Summary of Key Points
Tencent is clearly falling behind ByteDance and Alibaba in the AI competition, with Pony Ma admitting multiple times that the company has "slowed down." To catch up, Tencent hired 28-year-old former OpenAI scientist Yao Shunyu as the Chief AI Scientist, granting him special authority to drive reforms (such as shutting down old laboratories and adjusting model development strategies). However, no impressive results have yet been achieved. Unlike during the mobile internet era when Zhang Xiaolong single-handedly created WeChat, the AI era requires a coordinated effort from an entire organization, including a supportive ecosystem and sufficient financial resources. Although WeChat boasts 1.38 billion active users and a wide range of use cases, Zhang Xiaolong's philosophy of restraint has prevented Tencent from fully investing in AI. If Tencent fails to make progress in AI, it may end up in the same precarious position as the "third-place player" during the mobile internet era.
I. Tencent's AI Progress: Indeed Slowing Down, Under Similar Pressure to 2010
The industry generally acknowledges that ByteDance has DouBao (a popular consumer-facing product) and Alibaba has Tongyi Qianwen (a strong business-oriented solution). Tencent lacks significant influence in three key areas: foundational models (such as large language models), computing power (servers necessary for AI), and consumer apps (similar to DouBao). Pony Ma has publicly acknowledged the company's lag at least twice since 2026, a situation reminiscent of the rise of WeChat in 2010—Tencent almost fell behind at the starting line of the mobile internet revolution.
For example, while ByteDance's DouBao has over 100 million monthly active users and is widely used in business services, Tencent's Hunyuan model was previously focused on competing for high test scores rather than providing tangible value to users. It wasn't until Yao Shunyu joined that Tencent shifted its approach to identifying valuable use cases.
II. Yao Shunyu: A Key Player for Tencent, but Not a Savior
Yao Shunyu is a talented individual who entered Tsinghua University's renowned AI program and later obtained his Ph.D. from Princeton, with experience at OpenAI, where he contributed to the transition from chatbots to AI agents. Tencent has granted him unique authority, allowing him to report directly to President Liu Chiping on strategic matters and oversee both AI infrastructure and large language model development. His first half-year at Tencent included shutting down an outdated AI lab, stopping the excessive focus on model rankings, and reorganizing data review processes. The newly launched Hy3 preview in April this year bears his influence, but its performance has not yet been groundbreaking. Pony Ma described the situation as follows: "We thought we were on board the ship, only to realize it was leaking; now we're on the ship but can't sit comfortably."
Yao Shunyu's book, "The Second Half," highlights the focus on finding practical use cases for AI rather than simply competing in model performance. With WeChat, games, and advertising platforms, Tencent has a strong foundation that could be enhanced by AI. However, Zhang Xiaolong is cautious about overinvesting in AI, fearing it might undermine the success of these core businesses.
III. The Difference Between the AI Era and the Mobile Internet Era: From "Individual Heroes" to "Collective Efforts"
In the mobile internet era, Zhang Xiaolong was a pivotal figure for Tencent. By acquiring Foxmail and providing him with ample freedom (such as setting up a development center in Guangzhou), Tencent was able to succeed with WeChat. Today, AI requires a different approach:
- No individual heroes: AI projects are the result of collective efforts; users don't directly interact with developers, and products aren't tied to specific individuals like WeChat was with Zhang Xiaolong.
- Collective effort needed: AI competition relies on capital, computing power, and collaborative development. Yao Shunyu emphasizes building a balanced ecosystem consisting of foundational technology, valuable products, and cutting-edge research.
Therefore, Tencent cannot rely solely on Yao Shunyu; it must make comprehensive organizational changes and allocate resources to succeed in AI.
IV. WeChat: A Powerful Tool, but Zhang Xiaolong Hesitates to Force AI Integration
WeChat is a significant asset with 1.38 billion active users and valuable data. When rumors of AI integration surfaced in early June, Tencent's stock price soared 10%. However, Zhang Xiaolong is wary of adding a centralized AI component that could disrupt the user experience. WeChat is a daily essential tool, and any mistakes could have a significant impact.
V. The Transition from Old Heroes to New Leaders: Organizations Matter More
In large companies, new leaders often emerge as old ones retire. For example, Liu Chengmin, who helped Tencent go public and profit from mobile value-added services, was initially skeptical about mobile browsers but was persuaded by Pony Ma to develop a precursor to WeChat. Similarly, Lin Junyang's departure from Alibaba did not hinder the company's AI progress. These examples show that organizational priorities are more important than individual achievements.
To win in the AI era, Tencent must make broader changes, such as breaking down departmental barriers and optimizing evaluation systems, to fully leverage its strengths in AI.
Conclusion: A Difficult Path for Tencent
Tencent faces a challenging situation: it has a powerful platform but lacks momentum due to Zhang Xiaolong's caution. While hiring a talented scientist like Yao Shunyu is a step forward, success requires organizational adjustments and sustained investment. If Tencent fails to transform effectively, it may fall behind in the AI race, similar to its past position as a "third-place player" in the mobile internet era.
(The full text ends here.)