New Stock Preview | Three years of accumulated losses of 2.4 billion, can "AI new star" DeepTech successfully pass through the 18C?

Recently, Di Pu Technology Co., Ltd. (referred to as Di Pu Technology) submitted an application for listing on the main board of the Hong Kong Stock Exchange. CITIC SEC, CMBC CAPITAL, Guotai Junan International, Puyin International, and BOCOM INTL are its joint sponsors. It is reported that Di Pu Technology has completed approximately 9 rounds of financing, with a total amount of about 2 billion yuan, including investors such as Gao Ling, IDG Capital, Shanghai AI Investment Fund, Xingye Asset Management, Guotai Junan, and BOCOM INTL. If the company successfully goes public, it is expected to become the first enterprise-level AI listed company on the Hong Kong Stock Exchange by 2025. At a time when domestic open-source large models represented by DeepSeek are hot, the companies heading for a Hong Kong IPO with the AI "halo" will undoubtedly attract a lot of market attention. However, from a performance perspective, the challenge of profitability may become a "roadblock" for Di Pu Technology. Three-year accumulated losses of 2.4 billion The prospectus shows that Di Pu Technology focuses on providing cutting-edge artificial intelligence solutions for enterprises, helping enterprises integrate data, decision-making, and operations on a large scale efficiently. The company has two main infrastructure, the AI-Ready FastData Foil data fusion platform and the Deepexi enterprise-level large model platform, to deploy and implement Agentic artificial intelligence applications in enterprises. From 2022 to 2024, the company's revenue continued to grow, reaching approximately 100 million yuan, 129 million yuan, and 243 million yuan respectively, with a compound annual growth rate of 55.5%. However, the company incurred losses of approximately 655 million yuan, 503 million yuan, and 1.255 billion yuan each year, accumulating losses of 2.4 billion yuan over three years. The substantial loss in 2024 is attributed to share payment expenses and the fair value change of shares with preferential rights. It is understood that Di Pu Technology commercializes its solutions based on transaction models, starting in 2019 and 2023 with the commercialization of the FastData and FastAGI solutions. In 2024, the revenue from the FastData enterprise-level data intelligence solution and the FastAGI enterprise-level artificial intelligence solution accounted for 62.8% and 37.2% of total revenue, respectively, with the latter growing rapidly. In the FastData enterprise-level data intelligence solution, using the company's AI-Ready FastData Foil data fusion platform, the standardization and unified governance of multi-modal data of enterprise customers (such as knowledge, documents, image sets, formulas) reduces data development work and costs significantly, while greatly improving data analysis efficiency and business response speed. In the FastAGI enterprise-level artificial intelligence solution, based on the company's Deepexi enterprise-level large model platform, it provides Agentic artificial intelligence applications for enterprise customers in various industries, including consumer retail, manufacturing, healthcare, and transportation. (Deepexi Product Logic Architecture) As an AI company, Di Pu's research and development expenses account for the majority of total expenditures. From 2022 to 2024, Di Pu Technology's research and development expenses were 94.2 million yuan, 82.3 million yuan, and 81.4 million yuan, accounting for 93.7%, 63.8%, and 33.5% of total revenue for each respective year. During the reporting period, the number of company customers significantly increased. As of the end of 2022, 2023, and 2024, Di Pu Technology had accumulated services for 129, 178, and 245 customers respectively. As of December 31, 2024, out of 245 enterprise users, 81 were repeat customers, accounting for 33.1% of the company's customer base. Financially, due to continuous resource investment and issuing shares with preferential rights, the company's current liabilities and total liabilities remained at high levels. The prospectus shows that as of the end of 2022, 2023, and 2024, the company's net current liabilities were 2.039 billion yuan, 2.449 billion yuan, and 3.687 billion yuan, and net liabilities were 2.013 billion yuan, 2.424 billion yuan, and 3.676 billion yuan respectively. The year AI applications land commercially, the industry still faces commercialization difficulties After decades of significant technological evolution, artificial intelligence has evolved from early machine learning-based discriminative artificial intelligence to today's large model artificial intelligence, with intelligent bodies, large models, and other generative AI applications appearing rapidly. Industry insiders point out that 2025 is expected to be the year when AI applications are commercially landed on a large scale, accelerating the digital transformation of enterprises