China's appetite for artificial intelligence data is exploding at a scale that is beginning to reshape the economy, with daily token usage soaring from just over 1 trillion at the start of 2025 to 100 trillion by year-end, showing "exponential growth", officials said on Wednesday.

The figure rose further to 140 trillion by March this year, according to Liu Liehong, head of the National Data Administration, underscoring the accelerating value of data as a production factor and the rapid emergence of new forms of the "intelligent economy".

The figures, disclosed at the opening of the 9th Digital China Summit in Fuzhou, Fujian province on Wednesday, reflect not only rising demand for AI services but also the emergence of what officials describe as the token economy, where computing activity becomes measurable, tradable and increasingly central to economic output.

China generated 52.26 zettabytes of data in 2025, up 27.28 percent from a year earlier, with AI and system software accounting for 26.92 zettabytes — overtaking internet-of-things sensing data for the first time. The shift highlights a structural pivot from passive data collection to AI-driven data creation.

"Data is no longer just fuel for training models. It has become a core production factor driving iteration and real-world deployment,"Liu said, pointing to applications across healthcare, finance, manufacturing and urban governance.

A notable inflection point has emerged in how data is used. Of the 199.48 exabytes deployed for AI training and inference in 2025, inference data — generated when models are actively used — exceeded training data for the first time, reaching 101.34 exabytes. The trend suggests China's AI industry is moving beyond model building toward large-scale commercialization.

Industry experts said the rapid scaling of token usage could have far-reaching implications, from pricing models for AI services to global competition over data resources.

As AI systems shift from generating content to executing decisions — and from digital environments into the physical world — the volume, variety and velocity of data are expected to accelerate further, cementing data as a central pillar of economic growth in the AI era.

Behind the numbers, a new layer of industrial infrastructure is also taking shape.

Fujian-based Joyful Embodied is building what it describes as a large-scale robotic data collection facility, designed to operate around the clock capturing high-precision industrial data. Using arrays of cameras and sensors, robots replicate and record complex physical tasks, converting them into structured datasets for training embodied AI systems.

Chen Yishi, president of Joyful Embodied, said its automated pipelines — combining perception, computation and execution — can achieve data collection efficiency of up to 95 percent, significantly above industry norms. The goal is to standardize data generation and improve the usability and transferability of datasets across applications.

Such efforts point to a growing race not only for computing power and algorithms, but also for high-quality data — increasingly seen as the bottleneck in advancing next-generation AI, particularly in robotics and embodied intelligence.

Chen said that the company is also expanding overseas, partnering with firms in Indonesia and eyeing broader Southeast Asian markets, reflecting a wider push by Chinese AI companies to export both technology and data infrastructure.