Since 2026, with the accelerated penetration of AI technology across all scenarios, coupled with stimuli from major players like ByteDance and Alibaba accelerating the commercialization of AI, AI applications have rapidly become a core theme in the capital markets. Sub-sectors such as AI+marketing, AI+healthcare, and AI+pharmaceuticals have taken turns performing strongly. Leading companies including Bluefocus Intelligent Communications Group Co.,Ltd. (300058.SZ), Easy Click Worldwide Network Technology Co.,Ltd. (301171.SZ), Jiangsu Eazytec Co.,Ltd. (688258.SH), and Jiangxi Geto New Materials Corporation Limited (300986.SZ) have seen their stock prices double within just ten trading days, emerging as the brightest stars in this market rally. This surge in AI application stocks is no coincidence: following three years of capital frenzy in the AI hardware industry chain, the market's core expectation for 2026 has shifted from the "computing infrastructure race" to the "realization of application deployment." This transition in investment logic, from "hardware euphoria" to "application rise," has been foreshadowed in the market, awaiting only the dual catalysts of performance and practical scenario implementation to fully emerge.
Amid the industrial wave where AI applications are poised to become a main theme in capital markets, companies are actively seizing the window of opportunity for "application deployment realization" to accelerate their capitalization processes. Among them is Beijing Deep Intelligence Technology Co., Ltd. (hereinafter referred to as Deep Intelligence), which has initiated a listing application in Hong Kong. It is understood that Deep Intelligence submitted its second application for a main board listing on the Hong Kong Stock Exchange on December 1, 2025, with ICBC International as the sole sponsor. Notably, the company had initially submitted its application on May 28, 2025. According to Frost & Sullivan data, as a decision-type AI technology enterprise, Deep Intelligence ranked first in China's marketing and sales decision AI application market based on 2024 revenue, holding a market share of 2.6%. Clearly, in this specific niche of marketing and sales decision-making, Deep Intelligence has established itself as an industry leader.
However, financial reports reveal significant volatility in Deep Intelligence's performance. The prospectus shows that Deep Intelligence's revenues for 2022 to 2024 were approximately RMB 543 million, RMB 611 million, and RMB 538 million, respectively. Adjusted net profits for the same periods were approximately RMB 59.362 million, RMB 70.814 million, and RMB 21.52 million. In the first half of 2025, Deep Intelligence's revenue was RMB 277 million, a year-on-year increase of 5.73%, while its adjusted net profit was RMB 13.367 million, surging by 760.72%. Why does Deep Intelligence, as an industry leader, exhibit such significant performance volatility? What risks and challenges does its business operations face? Driven by ongoing AI industry trends, what valuation might the capital market assign to it? These are core issues that investors need to penetrate beyond the surface and analyze in depth.
An imbalanced business structure causes revenue to fluctuate with macroeconomic changes. Reviewing Deep Intelligence's capitalization journey reveals a rocky path. In 2016, the company first attempted capitalization by applying for listing on the New Third Board, which was accepted. However, due to issues like insufficient liquidity and low valuation benchmarks in the New Third Board market at the time, the company ultimately voluntarily withdrew its application. In 2022, Deep Intelligence shifted its focus to a charge on the A-share ChiNext board. Yet, during the review process, it faced three consecutive rounds of inquiries. After nearly two years of back-and-forth, it finally voluntarily withdrew its listing application in June 2024, ending its first ChiNext attempt in failure. Now, will Deep Intelligence succeed in its shift to the Hong Kong market?
With a history tracing back to 2009, Deep Intelligence was one of the earliest companies in China to apply AI technology to the digital transformation of enterprise marketing and sales. As early as 2011, Deep Intelligence developed its proprietary AI decision-making platform for advertising placement, AlphaDesk, specifically designed for digital advertising scenarios involving large data volumes and real-time execution. By 2017, Deep Intelligence launched its proprietary enterprise customer relationship management AI decision-making platform, AlphaData, aimed at enhancing how enterprises manage their own customer data and extract value from it. AlphaData enables enterprises to implement data-driven, personalized lifecycle strategies, thereby driving end-user growth, improving retention rates, and enhancing long-term end-user value. Following the global sweep of the generative AI trend, Deep Intelligence launched its enterprise AI agent system, Deep Agent, in February 2025. This system consists of a series of agent products that utilize agent technology and industry-specific proprietary knowledge to perform tasks, addressing various marketing and sales scenarios. Thus, Deep Intelligence's business layout now encompasses two decision-type AI platforms—advertising placement and enterprise CRM—alongside the generative AI platform of intelligent agents, forming a closed-loop deployment concerning AI technology.
However, looking at the revenue structure, Deep Intelligence's current income primarily comes from two sources: intelligent advertising placement and intelligent data management. In 2024, these accounted for 84.6% and 14.5% of revenue, respectively, with intelligent advertising placement being the core business. Revenue from intelligent agents remains relatively small. As of November 25, 2025, Deep Intelligence had signed 17 contracts related to Deep Agent, with a total contract value of only about RMB 11.9 million. From a client perspective, the prospectus indicates that since its establishment, Deep Intelligence has served approximately 530 end clients from various industries, including 89 Fortune Global 500 companies. These end clients are distributed across e-commerce, FMCG, automotive, retail, beauty, and travel & catering sectors, among others. Among these, FMCG, retail & beauty, and the internet are Deep Intelligence's strong suits. In 2024, revenue from these two major sectors accounted for 36.7% and 30.6% of total revenue, respectively, combining for over 66%. Revenue from the energy sector followed at 13.3%, while no other sector contributed more than 10%.
