On February 10, 2026, Shenzhen Zeroth AI Technology Co., Ltd., an AI company founded just over five years ago, formally submitted its listing application to the Hong Kong Stock Exchange, with CMBC Capital as the exclusive sponsor. According to Frost & Sullivan, by revenue in 2024, Zeroth AI has become the fifth-largest provider of enterprise-grade AI agent solutions in China, with a market share of 3.0%. Centered on its self-developed“BlackZero”platform and focused on AI assistants for contract and investment scenarios, the young company has stepped into the capital market with a strong financial track record of net profit more than doubling over three years.

I. Core Business: Built on BlackZero, Entering High-Value Scenarios via AaaS Model

Zeroth AI’s core product logic is clear: instead of pursuing large-scale general-purpose AI, it focuses on two high-value, compliance-intensive vertical scenarios–legal contracts and investment decisions–and delivers quantifiable business outcomes under the“AI-as-a-Service”(AaaS) model.

The company’s technical foundation is the self-developed BlackZero platform, designed around the principle of“large models for reasoning, small models for execution”. Through multi-agent collaboration, complex tasks are split into parallel-processing modules, enabling open semantic understanding while achieving high-precision operations in specific domains. This architecture allows Zeroth AI to build its own technological moat without falling into homogeneous competition relying on open-source large models.

The revenue structure clearly reflects its business focus. Contract Assistant is the core cash cow, contributing RMB 77.3 million and RMB 123.1 million in revenue in 2023 and 2024, accounting for 89.4% and 72.8% of total revenue respectively. Investment Assistant has shown strong growth momentum, with its revenue share rising rapidly from 7.0% in 2023 to 24.1% in 2024. In early 2025, the company launched Brand Assistant to enter the consumer brand public opinion monitoring sector, which generated RMB 4.2 million in revenue in the first nine months, serving as an early signal of the third growth curve.

For delivery, the company adopts a dual-track strategy:“data-supported deployment + on-premises deployment”. The former charges by project, with revenue linked to clients’claimed amounts; the latter charges fixed annual fees through software licenses. This combination ensures high average client value under the project model (average annual contribution per Contract Assistant client exceeded RMB 10 million in 2023) while reserving room for the standardization and scaling of products in the future.

II. Financial Perspective: Structural Optimization and Concerns Amid High Growth

Zeroth AI’s financials exhibit typical characteristics of a growing AI enterprise: rapid revenue expansion and synchronized profit growth, yet high customer concentration remains a prominent issue.

Revenue and Profit: The company’s revenue increased from RMB 86.5 million in 2023 to RMB 169 million in 2024, representing a year-on-year growth of 95.5%; it reached RMB 186 million in the first nine months of 2025, exceeding the full-year figure of 2024. Net profit rose in tandem, from RMB 21.26 million in 2023 to RMB 31.641 million in 2024, and reached RMB 40.413 million in the first nine months of 2025, up 63.1% year-on-year. Gross margin remained stable at 39%–42%, reaching 41.6% in the first nine months of 2025, demonstrating mature cost control capabilities.

R&D investment is a key window to observe the long-term value of this AI company. R&D expenditure reached RMB 15.482 million in the first nine months of 2025, accounting for 51.7% of total selling, administrative and R&D expenses, a significant increase from 37.0% in 2023. R&D personnel accounted for over 40% of total staff, and the development of core AI models and algorithms was fully in-house, with outsourcing only involving peripheral systems.

Customer structure is the company’s most notable weakness. In 2023, 2024 and the first nine months of 2025, the top five customers contributed 96.9%, 94.7% and 85.8% of total revenue respectively, with the largest customer alone accounting for 58.4%, 44.0% and 34.7%. Although dependency has been declining year by year, concentration remains extremely high.

III. Equity Structure and Market Prospects

As of the Latest Practicable Date, founder Mr. Qiao controlled approximately 55.7% of the company’s shares through Shenzhen Qingxin Management and direct personal holdings, making him the controlling shareholder. The company had certain equity nominee arrangements in 2023 and 2025, all of which had been cleaned up before the submission of this document.

The company completed its angel round financing in 2021, introducing Ziyi Yunshang as an early investor. Following the equity restructuring in September 2025, Ziyi Yunshang partially exited, and some of its beneficial owners continued to hold shares through newly established special purpose entities Tianyunlu and Xingqiongwujie.

China’s enterprise AI agent solution market, where Zeroth AI operates, is on the eve of explosive growth. According to Frost & Sullivan, the market size is projected to surge from RMB 5.6 billion in 2024 to RMB 59.1 billion in 2029 at a CAGR of 60.2%, expanding nearly 10 times within a decade.

However, high-growth tracks never lack competition. Zeroth AI currently ranks fifth with a 3.0% market share. Its core competitiveness lies in its deep understanding of vertical scenarios and delivery reputation. The number of Contract Assistant clients rose from 8 in 2023 to 26 in the first nine months of 2025, and the average number of projects per client increased from 1.1 to 2.0, indicating strong client repurchase and demand expansion capabilities.

Nevertheless, this reputation remains highly concentrated in a single product; the scale of Investment Assistant and Brand Assistant is insufficient to support the second growth curve. Second, clients are highly concentrated in investment institutions, and its anti-cyclical resilience yet to be tested. Third, the penetration rate of the software license model remains low, with only 13 clients in the first nine months of 2025, and the scaling capacity of standardized products has not been proven.