. The vertical industry model is progressing smoothly and has landed on many benchmark customers. It has signed more than 20 contracts in financial and other industries, with a total contract amount of about 50 million yuan. During the reporting period, the Tuotianda model focused on advantageous industry applications, integrated the company's comprehensive R&D strength and continuous innovation ability in key technologies over the years, used its own high-quality data for pre-training, and carried out a series of iterative innovations in response to the implementation needs of segmented field scenarios, which is expected to contribute to the company's additional performance.

With a high-quality A1 corpus, the vector database digs deep into the “gold mine of multi-modal data”. As the “fuel” for AI development, data provides an important impetus for the development of the industry. T Ceneri Wenfeng quantitative data base: As an industry-leading provider of artificial intelligence, big data and data security products and services, since the construction of big data centers in 2010, Tolls has accumulated operable big data resources with high commercial value; collected and cleaned millions of collection sites, further expanded investment in collection points and data labeling services to ensure that the collection sources are highly authoritative and trustworthy; it has the highest scale in the industry and has maintained rapid growth with an average update rate of more than 500 million pieces per day . ② Selected for the High Quality Artificial Intelligence Model Corpus: On March 29, the first batch of 100 high-quality data sets of the AI Big Model Corpus was led by the Beijing International Big Data Exchange and the Beijing Artificial Intelligence Industry Alliance and jointly released by enterprises and institutions. Tolls was selected as a co-builder to further explore the potential and value of high-quality data, provide more accurate and reliable data services for big model training, and have provided high-quality Chinese pre-training data sets for Xiaomi, YunLChuan, and ?$#@$. ③ Intelligent data infrastructure -- Haibei vector database: Along with the surge in enterprise multi-modal data, in 2024, the company launched the Haibei vector database, which is driven by “cross-modal data retrieval + enterprise-level RAG” dual engine, which solves the processing methods of “using and not using” many enterprises, turning “data baggage” into a “data gold mine”; through changes in data retrieval methods, “humans adapt to machines” and become a new infrastructure for enterprise data value mining, widely serving secure big data, media big data, and government big data The army is big Many industry segments, such as data, have been selected by many government clients such as the Ministry of Public Security, Xinhua News Agency, the General Administration of Market Supervision and Administration, the General Administration of Customs, the Patent Office, and the Trademark Office.

Refine data asset management to build a “one-stop” AI big model industry implementation. Based on the Tuotian model and AIAgent, the company implemented functional iterations and launched four major data asset platforms: Digital Star Industrial Brain (enterprise service), several information platforms (media), Internet inspection big data platform (public opinion), and Tianmu open source intelligence platform (security) to gradually monetize data assets from resource to operational. We believe that the company can rely on massive high-quality data resources and rely on the “1+1=N” data element business model to help users improve experience, control risk, reduce costs and increase efficiency, seize a larger market share of data services, and continue to accelerate data authorization, registration, and circulation to maximize the value release of data resources.

Investment suggestion: As a leader in semantic intelligence in China, the company relies on the six major advantages of “high quality data set+intelligent data infrastructure+self-developed AI model+data security and controllability+efficient AI engineering+deep customer base” to continue to focus on the scenario-based application and industry implementation of big models. In the future, along with the deep integration of AI and industry and business expectations, the company is expected to take advantage of the AI industry's growth dividends and set sail to build a new growth curve. The company's net profit for 2024-2026 is estimated to be 236/345/424 million yuan, respectively, and EPS is 0.30/0.43/0.53 yuan, respectively. Corresponding to the current PE of 48/33/27 times, respectively, this is the first coverage, giving it a “buy” rating.

Risk warning: risk of policy changes, risk of increased industry competition, new projects falling short of expectations