Woofun AI reports that SenseTime, an artificial intelligence enterprise affiliated with the Chinese Academy of Sciences, officially listed on the Hong Kong Stock Exchange on June 26, establishing itself as the first stock dedicated to large-model decision-making. The company's shares debuted at HK$110, representing an 81.22% premium over the issue price of HK$60.70, with the market capitalization reaching HK$19.076 billion by 10 a.m. This listing event marks a significant milestone for the Beijing-based entity, which has positioned itself at the forefront of the decision intelligence sector. The strong market reception was driven by an oversubscribed public offering that received 5,966.78 times the requested subscriptions, while the international offering garnered 20.70 times the demand. Based on the issue price, the company successfully raised net funds totaling HK$827 million through this issuance.

Founded in 2017 and originating from the Institute of Automation of the Chinese Academy of Sciences, SenseTime specializes in complex data analysis and AI-assisted decision-making services for government agencies and enterprises. The company has accumulated a client base exceeding 650 entities. According to CIC Consulting, SenseTime secured the top position among Chinese enterprise-level large-model-driven decision intelligence service providers in 2024, capturing an 11.4% market share. In the broader Chinese enterprise-level large-model market, the firm ranked seventh with a 2.4% share. The competitive landscape includes major domestic players such as Huawei, YunzhiSheng, iFlytek, Baidu, Alibaba, Torus, Zhipu, and DripTech. SenseTime's core offering revolves around its self-developed 'DIOS Platform,' which serves as the foundational infrastructure for its data governance, model training, and application development capabilities.

The leadership team possesses deep academic and research roots within the Chinese Academy of Sciences system. Chairman and Co-founder Wang Lei is a researcher at the Institute of Automation, while CEO and Co-founder Luo Yin, a senior engineer, also worked at the same institute prior to co-founding the company. Co-founder Zeng Dajun completed the Young Scientists Program at the University of Science and Technology of China and earned a Ph.D. from Carnegie Mellon University. As of December 31, 2025, the R&D team comprised 250 members, representing 43.9% of the total workforce. Within this technical cohort, 66 individuals held master's or doctoral degrees, with core personnel recruited from various Chinese Academy of Sciences research institutes, as well as prestigious universities including Tsinghua University and Peking University. This heavy investment in human capital underscores the company's commitment to technological iteration.

Financially, SenseTime has demonstrated a trajectory of revenue growth alongside improving gross margins, though it has not yet achieved profitability. Revenues climbed from RMB 250 million in 2023 to RMB 318 million in 2024, and further to RMB 405 million in 2025.

Concurrently, gross profit margins expanded from 44.0% in 2023 to 50.4% in 2024, and reached 51.2% in 2025. Despite these positive indicators, the company reported net losses of RMB 260 million in 2023, RMB 157 million in 2024, and RMB 166 million in 2025. After adjustments, the losses narrowed to RMB 186 million, RMB 115 million, and RMB 101 million respectively, indicating a gradual reduction in the severity of losses. R&D expenses fluctuated during this period, standing at RMB 180 million in 2023, dipping to RMB 131 million in 2024, and rising again to RMB 188 million in 2025, with a primary focus on the continuous iteration of the DIOS Platform and its core components.

The revenue structure of SenseTime has undergone a strategic shift, reflecting a pivot from public services toward commercial enterprises. In 2023, public services constituted the largest revenue segment at 50.7%, but this proportion declined to 47.2% in 2024 and further dropped to 36.5% in 2025. Conversely, the commercial enterprise segment experienced significant expansion, growing from 14.5% of total revenue in 2023 to 31.9% in 2025, emerging as the second-largest revenue source. The media and communications segment remained relatively stable, fluctuating between 26.8% and 30.5%, while other business categories saw their share diminish from 4.3% in 2023 to 1.6% in 2025. This structural evolution highlights the company's efforts to diversify its income streams beyond government contracts.

Market concentration in the enterprise-level large-model and decision intelligence sector remains relatively low, with the top five participants accounting for only 38.6% of the market share based on 2024 revenues. SenseTime's leading position in the decision intelligence niche corresponds to revenues of approximately RMB 310 million, while its seventh-place ranking in the broader market also aligns with this revenue figure. The estimated top six competitors include SenseTime, Huawei, YunzhiSheng, iFlytek, Baidu, and Alibaba, while Torus, Zhipu, and DripTech occupy the eighth to tenth positions. This fragmented landscape suggests ample room for growth but also intensifies competition among established and emerging players.

Woofun AI data shows that while SenseTime maintains sufficient cash on hand, its liquidity position is deteriorating. Cash and cash equivalents decreased from RMB 524 million in 2023 to RMB 445 million in 2024, and further declined to RMB 325 million in 2025. Simultaneously, trade receivables doubled from RMB 124 million in 2023 to RMB 248 million in 2025, signaling potential challenges in cash collection. On the liability front, the company incurred additional interest-bearing bank loans and other borrowings totaling approximately RMB 29 million in 2025. Non-current liabilities surged significantly, rising from RMB 3.36 million in 2024 to nearly RMB 100 million in 2025. These financial dynamics indicate increasing leverage and pressure on working capital as the company scales its operations.

