For two decades, 360doc stood as a quiet monolith of the Chinese internet, serving as a 'personal digital library' for over 100 million users. Launched in 2005 by founder Cai Zhi, the platform allowed users to curate, store, and organize over 1.1 billion articles, creating a massive repository of human-selected knowledge. However, the company has announced that its servers will go dark on May 1st, 2026, marking the end of an era for the Web 2.0 generation.
The decline of 360doc is a casualty of the shift from search-driven information gathering to AI-generated answers. During its peak in 2017, the company generated nearly 40 million yuan in annual revenue, largely fueled by advertising and traffic from search engines like Baidu. In recent years, that traffic has evaporated as users migrate toward AI agents like ChatGPT, DeepSeek, and Kimi, which provide direct solutions rather than lists of archived documents.
Financial reports reveal a stark downward spiral, with net profits plummeting into negative territory and revenue nearly halving in the last year alone. Founder Cai Zhi attempted to offload the platform’s assets—including its core technology and massive data pool—to a strategic partner for free. This desperate gambit failed because potential buyers were more interested in 'stripping the data for AI training' rather than maintaining the service for its loyal community.
Cai Zhi reflects on this transition with a mix of pragmatism and melancholy, noting that her team lacked the resources to bridge the gap into the AI age. She described the 1.1 billion articles on the platform as 'human native data'—information curated by people before the internet was saturated with AI-generated content. In today’s market, she lamented, this raw data is often perceived as more valuable for training models than the users themselves are for generating revenue.
The closure of 360doc signals a broader industry pivot toward 'Personal AI Assistants' or 'Second Brains.' Tech giants like Google with NotebookLM and Microsoft with Copilot are redefining knowledge management from passive storage to active synthesis. As static libraries vanish, the future belongs to decentralized, privacy-focused AI agents that do not just store information, but understand, remember, and execute tasks on behalf of their users.