The developments that have emerged over the past two years are truly remarkable. Due to the surge in global artificial intelligence applications, we can see that token consumption has increased nearly 300-fold. AI usage is accelerating in every area, from text generation to video processing and smart shopping experiences.

Behind this growth lies tremendous computing power. At present, more than 30 million AI agents worldwide are working together, and this number is increasing every day. China’s domestically developed large language models are seeing a significant rise, especially in API calls.

To understand how important energy has become, it is enough to look at electricity consumption data. The electricity produced by China in 2024 has reached roughly the 10 trillion kilowatt-hours level, surpassing the combined output of Amerika and the European Union. How valuable kilowatt-hours of electricity have become is reflected in figures that now exceed the billions. This is not just a number, but also an indicator of industrial strength and serious investment in AI development.

China’s strategy is quite interesting. Thanks to low-cost token production, it continues to ensure a steady supply for the global market despite the United States’ tariff barriers. Models such as Zhipu’s GLM Coding Plan and Kimi K2.5 are rapidly gaining popularity. This looks like a digital version of the global expansion previously enabled by cheap labor and textile exports.

Currently, China is playing a kind of subsidizing role for Silicon Valley’s application layer. With low-cost tokens, new AI applications can be developed more easily, and users can access services at lower costs. This dynamic is likely to play an important role in shaping the global AI market in the coming years.