Cash flow analysis reveals persistent operating deficits. Net cash flows from operating activities were negative for three consecutive years, amounting to RMB -183 million in 2023, RMB -135 million in 2024, and RMB -188 million in 2025, with the magnitude of the outflow increasing in the most recent year. Investing activities showed a net inflow of RMB 666 million in 2023, but shifted to net outflows of RMB -75 million in 2024 and RMB -14 million in 2025. Financing activities generated net cash flows of RMB -8 million in 2023, followed by inflows of RMB 131 million in 2024 and RMB 81 million in 2025. The reliance on financing to offset operating losses underscores the company's pre-profitability stage and the need for sustained capital injection.

Customer concentration remains a notable factor in SenseTime's business model, although dependence on any single client has decreased. The largest customer contributed 24.3% of revenue in 2023 (RMB 61 million), 19.9% in 2024 (RMB 63 million), and 19.1% in 2025 (RMB 78 million). The top five customers collectively accounted for 48.0% of revenue in 2023, dropping to 32.5% in 2024, and rising slightly to 37.6% in 2025, with corresponding amounts of RMB 120 million, RMB 103 million, and RMB 152 million. While the top customers still represent a significant portion of revenue, ranging between 30% and 50%, the trend indicates a gradual diversification of the client base. The primary customer segments include media and integrated media companies, government agencies, and commercial enterprises.

On the supply chain side, SenseTime's operations are heavily dependent on computing resources and infrastructure. Procurement focuses on three main categories: cloud computing services, hardware devices, and commercial software. For cloud computing, the company relies on providers such as Alibaba Cloud and Volcano Engine to support model training and inference. Hardware purchases encompass GPUs, servers, network equipment, and storage systems.

Additionally, the firm acquires commercial software solutions including CRM, ERP, and network security tools to facilitate business operations and internal management. Supplier concentration is relatively low, with the largest supplier accounting for 7.1% of purchases in 2023, 8.0% in 2024, and 8.0% in 2025. The top five suppliers represented 25.4%, 26.8%, and 24.0% of total purchases in these respective years, reflecting a diversified procurement strategy.

Technologically, SenseTime's competitive edge lies in its self-developed artificial intelligence data operating system, DIOS. This platform integrates data governance, model training, domain knowledge management, and application development into a unified framework. X-Data organizes scattered and heterogeneous raw data into AI-ready datasets, while the Yiyi Large Model provides capabilities in semantic understanding, logical reasoning, and multi-modal processing. DI-Brain focuses on the orchestration of intelligent agents, enabling the integration of models, tool calls, and decision-making processes. The Yiyi Large Model is specifically tailored for enterprise scenarios, supporting over 50 languages and handling images, videos, and text. The company also retains the flexibility to integrate third-party models such as DeepSeek and Tongyi Qianwen based on specific customer requirements.

Intellectual property assets form a critical component of SenseTime's valuation. As of December 31, 2025, the company held 154 registered patents and 108 patent applications in China, totaling 262 patents and applications. Of these, 212 are exclusively owned by SenseTime, while 50 are jointly owned, primarily through collaborations with the Institute of Automation of the Chinese Academy of Sciences and the China Academy of Chinese Medical Sciences.

Additionally, the firm owns 439 software copyrights, 151 registered trademarks, and 10 registered domain names. This robust IP portfolio reinforces the company's technological moat and supports its long-term innovation strategy.

The equity structure of SenseTime is characterized by a dispersed shareholding pattern without an absolute controlling shareholder. Zhongke Sanshi is the largest shareholder with a 16.57% stake, while institutions such as Zhongzi Investment, Zhongke Youcai, and the State Development Bank Manufacturing Transformation and Upgrading Fund each hold approximately 5%. Other significant shareholders include China Internet Investment Corporation, Shenzhen Venture Capital Group, Sequoia China, Zhongguancun Science City, and CCTV Integrated Media. The founding team retains substantial influence through shareholding platforms; Chairman Wang Lei, CEO Luo Yin, and Co-founder Zeng Dajun hold approximately 1.78%, 4.64%, and 2.42% of shares respectively, collectively controlling about 30.66% of voting rights. This structure ensures that the core management team maintains leadership in major decision-making processes.

Compensation for senior management consists of fixed salaries, performance bonuses, and stock-based payments. Total director salaries were RMB 26.87 million in 2023, RMB 21.74 million in 2024, and RMB 18.85 million in 2025. The five highest-paid individuals received RMB 36.85 million, RMB 29.06 million, and RMB 21.96 million in the same periods. In 2025, Chairman Wang Lei and CEO Luo Yin earned approximately RMB 1.83 million and RMB 1.82 million respectively. These figures reflect a compensation structure aligned with the company's growth stage and market conditions.

The enterprise-level large-model and decision intelligence sector continues to attract new entrants, yet market concentration remains low, and the industry landscape is still evolving. SenseTime's strategic shift from public services to commercial enterprises is evident in its changing revenue structure.

However, the company remains in a phase of continuous investment, having not yet achieved profitability. Key challenges ahead include managing customer concentration, controlling the cost of hash rate, and accelerating the commercialization of large models. The successful listing on the HKEX positions SenseTime as a pioneer in the capital market for decision intelligence, but its long-term viability will depend on its ability to translate technological innovation into sustainable commercial value. Whether the market will fully recognize its business model and the extent to which enterprise-level large models can realize their potential will become clearer as the company progresses post-listing.