April 30, 2026 – At the 9th Digital China Construction Summit, China’s “National Data Resource Survey Report (2025)” was officially unveiled, revealing a significant surge in the role of data resources in driving artificial intelligence (AI) innovation across the country.
The report highlighted that China’s data ecosystem has undergone a transformative shift, with token usage—the foundational units of AI data processing—soaring to unprecedented levels. By the end of 2025, daily token calls skyrocketed from over 1 trillion at the start of the year to a staggering 100 trillion, marking exponential growth. Annual cumulative token calls reached approximately 21,100 trillion, underscoring the nation’s rapid adoption of AI technologies.
Tokens, defined as the smallest units of information processed by AI models, function as the building blocks for efficient data handling. These “text fragments” or “symbolic units” enable large language models to break down complex inputs—such as user queries, generated text, or image recognition tasks—into manageable components. Unlike the internet era, where data flow was measured in bandwidth or traffic, the AI age prioritizes token consumption as a critical metric for technological advancement.
Every character typed by a user, every paragraph generated by an AI model, and every image analyzed by machine learning systems relies on token processing. This shift reflects China’s strategic focus on leveraging data resources to fuel AI breakthroughs, with applications spanning industries from healthcare to autonomous vehicles.
The report also emphasized the growing importance of optimizing token efficiency to reduce computational costs and enhance model performance. As China continues to expand its AI infrastructure, experts predict token usage will remain a central indicator of the sector’s growth and innovation.
