XunCe's first financial report after going public: Revenue increased by 103% in 2025 and achieved semi-annual profit

A scarce target in the Token economy wave.

According to news you saw before IPO, Shenzhen Xun Kae Technology Co., Ltd. (hereinafter “Xun Kae Technology,” 3317.HK) released its annual performance announcement for the period from January 1, 2025 to December 31, 2025 on March 27, 2025. This is also its first financial report published since it became the “Token No. 1 stock” on December 30, 2025.

The financial report shows that Xun Kae Technology’s 202****5 revenue was RMB 1.29B (the same below), up 103.28% year over year from 2024, successfully crossing the key “billion-yuan revenue” threshold. This marks that it has officially moved from an early technology-driven startup stage into a new era of platform-based development that can be scaled and replicated. It is worth noting that in the first half of 2025, Xun Kae Technology achieved revenue of RMB 198 million, while revenue jumped to RMB 1.09B in the second half, up 449.32% quarter over quarter, demonstrating strong growth momentum.

In terms of profitability, in 2025, Xun Kae Technology’s consolidated gross profit was approximately RMB 792 million, up 63.44% year over year, with a consolidated gross profit margin of 61.66%. At the same time, Xun Kae Technology’s adjusted net loss in 2025 narrowed sharply by 33.41% year over year to RMB 54 million. Behind the narrowing loss is a strong confirmation that the scale effects brought by rapid business development are gradually emerging, human resource efficiency has been greatly improved, and operating quality has continued to be optimized.

One point that needs to be highlighted is that based on the quarter-over-quarter changes in adjusted net profit, Xun Kae Technology’s adjusted net loss in the first half of 2025 was RMB 105 million, while it achieved adjusted net profit of RMB 50 million in the second half—successfully迎来 a performance turning point and for the first time achieving positive semiannual profitability. This shift indicates that while Xun Kae Technology maintains rapid expansion, its profitability has begun to be meaningfully released, and the scalability of its business model and its profit potential have been preliminarily validated.

As of December 31, 2025, Xun Kae Technology’s audited consolidated cash and cash equivalents were approximately RMB 1.09B, with ample cash on hand.

ARPU and human-efficiency levels both doubled

Accelerating penetration across diversified scenarios, with business structure becoming more robust

In 2025, Xun Kae Technology, centered on its AI Data Agent, continued to strengthen millisecond-level real-time data processing capabilities. It built an end-to-end technical system covering data acquisition, cleaning, standardization, real-time computation through to large-model tuning, delivering data infrastructure and data analytics solutions to customers across industries, and enabling seamless deployment in customers’ private cloud and on-premises systems.

Currently, Xun Kae Technology has coverage across 9 major industries. Compared with Palantir’s 17 industries, there is broad room for horizontal expansion, and it is also accelerating deep expansion into nationally prioritized industries such as asset management, telecom, power, urban management, high-end manufacturing, healthcare, energy, robotics training platforms, and commercial aerospace. Among these, every time it enters a new industry, Xun Kae Technology first uses 3–5 years to accumulate industry data, and then it can achieve rapid replication, promotion, and scaling to customers within the same industry.

It’s undeniable that in the present era where AI inference is taking hold, Tokens are evolving from “fuel” into “hard currency.” How to maximize the value of every Token has become the core proposition in the large-model inference stage. Today, general-purpose large models commonly adopt a strategy of “trading compute for accuracy,” and every inference is accompanied by a large consumption of ineffective Tokens. Once an inference fails, all the Tokens invested in the earlier stage become worthless—this is a shared dilemma faced by general AI.

By contrast, niche AI solutions install an “extra brain” for general-purpose large models using industry data—its core is to optimize the inference path with business models, judge task feasibility in advance, and avoid Token waste at the source. Xun Kae Technology has been deeply focused for many years on specialized niche data modeling. With its years of accumulation of high-quality, scenario-based niche data, it is effectively like equipping an “efficiency enhancer” on every Token call—trading off Token consumption for higher-accuracy results, achieving the highest level of output certainty. Even more importantly, Xun Kae Technology is building end-to-end data measurement, billing, and settlement capabilities, so that every data call can be quantified and priced, and by improving the effectiveness of each Token, it brings customers greater business value**.

Guided by this approach, Xun Kae Technology’s platform uses a “Lego-style” modular architecture, enabling customers to flexibly combine modules according to their own needs and achieve “on-demand assembly, and change whenever needed,” building deep stickiness. The company also follows flexible configuration in its pricing and charging models: based on dimensions such as number of modules and processing speed, it sells through subscription, transaction-based, and Token-based pricing models to accurately match customer needs.

At present, Xun Kae Technology is fully building an end-to-end data measurement and settlement system, exploring pricing by dimensions such as the number of large-model calls and the number of module applications, so that customers pay for “effective Tokens” rather than paying for compute consumption.

In terms of the business model, Xun Kae Technology’s solutions are deeply deployed in customers’ private clouds or on-premises systems, acting as the “data hub,” and accelerating its transition from a traditional tool provider to an “AI Agent.” By embedding AI deeply into core business processes, Xun Kae Technology’s value positioning upgrades from “tools” to “AI Agent,” strengthening customers’ willingness to pay.

In 2025, Xun Kae Technology had 230 customers. ARPU increased sharply from RMB 2.72 million in 2024 to RMB 5.59 million in 2025, up 105.04% year over year. Customer value has been deeply unlocked, and its ongoing enhancement of commercialization capabilities has continued.

