Is Tencent Yuanbao about to overtake on a curve?

Author | Huang Yu

Editor | Zhou Zhiyu

As the first AI-native app under Tencent, the usually "laid-back" Yuanbao has started to actively attract users after the explosive popularity of DeepSeek.

From integrating the DeepSeek large model to the announcement on February 18 that Tencent would utilize Yuanbao to support WeChat search, Yuanbao's ambition to achieve a competitive edge in the fiercely competitive AI application market through a combination of "technical complementarity + traffic synergy" is becoming increasingly evident.

According to Wall Street Insights from Tencent, WeChat search has recently begun a gray-scale test of AI search, and user enthusiasm for the DeepSeek model has far exceeded expectations. To allow more users to conveniently and stably experience and use DeepSeek, Tencent has urgently coordinated and deployed internally to utilize Yuanbao to support WeChat search, providing users with more entry points and choices.

In other words, although the number of users currently covered by the gray-scale test of WeChat AI search is still very small, the call volume has still resulted in significant computational resource consumption. With over 1.3 billion monthly active users, WeChat clearly cannot meet the AI search demand of all users in the short term, so it can only choose to allocate massive traffic to its "brother app."

According to Wall Street Insights, starting from February 18, users who have not yet been included in the gray-scale AI search test will have the opportunity to see "Download Yuanbao (Experience DeepSeek-R1)" when they click on the search bar at the top of the WeChat dialogue box and enter WeChat search.

As the core entry point of Tencent's ecosystem, WeChat's traffic allocation strategy has always been cautious, focusing more on optimizing its own functions rather than actively directing traffic to other applications.

This action of diverting traffic to download Yuanbao through the WeChat search entry reflects a significant shift in Tencent's AI strategy, marking that Yuanbao will more proactively attract users and indicating Tencent's ambition in C-end AI scenarios.

More than two years after ChatGPT ignited the AI large model craze, it is evident that the battlefield in the AI field has shifted from foundational technology research to application implementation.

Insiders at Tencent told Wall Street Insights that Tencent is continuously promoting product layout and upgrades in the AI era through efficient organizational changes.

At Tencent's annual meeting earlier this year, Ma Huateng revealed that the TEG (Technical Engineering Group) responsible for developing Tencent's mixed large model underwent structural adjustments to focus more on building the technical foundation, while productization is hoped to be advanced by other business groups.

According to Wall Street Insights, in January this year, Yuanbao was split from TEG to CSIG (Cloud and Smart Industry Group), with the Tencent Meeting team in charge. This move allows Yuanbao to focus more on market implementation, while TEG concentrates on enhancing model performance.

Additionally, Tencent Meeting is currently Tencent's flagship product for B2B, and handing Yuanbao over to its team also demonstrates Tencent's emphasis on the commercialization of Yuanbao.

Since the beginning of this year, Tencent's AI strategy has been accelerating, and a new product matrix is gradually emerging According to information obtained from Tencent internally by Wall Street News, following the transfer of Tencent Yuanbao from TEG to CSIG, more products and applications such as QQ Browser, Sogou Input Method, and ima will also be integrated into CSIG, collectively forming a new product portfolio for Tencent in the era of large models.

Correspondingly, a series of organizational adjustments are underway, with the teams and organizations for products like QQ Browser, Sogou Input Method, and ima being shifted from PCG (Platform and Content Group) to CSIG.

After the Spring Festival, all of Tencent's products are firmly and agilely advancing towards intelligent upgrades. Yuanbao, ima, QQ Browser, Tencent Docs, Tencent Maps, QQ Music, and others have announced simultaneous support for the "Hunyuan Large Model and DeepSeek-R1 Model" dual-engine, while WeChat Search has also launched the "AI Search" feature and integrated DeepSeek-R1, with Sogou Input Method also under evaluation and planning.

Tencent's strategy of choosing the "Hunyuan + DeepSeek" dual-engine has emerged, and among the AI product matrix it is building, Yuanbao's position as a pioneer of future AI products is undoubtedly established.

Tencent hopes that Yuanbao can become an "all-purpose AI assistant" in everyone's pocket, while from the perspective of AI search capabilities, Yuanbao also acts as a connector, linking all of Tencent's internal ecosystems, including WeChat Official Accounts and WeChat Work.

However, unlike many large companies competing for users and traffic, Tencent Yuanbao, which launched in May last year, does not seem to be in a hurry to attract users. When it first launched, Liu Yuhong, Vice President of Tencent Cloud and head of Tencent's Hunyuan Large Model, revealed that user acquisition through advertising was not the main goal for Yuanbao at this stage of development.

Against this backdrop, the user scale of Tencent Yuanbao has somewhat lagged behind. According to QuestMobile data, as of December last year, Doubao, a product under ByteDance, had already become the AI-native application with the highest monthly active users in China, reaching 75.23 million, while Tencent Yuanbao only had 2.11 million, ranking eighth.

Now, with the model equality brought by DeepSeek, the entire AI industry has been reshuffled, making the competition for AIGC products seem to return to the same starting line, and the so-called first-mover advantage appears to be gone. With Tencent's super traffic pool, Yuanbao now has the opportunity for a turnaround