China's Daily AI Token Consumption Surges 300-Fold Amid Rapid Application Growth
China has experienced a dramatic increase in artificial intelligence activity, with daily token consumption rising more than 300-fold over the past 18 months. Liu Liehong, head of the National Data Bureau, announced at a State Council press conference that daily token usage escalated from 100 billion at the beginning of 2024 to over 30 trillion by June. This surge underscores the rapid expansion of AI applications across the country. Alongside increased computational demand, significant progress has been made in developing high-quality datasets. China has compiled over 35,000 datasets totaling more than 400 petabytes. The growing need for data in AI model training has stimulated the data trading market, with cumulative transactions reaching nearly 4 billion yuan. The Beijing Data Exchange has emerged as a dominant platform, handling nearly 80% of these trades, a substantial increase from 10% the previous year. Furthermore, Chinese-language data has become central to domestic model development, with most local AI models now trained on datasets comprising 60% to 80% Chinese content. These developments highlight China's strategic focus on building robust data infrastructure and leveraging native language resources to advance its artificial intelligence capabilities and support the tech sector's foundational cloud services.
Wire timeline
China's Daily AI Token Consumption Surges 300-Fold Amid Rapid Application Growth
China has experienced a dramatic increase in artificial intelligence activity, with daily token consumption rising more than 300-fold over the past 18 months. Liu Liehong, head of the National Data Bureau, announced at a State Council press conference that daily token usage escalated from 100 billion at the beginning of 2024 to over 30 trillion by June. This surge underscores the rapid expansion of AI applications across the country. Alongside increased computational demand, significant progress has been made in developing high-quality datasets. China has compiled over 35,000 datasets totaling more than 400 petabytes. The growing need for data in AI model training has stimulated the data trading market, with cumulative transactions reaching nearly 4 billion yuan. The Beijing Data Exchange has emerged as a dominant platform, handling nearly 80% of these trades, a substantial increase from 10% the previous year. Furthermore, Chinese-language data has become central to domestic model development, with most local AI models now trained on datasets comprising 60% to 80% Chinese content. These developments highlight China's strategic focus on building robust data infrastructure and leveraging native language resources to advance its artificial intelligence capabilities and support the tech sector's foundational cloud services.
TechNode