A surge in demand for data infrastructure during the artificial intelligence boom is pushing Shenzhen-based Xunce Technology into the spotlight.
Often described by analysts as China’s equivalent to Palantir, the company said its revenue nearly doubled in 2025, surpassing earlier expectations and highlighting how the race around large language models is shifting toward a new battleground: data infrastructure.
Xunce released a profit forecast on March 6, projecting full-year 2025 revenue of about RMB 1.283 billion (USD 179.6 million), up 102.95% year-on-year. Adjusted net loss is expected to narrow to roughly RMB 55 million (USD 7.7 million), compared with RMB 82 million (USD 11.5 million) in 2024.
The company attributed the remaining losses primarily to a period of rapid expansion and sustained R&D spending. It is also extending its AI infrastructure beyond its initial domains into emerging areas such as robotics data platforms and commercial aviation.
Xunce listed on the Hong Kong Stock Exchange on December 30, 2025. It positioned itself as the first publicly traded company built around a data agent architecture designed for large AI models.
Financial momentum accelerated sharply in the second half of the year. According to the company’s earlier prospectus, Xunce generated RMB 198 million (USD 27.7 million) in revenue during the first half of 2025. That implies more than RMB 1 billion (USD 140 million) in revenue in the second half alone, representing a sequential increase of about 448%.
In its earnings notice, the company said the revenue growth reflects rising demand for data capabilities as AI models move from experimentation to real-world deployment. With its AI data agent platform at the core, Xunce has built an end-to-end technology stack covering data acquisition, cleansing, standardization, real-time processing, and model optimization. The system is designed to process data with millisecond-level latency.
Leveraging that infrastructure, Xunce said its AI platforms and solutions are increasingly deployed across a wider range of industry scenarios, while existing customer partnerships deepen.
The same day it released the profit forecast, the company was also added to the Hong Kong Stock Connect program, effective March 9. Inclusion allows mainland Chinese investors to trade its shares through the cross-border channel, a development that often increases trading liquidity and investor visibility.
The move may also reflect growing confidence among capital market participants in Xunce’s long-term prospects, while potentially supporting its valuation as access to cross-border investors expands.
Analysts draw parallels with Palantir
Over the past month, several domestic and international investment banks have initiated coverage of Xunce, broadly assigning “buy” or “overweight” ratings.
Deutsche Bank, for example, said the company’s growth trajectory resembles the early strategy of Palantir. Rather than scaling through low-barrier applications, Xunce first commercialized its technology in sectors with complex data environments and high technical requirements.
Palantir began by working with military and government intelligence agencies before expanding into finance, healthcare, and manufacturing. Over time, it developed into one of the world’s most prominent AI data platforms.
Xunce followed a similar path. It initially focused on the asset management industry, a sector known for highly complex datasets and strict requirements for accuracy and timeliness.
Successfully commercializing technology in such environments can create a durable competitive moat. The real-time data processing capabilities Xunce developed in asset management, including the ability to manage high concurrency, low latency, and strong data consistency, can transfer to other industries.
Deutsche Bank estimates that revenue from industries outside asset management could grow at a compound annual rate of 109% between 2024–2027. Telecommunications, urban governance, and manufacturing are expected to become key growth drivers.
The company’s customer base has already expanded beyond asset management to include financial services, city management, manufacturing operations, and telecommunications. Clients include China’s three state-owned telecommunications operators.
In 2024, industry application revenue from sectors outside asset management accounted for 61.3% of Xunce’s total revenue, up from 34.1% in 2023 and 25.6% in 2022.
Modular architecture as a moat
Analysts at Guotai Haitong also highlighted the company’s technological differentiation, particularly its modular product architecture.
The design gives Xunce’s platform a high degree of flexibility. Individual modules can be combined to meet the needs of different industries or configured into customized solutions for specific clients.
According to the company’s prospectus, Xunce had developed more than 300 modules as of June 30, 2025, spanning real-time data infrastructure and data analytics capabilities.
The number of modules increased from 152 at the end of 2022 to 318 at the end of 2024, and 332 by mid-2025.
The modular approach echoes Palantir’s product strategy. Its core platforms, Foundry for commercial clients and Gotham for government users, are built around component-based architecture in which individual modules provide specific functions and can be assembled based on client requirements. Palantir’s Apollo platform enables remote updates, centralized management, and compliant deployment of software across global customers operating in heterogeneous environments.
For Xunce, the growing module library could also increase revenue per user. Deutsche Bank estimates that average revenue per user could grow at a compound annual rate of 83% between 2024–2027 as clients adopt more modules and expand usage across deeper project deployments.
Taken together, analysts increasingly frame Xunce not as a traditional software company but as a foundational data infrastructure platform for the AI era. That perspective partly explains why Guotai Haitong assigned the company a target price of HKD 104.78 (USD 13.4), implying nearly 40% upside from its recent share price.
As competition around large AI models matures, the focus may gradually shift from algorithms to the data systems that power them. Companies capable of building those underlying rails may shape the next phase of the AI sector.
This article was adapted based on a feature originally written by Stone Jin and published on IPO Zaozhidao. KrASIA is authorized to translate, adapt, and publish its contents.