The power structure of China's internet is not static; it is continuously reshaped by technology and business models.

Around the year 2000 was the portal era, with Sina, Sohu, and NetEase controlling the information entry points; from 2010 to 2015, the BAT era began, where search, e-commerce, and social media achieved systematic integration of traffic, with Baidu, Alibaba, and Tencent establishing dominant positions. After 2020, the landscape began to shift, with the new generation of companies like ByteDance, Meituan, Pinduoduo, and Didi rapidly rising through algorithms, local life services, and social e-commerce, breaking the monopoly of BAT.

By 2025, ByteDance is expected to reach the top. This is not accidental; rather, it is due to short videos and content streams becoming the new traffic infrastructure, with algorithmic recommendations replacing traditional entry points, thus reconfiguring the way user attention is distributed.

Looking back over the past twenty years, the competitive storyline is very clear: early competition focused on information distribution (portals, search), then shifted to connection and transactions (social, e-commerce, O2O), and today we enter a new stage driven by models and computing power. Baidu's decline reflects the marginalization of traditional search entry points; Meituan, JD.com, and Pinduoduo can still maintain their positions at the forefront because they are deeply embedded in the industrial chain, addressing real-life and transaction problems with greater vitality, but this also means long-term low margins and high-intensity operations.

The ones truly crossing the cycle are still Alibaba and Tencent. For twenty years, they have relied on ecosystem and infrastructure capabilities to remain at the top. The emergence of SHEIN at the tail end of the list sends a clear signal: the next phase of growth for China's internet is shifting from local competition to global expansion.

If the past 20 years were defined by "traffic is king" and "mobile is king," the next 10 years are likely to be characterized by "computing power and models are king." Within this framework, the infrastructure advantages of Alibaba Cloud and Tencent Cloud will continue to support their leading positions; ByteDance, with its algorithms, content, and globalization capabilities, will deepen its competitive moat.

At the same time, the real variable comes from the AI Native new species that have yet to fully emerge. Just as no one could foresee ByteDance and Meituan in 2010, the AI era will also inevitably give rise to new giants, possibly from embodied intelligence or from globally oriented AI-native applications. History has proven that the internet landscape is never maintained but is always redefined.