#Gate广场AI测评官

Oh my goodness! China's daily Token calls exceed 1.4 quadrillion—has the era of AI for all begun?

I just saw a news report where Liu Liehong, director of the National Data Administration, introduced today (the 24th) that by the end of 2025, China has built over 100,000 high-quality datasets with a total volume exceeding 890PB (a computer storage capacity unit), equivalent to approximately 310 times the total digital resources of China National Library. By March of this year, China's daily average Token (词元) calls exceeded 1.4 quadrillion, representing a growth of over 1000 times compared to 100 billion at the beginning of 2024, and compared to 1 quadrillion at the end of 2025, growing by more than 40% in just three months. What does this mean for us? Let's talk about it.

🙎 First, let me clarify a piece of knowledge that everyone might be confused about but feels embarrassed to ask—what is a Token?

The Token here is different from the "token" or "cryptocurrency" we commonly think of. In the world of AI, Token is the smallest unit of measurement for how a model processes information. Whether it's user queries or a piece of code generated by AI, everything ultimately needs to be broken down into Tokens to complete calculations. Because of this, Token call volume has become a key indicator for measuring AI model activity and industry value. The higher the Token call volume, the more the model is being used, and the greater the actual value it creates. Simultaneously with releasing the data, the government has also given Token a Chinese name—"词元" (word element).

🔍 So what does this data mean?

Behind the 1.4 quadrillion Tokens lies an important achievement of China's data elements market allocation reform. In the past, high-quality data was locked away in "information silos" of governments, universities, and enterprises, causing large models to "struggle without sufficient resources." However, as data confirmation of rights, pricing, and transaction mechanisms improve, data assets are beginning to flow freely. The flow of data creates a flywheel effect of "data becoming more active the more it's used, models becoming smarter the more they're used." Liu Liehong, director of the National Data Administration, stated: "1.4 quadrillion Tokens is a 'thermometer' of data supply and value release." This marks that China has successfully mapped out the transformation pathway from data as "resources" to "assets" to "capital," and the value of data elements is accelerating its release. More importantly, Chinese AI is transitioning from "technological following" to "ecosystem leadership." From DeepSeek's open-source model going viral to Seedance2.0 becoming a global hit, to OpenClaw sparking a "lobster farming" craze, Chinese AI enterprises are forming their own technical standards and application ecosystems.

💡 Envisioning the future—how cool would life in an era of "universal lobster farming" and universal AI be!

The rapid growth in daily Token calls might also be inseparable from the "lobster craze" at the beginning of this year. Next, we may enter an era of universal "lobster farming" and universal "AI." Imagine this: in the future, when your "lobster" is "well-raised" (having fully understood your life patterns), the moment you wake up and open your phone each morning, you'll discover that your "lobster" has already ordered your breakfast delivery, replied to morning greetings from friends in social groups, and prepared your daily curated news. While you're brushing your teeth and washing your face listening to your "lobster" read the news aloud, it's simultaneously working tirelessly in the background ordering your Didi ride—it's as if you've hired a personal assistant that never disrupts your daily life. How cool would that be! Let's witness the future together!