China’s Daily Token Usage Surges 40% in Just Three Months

Select Language:

China’s use of artificial intelligence is accelerating rapidly, with recent figures revealing a significant rise in the daily processing of tokens—the smallest units handled by large language models—over the past three months.

This month, the average daily token usage in China has surpassed 140 trillion, which is more than 1,000 times greater than the 100 billion recorded at the beginning of 2024. Additionally, there’s been a 40% increase from 100 trillion tokens at the end of last year, according to Liu Liehong, head of the National Data Administration, speaking in Beijing.

Liu pointed out that tokens can now be quantified, priced, and traded, leading to the emergence of a new valuation system that revolves around their utilization, distribution, and settlement. This development is becoming a crucial avenue for transforming the AI industry into a commercial enterprise.

The notable surge in daily token activity indicates that China’s artificial intelligence sector is entering a phase of rapid expansion, significantly heightening its global competitiveness. It also reflects an increase in data set supplies—these are the organized collections of information used to train large language models—and underscores the value generated from data elements. This cycle of data utilization is fueling ongoing innovation and growth in AI.

Earlier this month, the Government Work Report emphasized initiatives aimed at fostering an intelligent economy and expanding the “AI +” strategy, which seeks to promote AI integration across various industries. The report also highlighted plans to develop and leverage high-quality data resources and datasets.

2026 marks the beginning of China’s 15th Five-Year Plan and has been officially designated as the “Year of Data Element Value Release.” By the end of last year, over 100,000 high-quality datasets had been created nationwide, totaling more than 890 petabytes—roughly 310 times the digital material stored in the country’s National Library.

Liu stated that the next phase involves advancing innovation through data-driven AI development and collaborating with industry stakeholders to implement a new action plan focused on high-quality dataset creation. This will include strengthening foundational infrastructure, expanding capacity, overcoming annotation challenges, enhancing data quality, improving efficiency, and unlocking the value of data for practical applications.

The goal is to develop datasets that are not only high in quality but also practical and ready for AI deployment. Additionally, efforts will focus on establishing a unified national data property registration system, rolling out policies for a comprehensive data market, and fostering the development of essential data infrastructure. These measures aim to provide robust support for maximizing the value of data elements.