by deeply integrating business scenarios to serve the real economy. The "Generative Artificial Intelligence Application Development Report (2024)" shows that the global generative artificial intelligence industry is booming, driving rapid growth in the industry's scale. As of June 2024, the user base of generative artificial intelligence products in China reached 230 million people, and by the end of 2024, a total of 309 generative artificial intelligence products were filed in China. As artificial intelligence transitions to commercialization and applications, the market for enterprise-level artificial intelligence solutions is also booming. According to Frost & Sullivan data, the market size of enterprise-level artificial intelligence application solutions has grown from 107 billion yuan in 2020 to 386 billion yuan in 2024, with a CAGR of 37.8% from 2020 to 2024. It is expected that the market size will reach 239.4 billion yuan in 2029, with a further increase in CAGR from 2024 to 2029 to 44.0%. However, from the return of the technological frenzy to the landing of scenarios, commercialization remains a difficult path for most AI companies. For example, in the "AI Four Little Dragons," iFlytek, a company that has not been successful in going public, reported a net loss of 2.477 billion yuan in the first half of 2024 and still has not achieved profitability; Yuncong Technology reported a net loss of 359 million yuan in the first half of 2024, a slight increase from the same period in the previous year. AI companies have a strong technological nature, which also means that substantial research and development funds are required over the long term. For small companies like Di Pu Technology, their ability to raise funds is a significant challenge.Compared to large equipment manufacturers and internet giants with the same emphasis on AI, it is clearly outclassed.From the perspective of revenue, the value of a company's AI business ultimately needs to be reflected in the actual returns it can bring to customers. In terms of TOC, according to the statistics of the top 50 AI application revenues released by well-known investment firm A16Z, the most profitable AI applications are mainly concentrated in sub-fields such as image/video editing, beauty cameras, and ChatBot; while in B2B, the popularity of areas such as autonomous driving, automated production, financial risk management, intelligent education, and intelligent medical diagnosis continues to remain high. According to DiPu Technology, the company's commercial landing cases currently include collaborating with a leading company in the footwear and clothing industry to build an AI training and promotion platform, realizing three major application scenarios: R&D design assistant (text and image), retail operation intelligent assistant (frontline-oriented), and retail supply chain assistant (management-oriented); collaborating with a leading company in the domestic engineering design service industry to realize two major application scenarios: a Q&A assistant and a design assistant; and collaborating with a well-known commercial real estate brand to create a smart business platform, forming a product combination of "DM DiPu Intelligent Marketing Platform + DC DiPu Omni-channel Transaction Platform + DO DiPu Intelligent Operation Analysis Platform." The IPO prospectus shows that DiPu Technology has already achieved scale commercial landing in multiple vertical industries, including consumer retail, manufacturing, medical, and transportation. With the core capabilities of the FastData real-time smart lake warehouse platform and the Deepexi DiPu enterprise large model, as well as the rich landing practice results in industrial scenarios, DiPu Technology has been selected in the IDC "China Data Intelligence Market Ecological Map V5.0" in the areas of data storage, data management and analysis, and industrial industry applications. In IDC's report on "Data Intelligence Market Trends Analysis," DiPu Technology is also included in the "Data Intelligence Core Technology Trends and Representative Manufacturers Chart V2.0," and has been selected as a representative manufacturer in the Product Building Blocks category. According to data, the competition in the Chinese enterprise large model artificial intelligence application solutions market is relatively concentrated, with the top five providers accounting for 39.1% of the market share by 2024 revenue. Among them, DiPu Technology ranks fifth, with a market share of 4.2%, expected to reach 4.4% by 2025. Overall, as a "small but beautiful" AI company, if DiPu Technology can successfully go public, its stock price is expected to benefit from AI favorability catalysts and achieve significant gains. However, from a long-term perspective, with the company's current profit situation and business scale not having a clear advantage in the increasingly competitive market, DiPu Technology may need to tell a more compelling story to firmly establish its position in the market.