China's data industry reached 6.78 trillion yuan (about $1 trillion) in 2025, and the country is now positioning high-quality data as the foundation for its AI ambitions. At the 2026 China International Big Data Industry Expo in Guiyang, which drew 372 companies and over 16,000 registered guests, officials and industry leaders laid out how data collection, standardization, and circulation will drive the next phase of AI development.
"Wherever AI advances, we push forward the construction and application of high-quality datasets," Liu Liehong, head of China's National Data Administration, said at the expo's opening ceremony. The event, themed "Token: A New Path to Value of Data Elements," ran from Friday through Sunday and showcased new products in the token economy, computing power, large model innovation, and data element circulation.
Data infrastructure upgrades traditional industries
The expo highlighted how data and AI are reshaping conventional sectors, with energy a leading example. Guizhou Wujiang Hydropower Development Co. has connected hydropower stations, wind farms, and solar plants along the Wujiang River through a unified digital system, addressing a long-standing problem: renewable energy output swings wildly with weather conditions.
"We're bringing together hydropower, wind and solar into one smart dispatch platform," said Fan Sheng, deputy general manager of the company.
The company cleaned and standardized data from meteorological, hydrological, and grid operations, then linked that data to the grid dispatch system. Technical teams trained multiple models using years of local data, and the models can now predict conditions up to 15 days ahead. Test data shows wind forecast accuracy at 91.95 percent, solar forecast accuracy at 94.54 percent, and cascade hydropower inflow forecast accuracy at 94.1 percent - an improvement of about 4 percentage points year on year.
For IT professionals, this points to a practical reality: the value of AI in industrial settings depends on the quality of the data pipeline feeding it. Duan Zhongxian, a professor at Guizhou University, put it directly: "AI's edge in industrial upgrading goes beyond technology-it lies in its reach as a foundational infrastructure." He advocated for integrating computing power with industry, agriculture, and cultural tourism.
New jobs and business models emerge from AI data work
The data-AI relationship is also generating entirely new employment categories. Embodied intelligence - AI with a physical body that perceives and acts like a human - requires massive amounts of real-world behavioral data. In Guiyang, 21-year-old Wang Junbao, a graduate of Guizhou Tongren Data Vocational College, works as an embodied intelligence data collector. Wearing a panoramic data collection headset, he records household tasks from a first-person perspective: how to grasp objects, how to exert force, how to navigate obstacles. That data becomes training material for robots.
"This industry has great potential, and it's deeply rewarding to contribute to the progress of embodied intelligence," Wang said. He joined Shanghai Benyuan Digital Technology Co. in June after a three-month internship. "Five or six of my classmates stayed with the company after their internships."
Duan Annan, general manager of the company's Guiyang branch, said the base employs 400 on-site staff plus 800 collectors nationwide. "Real-world scenes, real interactions, and real action data are the core bottlenecks constraining model deployment," he said. The company has supplied over 100,000 hours of effective collection data for domestic embodied intelligence models.
The expo also showed how AI is reshaping business structure itself. Many exhibitors were micro-enterprises with three to five employees, or even single-person companies. Qin Yongbin, vice president of Guizhou Minzu University, framed the shift in practical terms: "AI is not simply taking away jobs. It is reshaping existing work and creating entirely new professions." He predicted that while AI replaces routine tasks, it will generate more roles demanding creativity, decision-making, and complex teamwork.
For developers and IT professionals, the embodied intelligence data pipeline is worth watching. The skills involved - data collection, labeling, quality control, and model training - are directly relevant to anyone working in AI for IT & Development. The demand for structured, high-quality data is not a niche concern; it's becoming a core constraint across AI deployment.
International cooperation expands around data and AI
The expo drew participants from Singapore, Morocco, Canada, and beyond, with several announcing concrete cross-border projects. Dubuya Thulani, a Zimbabwean entrepreneur, launched Wazisana, a blockchain-based platform that connects African agricultural exporters with Chinese buyers. The platform requires farmers to upload videos of their production process, and AI processes that data into readable descriptions for potential clients.
"We have already exported coffee, red wine and avocados successfully, and we are already profitable," Thulani said. "Next, we want to expand our business and contribute to win-win cooperation between China and Africa through data and AI."
Syed Muhammad Ashiq Hussain Shah, agriculture minister for Pakistan's Punjab province, led a delegation of nearly 30 officials to study China's smart agriculture systems. His province is building an AI-powered digital monitoring network for farming. "We are focusing on finding partners in China and learning from China's experience," he said.
Cristian Davis Acosta, consul general of Chile in Chengdu, said Chile is developing Latam-GPT, an AI tool designed for Latin America. "China already has very advanced AI tools, so perhaps we can establish cooperation between institutions on both sides and exchange experiences," he said.
Muhammadou M.O. Kah, vice-chair of the United Nations Commission on Science and Technology for Development, urged open international cooperation. "Expanding access to digital public goods, open AI methodologies, and capacity building, so that every nation can actively participate in the data economy," he said.
Why this matters for IT and development professionals
The through-line of the expo is that data quality determines AI capability. Yang Jie, secretary-general of the World Data Organization, said the "richness, purity, authenticity and professionalism of data directly determine the cognitive depth, knowledge accuracy, and capabilities of large models." That's a direct signal for developers: the bottleneck in AI projects is often not model architecture but the data infrastructure around it.
For those working in IT and development, the practical takeaway is to invest in data engineering skills - cleaning, labeling, validation, and pipeline construction. The Guiyang expo shows that organizations are paying for exactly these capabilities, whether they're building energy forecasting systems, training embodied intelligence robots, or creating cross-border trade platforms. The demand for professionals who can build and maintain high-quality data pipelines is growing across industries, not just in tech companies.
Professionals looking to build these skills can explore AI Data Analysis Courses to understand how data quality directly impacts model performance. The work on display in Guiyang - from energy forecasting to embodied intelligence - is fundamentally about getting data right before the model ever runs.
Your membership also unlocks:
