BlockBeats news: On March 3, Yang Chaobin, CEO of Huawei's ICT BG, delivered a keynote speech at the Mobile World Congress, stating that the global daily Token consumption has increased nearly 300-fold over the past two years, fueling AI applications such as text-to-video and intelligent shopping, with over 30 million AI agents now working collaboratively worldwide.
Against the backdrop of Yang Chaobin’s remarks today, global API calls to China’s domestically developed large models are surging, with token exports driving expansion in computing capacity and consequently triggering a sharp rise in electricity demand. Yesterday, Musk retweeted and commented, “Electricity production is the single best metric for measuring industrial capacity.” China’s electricity generation in 2024 reached approximately 10 trillion kilowatt-hours—40% higher than the combined total of the United States and the European Union (27 countries), which stood at around 7 trillion kilowatt-hours. China surpassed the EU in electricity generation around 2007 and overtook the United States around 2010. Electricity generation not only reflects industrial strength but has also become a core metric for computing power, directly determining the upper limits of AI development. Among the three major bottlenecks to AI scaling—chips, electricity, and transformers—global data centers will first face acute electricity shortages.
Today, as the United States continues to impose tariff blockades, physical isolation alone is increasingly ineffective. While the U.S. can restrict logistics, it struggles to limit the steady supply of tokens produced by low-cost electricity to global markets. Recently, Zhipu’s AI programming subscription plan, GLM Coding Plan, sold out immediately upon launch—a rare occurrence for a paid Chinese-developed AI programming model. Kimi K2.5 also rose to the top of the global large model API aggregation platform OpenClaw’s usage ranking within just one week of its release, outperforming major models like Gemini and Claude. Chinese-developed large models are leveraging low prices as a catalyst, combined with high performance to retain users, delivering a dual strike on both price and performance in the AI market.
Beyond the intensifying U.S.-China AI competition, Token globalization fosters mutual benefit and shared development, as China’s low-cost Tokens once again indirectly alleviate U.S. inflation. Just as Chinese products leveraged low-cost labor and textiles to go global in the 2000s, today they rely on low-cost Tokens—a stable supply of just $0.10 per million Tokens, serving as an invisible subsidy underpinning Silicon Valley’s application layer.