Also worth paying attention to is that in 2025, Xun Kae Technology’s employee headcount was 448. Revenue generated per employee was RMB 2.87 million, up sharply from RMB 1.22 million in 2024 by 135%, and the human-efficiency level has doubled. This step-change improvement is mainly attributed to Xun Kae Technology’s platform-based and modular product architecture, as well as its AI-enabled data development system, which significantly lowers human investment while enabling solutions to be rapidly replicated and promoted.

As the scale of revenue breaks through the billion-yuan threshold, scale effects are gradually becoming more evident, and the dilution effect of fixed costs continues to strengthen. Xun Kae Technology’s revenue per employee is already significantly higher than that of traditional software companies, and also remains at a leading level among AI technology companies, fully confirming the platform advantages of its business model and operating efficiency. With the AI Data Agent strategy being pushed deeper and the business model upgrading, Xun Kae Technology’s human-efficiency is expected to remain at a high level, laying a solid foundation for releasing profitability.

In terms of global expansion, Xun Kae Technology also plans to increase the share of overseas revenue to 10–15% in 2026, and from 2027 to 2028 it will continue to advance its globalization strategy, opening up new space for long-term growth.

The development strategy will be deepened around five core directions

A scarce target in the Token economy wave

Looking ahead, Xun Kae Technology believes that the AI data sector is entering a deep intersection of five core trends, forming a solid underlying logic for long-term development:

First, core data infrastructure in the AI Agent era. AI is evolving from “conversation and interaction” to “autonomous execution.” The large-scale application of AI Agents places higher demands on data real-time capability, security, and quality. For more than a decade, our company has been deeply focused on real-time data, and has become the core hub supporting AI Agents’ perception of environments, tool invocation, and task execution.

Second, data refinement for AI niche models—from a key leap from general-purpose to specialized. The boundaries of general-purpose large-model capabilities are becoming clearer. Specialized models for vertical industries are becoming a key carrier for the industry’s intelligent upgrade. Xun Kae Technology has long served core customers in highly complex and highly compliant domains, and has the core capabilities to provide high-quality training datasets, domain knowledge graphs, and data labeling services for vertical model vendors—making it a key data infrastructure supplier for the niche-model ecosystem. Through professional data operations capabilities, Xun Kae Technology can enhance the business value of each unit of Token.

Third, the data foundation for Open Claw and the next generation of AI operating systems. In response to trends in next-generation AI operating systems such as Open Claw, Xun Kae Technology is proactively planning ahead: it is deeply involved in building the AI operating system ecosystem, jointly defining data interface standards and scheduling protocols; at the same time, relying on its real-time data infrastructure capabilities, it builds “Tokenized data” services, aiming to become an indispensable core data Token supplier within the new operating systems, providing stable, efficient, and trusted data fuel for AI ecosystems and niche model applications.

Fourth, Tokenized payments—building a business-model evolution toward “pricing by value.” The data element market is evolving from “non-standard transactions” to “standardized transactions.” Business models that charge based on Tokens used will become a new trend. Xun Kae Technology can build an end-to-end data measurement, billing, and settlement system for customers, and explore pricing by dimensions such as the number of large-model calls and the number of module applications, laying a technical foundation for innovative business models such as revenue sharing from data operations.

Fifth, data capitalization will become a core enterprise strategy. As the listing of data assets on balance sheets has moved from policy exploration to large-scale practice, enterprises’ rigid demand for infrastructure such as data governance, data standardization, and data quality control continues to be released. Xun Kae Technology’s real-time data platform, which has been deeply cultivated for ten years, is becoming the core technical foundation for data capitalization.

At the intersection of these five trends, “high-quality, structured, scenario-based data” is becoming a strategic resource that determines a company’s competitiveness in the AI era.** Especially in high-value vertical fields such as finance, energy, healthcare, consumer sectors, robotics, and city operations, what is needed is no longer rough, low-cost pipeline-style data labeling, but rather professional data operations capabilities that both understand industry business logic and understand the path to deploying AI, as well as streaming computation capabilities that can be embedded into real-time business processes. Relying on its end-to-end technical system, Xun Kae Technology is becoming an indispensable “data hub” in the process of intelligent transformation.

Meanwhile, as the focus of AI applications shifts from model training to inference deployment, the unit for measuring business value is undergoing a fundamental leap—from “compute consumption” to “Token consumption.” Xun Kae Technology’s deployment of a Token payment and outcome revenue-sharing business model is redefining the path to realizing data value in the AI era, becoming a scarce strategic asset in the Token economy wave.

Xun Kae Technology’s development strategy will be deepened around five core directions: first, deepen the evolution of its business model, accelerating the upgrade from project-based and subscription-based models to a Token-based payment and revenue-sharing model, so that returns are deeply tied to customer value; second, accelerate cross-industry replication, relying on real-time data processing capabilities to expand deep into key national areas such as power, telecom, healthcare, energy, robotics training platforms, and commercial aerospace; third, expand cutting-edge applications, continuously increasing R&D investment in emerging areas such as robotics data platforms and commercial aerospace to seize the commanding heights of future industries; fourth, steadily expand overseas business, advancing globalization under controllable risk; and fifth, build a strategic cooperation ecosystem, tightly linking with leading compute providers and algorithm companies at home and abroad, to provide customers with a one-stop solution from underlying compute to top-layer applications.