中国互联网：AI全面战争-我们对全栈竞争的看法

China Internet: AI Total War - our thoughts on full stack All’s fair in love and AI.One of the more interesting ongoing debates we’ve had withinvestors and industry contacts relates to the potential for AI-enabled services to shakeup the established order at the top of the Internet funnel. In this note we’ve outlined somethoughts on the state of play in China, and implications for our coverage. High level, weexpect the leading Chinese devs to remain fast followers on model development. But we Robin Zhu+852 2123 2659robin.zhu@bernsteinsg.com Charles Gou+852 2123 2618charles.gou@bernsteinsg.com Min-Joo Kang+852 2123 2644minjoo.kang@bernsteinsg.com AI chatbots have looked more like search engines than top-funnel disruptors.The fact Doubao hit 70mn DAUs (per QuestMobile) - and continues to grow rapidly - has focused investor attention on AI chatbot development. While it certainly behoves peersto accelerate development, an average of 10-12 minutes of use over 5-7 sessions a daymakes the AI chatbot collective look more like search engine replacements than a new top-funnel entrance for user traffic. The fact ChatGPT, Gemini, and other top Western leaders Charlie Peng+81 3 6777 6993charlie.peng@bernsteinsg.com Everyone wants to be a super app.WeChat serves as the template for wannabe superapps globally, and we expect “everyone uses WeChat because everyone uses WeChat” toremain a robust moat. Elsewhere, Alibaba’s strategy has pivoted to building a transaction-based ecosystem around Qwen, and its various consumer-facing ecosystems. Bytedancehas put resources into adding music, text-based content, payments, and an input method Alibaba: can agentic services catalyse engagement growth?On Thursday, Alibabaunveiled a series of upgrades to its Qwen app, leveraging Taobao, Alipay, Amap, andFliggy commercial inventory to enable Qwen users to transact in a broader range of ways.Alibaba’s shares had already front-run the event, and a consolidation period feels likely. Butthe announcement represents a positive strategic move in our view. The extent to which the Tencent: AI monetisation success vs. the AI narrative issue.Measured on earningsimpact, we think a solid case can be made that Tencent’s AI monetisation in gaming andads has been among the sector’s most successful to date. But the fact Tencent is perceivedas having fallen behind on AI chatbot development remains a notable sentiment overhang.Our bias is it will take much longer for peers to reprogramme top-funnel engagement than BERNSTEIN TICKER TABLE INVESTMENT IMPLICATIONS We’re generally bullish on the Chinese Internet companies’ ability to remain fast followers on AI, and ability to continue to winshare overseas. Domestically, the shift among China’s largest consumer Internet ecosystems towards becoming super-appecosystems feels like the most notable development to watch. To date, the top AI chatbots have shown engagement trendsthat resemble enhanced search engines rather than true top-funnel disruptors… whether new agentic services can catalyseengagement growth will likely be key to watch in 2026. We see expansion into alternate form factors like smart glasses and AI We remain bullish on Tencent as our top pick in the Internet sector, and on Alibaba as the leader in AI model development.Grumbling about the former’s lack of progress on chatbot development will probably continue until Vinces Yao and the recentlyreshuffled AI team delivers on their promise (and large comp packages). Sentiment headwinds aside, we think Tencent hasplenty of earnings levers to pull to deliver another year of 15-20% core earnings growth. For Alibaba, the latest Qwen updatestruck us as promising, though we’d advise caution after the recent jump in AI sentiment. $100-120 for e-commerce plus AI VALUATION COMPS TABLE DETAILS AI REMAINS THE DOMINANT NARRATIVE IN CHINA INTERNET In our outlook note for the China Internet sector in 2026 (LINK), we argued for a shift in the AI debate - away from superficial debates about AI capex dollars and chatbot MAUs to full stack development, and the ability of agentic AI services to drive realbusiness change. In this note we review the AI ecosystem landscape in China, and try to assess whether currently observablevectors for AI development pose a threat to the established order at the top of the funnel in China Internet. In recent years, we’ve argued on various occasions that LLM development as it stands represents a sustaining innovation forthe largest internet platforms, and that differentiated content and entrenched user intent represent the key to victory longer-term. This view largely persists: programming human users to behave differently can easily take more time and money thanprogramming new AI capabilities - see Alibaba spending RMB55bn and counting to persuade consumers to use a second food Our world view on AI development: consumer use cases matter as much as model development In Exhibit 2 we’ve reproduced a framework we drew u