For four days in late April, Yu Hao, the firebrand founder of home-appliance giant Dreame, launched a scorched-earth critique against Xiaohongshu. He labeled the lifestyle platform’s values “toxic” and its anonymous mechanisms a breeding ground for “evil.” Yet, when Xiaohongshu finally broke its silence, it didn't offer a rebuttal to Yu. Instead, it unveiled the most radical organizational restructuring in its thirteen-year history.
The overhaul signals a profound shift in power. Konan, the platform’s long-time product lead, has been elevated to President, centralizing control over the community, e-commerce, and commercialization. Simultaneously, a top-level AI department dubbed “Dots” was established, while the company’s international arm, Rednote, was spun off into an independent unit. Most tellingly, CEO Charlwin (Xing Shi) has receded to the strategic layer, leaving the day-to-day battle for survival to this newly unified command.
To understand this shift, one must recognize that Xiaohongshu’s core product is not content, but consumer decision-making. It occupies the lucrative “middle ground” of the internet: the space where a user moves from an idle interest to a firm intent to buy. While Douyin excels at mindless scrolling and Taobao at transactional efficiency, Xiaohongshu is where users go to be convinced. For a decade, this “seeding” (zhongcao) culture made the platform indispensable, but it also made it vulnerable.
The platform currently faces a structural trap. Approximately 80% of its 42 billion RMB revenue in 2025 came from advertising. When a platform relies so heavily on ad spend, its algorithms naturally prioritize high-engagement controversy over sober, authentic reviews. A post sparking a 300-comment argument generates more data and ad inventory than a quiet, honest product evaluation. This misalignment is exactly what critics like Yu Hao are tapping into—a system incentivized to favor noise over truth.
However, the greater existential threat is not the anger of brand founders, but the rise of Generative AI. For years, the “Xiaohongshu process” involved a user spending 30 minutes reading dozens of notes to aggregate a decision. Today, an AI can summarize those thousands of data points and provide a personalized recommendation in seconds. By doing so, AI effectively “hollows out” Xiaohongshu’s primary utility, rendering the manual research process obsolete.
The creation of “Dots” and the unification of the platform’s technical and commercial wings represent a defensive pivot. Xiaohongshu is attempting to transform itself from a “seed-planting” community into a “decision-making infrastructure.” By integrating AI deeply into its own corpus of human-generated notes, it hopes to provide the speed of an AI chatbot with the high-trust, human-centric data that general AI lacks.
This restructuring is a race against time. If Xiaohongshu can successfully bridge the gap between discovery and transaction—keeping the user within its ecosystem from the moment of inspiration to the point of purchase—it survives as a commerce titan. If it fails, it risks becoming a mere training set for the very AI models that will eventually replace it as the internet’s primary concierge.