On April 17, Sohu.com released an AI assistant named "Xiaohu" on Apple's App Store. The product's description repeatedly emphasizes its role as an "AI search engine." When asked "Who are you?", the assistant identifies itself as an AI search engine for video content developed by the Sohu Video team. As a veteran in China's internet industry with 28 years of operation, Sohu has witnessed both the internet and mobile internet eras. However, many long-standing websites like Sohu face a strategic disconnect, as user habits have shifted predominantly to smartphones, while their core strengths remain on the PC platform. Although these companies have made efforts in the mobile space, their flagship applications are absent from the top 50, and even the top 200, of the App Store's free rankings.
In terms of scale, Sohu is significantly larger than niche sites like 360doc Personal Library, affording it greater resources to adapt and experiment in emerging fields like AI. The AI sector has gained substantial traction not only in popularity but also in market valuation. Recent listings of AI firms have seen record-breaking market capitalizations, such as Zhipu AI exceeding HK$430 billion, MiniMax over HK$270 billion, Xunce Technology over HK$100 billion, and Biren Technology over HK$120 billion (data as of April 20, 2026). The allure of AI has attracted companies from diverse sectors—gaming, software, home appliances, hardware, automotive, and education—all eager to enter the field. By February 28, 2026, 796 generative AI services had completed filing procedures, with 481 AI applications or features registered.
Compared to internet firms quick to capitalize on trends, Sohu has a more measured approach. During the rise of products like ChatGPT, Gemini, Claude, Grok, and DeepSeek, many competitors rapidly integrated similar technologies, with some founders personally leading the charge, akin to the recent "lobster fever" trend. Sohu had previously launched two AI products: a consumer-facing tool called Simple AI and a B2B AI technology platform, Sohu AI Lab, neither of which had standalone apps. The "Xiaohu" app marks Sohu's first dedicated AI application.
The critical question is whether "Xiaohu" can catch up to established AI assistants like Doubao, Yuanbao, Wenxin, and Kimi, which boast mature user bases. One strategy is to learn from successful predecessors. Operators familiar with platforms like Douyin, Xiaohongshu, Video Accounts, and WeChat Official Accounts often emulate proven models. A successful product indicates existing user demand, providing a blueprint for others. Sohu's late entry in 2026, years after rivals, makes studying successful AI assistants essential.
Analyzing the growth paths of Doubao, Yuanbao, Wenxin, and Kimi reveals three key user acquisition strategies: First, leveraging internal product ecosystems for引流. For instance, Tencent's Yuanbao benefits from promotions across WeChat, QQ, Tencent News, Tencent Video, and QQ Browser. Similarly, Doubao is promoted within ByteDance's ecosystem, including Douyin and Jinri Toutiao. Giants like ByteDance, Tencent, and Baidu possess substantial internal traffic, reducing marketing costs compared to standalone AI products.
Second, aggressive external advertising. AI assistant ads are ubiquitous on ad-supported apps, app stores, news clients, literature sites, and video platforms. While users may find them intrusive, this saturation builds brand awareness.
Third, capitalizing on seasonal marketing opportunities. During the Chinese New Year, Doubao, Yuanbao, and Wenxin invested heavily in campaigns leveraging holiday user engagement. Despite criticism of users exploiting incentives and churning, these efforts boosted app metrics, with Yuanbao and Qianwen temporarily topping the App Store free charts, displacing Doubao. Without such investments, overtaking leaders would be challenging.
If "Xiaohu" adopts similar aggressive user acquisition and investment tactics, it could potentially achieve viral growth. Alternatively, it could identify unmet market needs. China's internet history includes latecomers like NetEase Cloud Music, which gained traction through curated playlists amid competition with Tencent Music Entertainment. Similarly, Jinri Toutiao rose prominence with personalized algorithm-based recommendations amid crowded news app markets.
For "Xiaohu," beyond marketing, product differentiation is crucial. What unique problems does it solve? What are its primary use cases? While most AI assistants pursue a "general-purpose" approach, creating a crowded field, niche applications remain underexplored. If the internet once offered opportunities for small and medium-sized players, the current landscape is less forgiving. Non-giants face an uphill battle, as established AI assistants from major firms enjoy strong branding, product quality, and user bases, rendering smaller competitors easily overlooked.
Can "Xiaohu" emulate the high-profile strategies of Doubao or Yuanbao? Likely not. Sohu's 2025 financial report shows total revenue of $584 million, with marketing services at $60 million and online games at $506 million. After adjustments, the company reported a non-GAAP loss of $51 million. In contrast, giants like Alibaba, JD.com, Tencent, and ByteDance invested nearly RMB 10 billion collectively in 2026 New Year marketing campaigns alone, with Yuanbao allocating RMB 1 billion, Baidu RMB 500 million, Qianwen App RMB 3 billion in user incentives, and Doubao significant sums. Sohu's financial scale precludes such massive marketing expenditures.
Product innovation is equally challenging. Any breakthrough would quickly attract imitation from giants, many of which already operate multiple AI products. ByteDance, known as an "app factory," has over 20 AI tools, while Alibaba boasts several AI offerings. In today's ecosystem, any innovation by large or small firms is closely monitored, increasing the difficulty of distinct advancement. "Xiaohu" must compellingly answer why users should choose it over more established alternatives.
Comments