China’s National Data Group Puts Telecom and Grid Giants Behind AI Training Data

Registration records show RMB30 billion in capital, with SASAC holding 20% and each of China’s three telecom carriers 15% of the new central SOE.

WHY IT MATTERS

The carriers and grid operators own 70% of a company whose registered scope covers data processing and AI software, putting their data close to China's AI developers.

China’s newest central state-owned enterprise has a mandate that reaches straight into the AI supply chain. China National Data Group Co., Ltd. was inaugurated on October 20 in Xiong’an New Area, Hebei province, with Vice Premier Ding Xuexiang attending, according to a Xinhua report carried on the State Council’s English website. Among its stated tasks is supplying high-quality datasets and corpora to AI companies.

A week later, the question is whether a single state company can actually pry that data loose. In an analysis published on October 26, the South China Morning Post reported that experts see the group’s success as depending on tearing down entrenched barriers to data sharing. Sun Yi, a professor at the University of Chinese Academy of Sciences, told People’s Daily the group could set unified data standards and coordinate data collection and procurement, supplying high-quality datasets for Chinese AI models and robots, the SCMP reported.

The ownership is now on record. Business registration data published on October 26 show registered capital of RMB30 billion, C114 reported. The State-owned Assets Supervision and Administration Commission (SASAC) holds 20% through a RMB6 billion contribution. China Mobile, China Telecom and China Unicom each put in RMB4.5 billion for 15%, while State Grid and China Southern Power Grid each put in RMB3.75 billion for 12.5%. China Xiong’an Group holds the remaining 10%, according to C114 and Caiwen, which cited Qichacha, a company-records database. Beijing Daily described the company as wholly state-owned.

The group becomes Beijing’s commercial arm for data under SASAC

The State Council approved the company, and SASAC exercises investor rights on its behalf, according to SASAC’s founding announcement dated October 20. Xinhua said the group ranks 28th on the roster of central enterprises and brings their number to 100. Xinhua said the founding fills a gap, since the state had no dedicated corporate vehicle for the data sector until now.

The leadership comes from the telecom sector and the regulator. Jian Qin, a former general manager of China Unicom, is chairman, and Chen Ronghui, a former deputy head of the National Data Administration, is general manager, China Daily reported. According to Zhang Xianghong, a member of the national data expert advisory committee, cited by China Daily, the administration, set up in 2023, will focus on basic data rules and policy design, while the company handles the development and market operation of data resources.

Ding, who called the venture a "major strategic decision," urged the company to "play a strategic, leading and supporting role in the data sector," according to the State Council report. An earlier SCMP report quoted Xinhua as saying the group’s tasks include "serving AI development and fostering a data industry ecosystem." Xinhua described Xiong’an, which has set its own targets for AI and robotics growth, as a testing ground for moving data across regions and departments.

Training data is where the mandate meets the sharing problem

The group will focus on Beijing’s "AI Plus" initiative and "vowed to provide high-quality datasets and corpora to further the development of the AI industry," People’s Daily Online reported, citing Xinhua. Analysts quoted there expect it to tackle barriers to data circulation, high data governance costs and a shortage of high-quality datasets.

The raw material is large. China’s total data output reached 52.26 zettabytes in 2025, up 27.28% from a year earlier and about 27.44% of the global total, according to a national survey released in April and cited by People’s Daily Online. Enterprises accounted for about 90% of the increase. China’s data industry grew 15.7% to RMB6.78 trillion in 2025 and counted 482,000 companies, the SCMP reported, citing the National Data Administration.

Zhang framed the company as a tool against fragmentation. "The group will promote the cross-industry and cross-regional flow and sharing of data resources," he said, according to China Daily, adding that it would help break down "data silos." The registered business scope points the same way: big data and data processing services, AI application software, cloud services and IT system integration, plus licensed value-added telecom and internet information services, C114 reported. The group holds no basic telecom license. One shareholder has said what it will bring: China Southern Power Grid will connect its energy-sector data-sharing infrastructure to the group’s national platform, Yu Yun, head of its digitization department, told People’s Daily, according to the SCMP. The National Data Administration has been working on the same problem from the rules side; in September it was moving toward embodied AI training standards.

First data products will show how far the group reaches

The national group arrives after provinces built their own. Shanghai’s data group has registered capital of RMB5 billion and Fujian’s RMB10 billion, Yicai noted, against RMB30 billion for the national company. The group also faces a balancing act between its state role and commercial operations. People’s Posts and Telecommunications News, a state-backed industry paper, warned in an editorial that "excessive administrative intervention could squeeze out innovation by private enterprises," the SCMP reported, and urged the group to act as an enabler rather than a replacement for existing businesses.

The first concrete signal will be the group’s first datasets or corpora offered to AI developers, and the terms on which they are sold. A product drawing on carrier data, or on the grid data China Southern Power Grid has promised, would show the shareholders are contributing more than money.