China's AI Inference Volume Surpasses Training Data for First Time
According to a report by the CCID released in October 2026, China's AI inference data volume has surpassed its training data volume for the first time. As of June 2026, the daily token call volume exceeded 175 trillion, marking a transition from model development to large-scale industrial application. This milestone indicates that artificial intelligence is shifting out of the lab-bound 'model-feeding' phase and entering a broad industrial adoption phase. It highlights the rapid commercial growth and strong market demand for AI integration across traditional manufacturing sectors. The daily token call volume has surged over 1,000 times compared to early 2024. Additionally, domestic computing power and large models have been deployed in over 1,000 application scenarios across 20 industries, including steel and automotive manufacturing.
## BACKGROUND
AI inference refers to the process where a trained model applies its learned patterns to make predictions or process new data in real-world applications, whereas training is the compute-heavy phase of teaching the model using massive datasets. A token is the basic unit of text or information that large language models process and generate, functioning as a measurable and tradable unit in the smart economy.