The primary reason for Deep Intelligence's significant performance volatility in 2024 was that certain overseas clients postponed advertising budgets due to adjustments in their marketing plans. Additionally, end clients in the consumer sector and traditional automotive clients tightened their budgets amid weak consumer demand and a more challenging macroeconomic environment. This led to a 6.53% decline in revenue from the intelligent advertising placement business to RMB 455 million, and a 34.52% decline in intelligent data management revenue to RMB 78.086 million during the reporting period. Consequently, the company's total revenue for 2024 decreased by approximately 12%, causing the adjusted net profit for the period to plummet by nearly 70%. Entering the first half of 2025, due to increased advertising budgets from domestic internet clients and some overseas clients deferring their advertising plans from 2024 to the first half of 2025, Deep Intelligence recorded revenue of RMB 277 million for the period, a year-on-year increase of 5.77%. On the basis of revenue growth, coupled with reductions in sales and R&D expenses, this further boosted the accelerated release of profits during the period, resulting in a massive 760.72% increase in adjusted net profit for the first half of 2025.
The significant volatility in Deep Intelligence's performance indicates that, as its business is heavily concentrated in the intelligent advertising placement sector, its results are susceptible to noticeable fluctuations based on macroeconomic conditions and changes in client budgets. This is particularly evident in the first half of 2025, where intelligent advertising placement accounted for 93.1% of the company's total revenue. Therefore, how to build a more robust business structure is a critical issue that Deep Intelligence needs to address.
Whether it can transform "sector红利 (dividends)" into "growth momentum" becomes the core challenge. In fact, the challenges Deep Intelligence faces in its business operations extend beyond this. The highly fragmented nature of the industry and intense market competition pose potential pressures on its growth. Although Frost & Sullivan data ranks Deep Intelligence first in China's marketing and sales decision AI application market by 2024 revenue, it is noteworthy that its corresponding market share is only 2.6%. The market shares of the second to fifth players are 2.4%, 2.1%, 1.9%, and 1.8%, respectively, indicating a highly fragmented market with numerous players and fierce competition; there are over a hundred participants in this field to date.
Secondly, Deep Intelligence is overly reliant on major clients. According to the prospectus, from 2022 to 2024, revenue from the top five clients accounted for 51.1%, 50.2%, and 54.6% of the total revenue for the respective years, showing an increasing trend. Revenue from the largest client accounted for 22.6%, 14.0%, and 17.8% of total revenue in those years. By the first half of 2025, the revenue contribution from the top five clients had further surged to 70.2%, with the largest client contributing 23.9% of revenue. Meanwhile, the number of end clients decreased from 119 in the first half of 2024 to 113 in the first half of 2025, indicating weak client expansion efforts. This highly concentrated client structure implies that if major clients adjust their cooperation strategies or reduce orders, the company's revenue could face a cliff-like decline.
Furthermore, Deep Intelligence's R&D investment intensity is insufficient, which may be becoming a短板 (shortcoming) constraining its technological iteration and the building of competitive barriers. Data shows that the company's R&D expenses for 2022-2024 were RMB 47 million, RMB 54 million, and RMB 56 million, respectively, accounting for only 8.6%, 8.8%, and 10.5% of total revenue during the same periods. This ratio further decreased to 8.4% in the first half of 2025. Compared horizontally with listed competitors, this level of investment may be inadequate to support deep technological cultivation in "decision AI + marketing scenarios," potentially leading to weakened technological barriers. This lag in technological investment becomes even more pronounced when contrasted with industry growth rates. According to Frost & Sullivan, the size of China's marketing and sales decision AI application market grew from RMB 11 billion in 2022 to RMB 20.3 billion in 2024, achieving a compound annual growth rate (CAGR) of 35.8%, indicating sustained rapid growth. However, Deep Intelligence's revenue growth rate during the same period lagged significantly behind the industry average, suggesting it has not fully capitalized on the market expansion红利 (dividend). This "growth misalignment"直观反映 (directly reflects) deficiencies in the company's client expansion efficiency and product differentiation capabilities, further corroborating the现状 (current situation) that its competitiveness needs enhancement.
Despite current growth pressures, the sector in which Deep Intelligence operates still demonstrates strong long-term potential. Frost & Sullivan predicts that China's marketing and sales decision AI application market will grow from RMB 20.3 billion in 2024 to RMB 94.4 billion in 2029, maintaining a high CAGR of 36.5%, indicating significant industry beta characteristics. This means the company remains in a favorable position, with opportunities to capture market share through strategic adjustments. However, the core challenge lies in whether it can transform the "sector红利 (dividend)" into its own "growth momentum." If Deep Intelligence cannot quickly optimize its business structure, reduce client concentration, increase R&D investment to build core competitive barriers, it may still face risks such as revenue volatility, profit contraction, and even loss of market share. The ability to successfully address these challenges will directly determine Deep Intelligence's valuation logic and premium space in the capital market: only by breaking through the current growth bottleneck of "walking on one leg" and resolving client dependency and profit pressures can it重塑 (reshape) the market's perception of it as the "leader in marketing decision AI." Otherwise, its valuation may continue to be pressured under the label of "high risk, low barriers."
Comments