Summary of Key Points
In the past two years, Tencent has faced criticism for its slow pace and cautious approach in the AI sector, particularly with its large-scale models and consumer-oriented product YuanBao, which has impacted its stock price. However, since the second half of 2025, Tencent has attempted to catch up with industry trends by reorganizing its AI efforts (assigning YuanBao to the team led by Tang Daosheng and bringing in renowned scientist Yao Shunyu), increasing AI investment (a 16% increase in capital expenditure in the first quarter), shifting focus to AI Agents (such as the launch of WorkBuddy), and preparing for the rollout of WeChat AI agents. Nevertheless, Tencent still faces challenges such as insufficient computing power, unclear commercialization strategies, and barriers to internal collaboration. Whether it can establish a strong position in the second half of the AI revolution remains uncertain.
Detailed Analysis
1. Why is Tencent's AI progress considered "slow"? Indeed, there have been signs of lagging behind:
- Late product launches: Its MixNan large-scale model was released half a year after Baidu's, and its consumer AI assistant YuanBao came out a year after ByteDance's DouBao, by which time DouBao had already surpassed 100 million monthly active users.
- Conservative investment: In 2025, Tencent's capital expenditure was only 79.2 billion yuan (a 3% increase year-on-year), while ByteDance planned to invest over 160 billion yuan in AI. Following the earnings report, Tencent's stock price fell by 6%, reflecting investors' skepticism about its commitment to AI.
- Computing power constraints: Tang Daosheng admitted that Tencent lacks sufficient GPUs, with limited resources prioritized for internal needs (MixNan training and WeChat functionality), leaving it unable to meet the demands of external clients. ByteDance can purchase large quantities of GPUs, which has eroded market patience.
In short, while other companies are aggressively investing in AI, Tencent's more frugal and slower approach has raised doubts.
2. The unexpected success of WorkBuddy: A change in strategy to avoid direct competition in the consumer segment
Failing to compete with DouBao in the consumer space, Tencent turned to developing AI Agents for business and productivity tools. WorkBuddy unexpectedly became a hit:
- Unintended success: It originated as an internal tool called CodeBuddy, which was later integrated into a desktop office assistant. Tang Daosheng said he never expected it to become so popular.
- No need to compete for consumer traffic: As a desktop office assistant, WorkBuddy focuses on helping users with tasks like document processing and coding, leveraging Tencent's strengths in business services. Its usage has increased tenfold in the past three months, making it the leading office AI tool in China.
- Agile development: The small, flat-team approach (3-5 members) allowed for rapid iteration and experimentation. This flexibility helped WorkBuddy capitalize on the early AI Agent trend.
This move represents a strategic shift from direct competition to finding new opportunities in the business sector.
3. WeChat AI agents: A potential game-changer, but with implementation challenges
The news of WeChat AI agents caused a significant stock price increase (10.46% daily, with a market value boost of HK$415.8 billion). Why is this such a big deal?
- WeChat's vast ecosystem: With 1.4 billion users, integrating AI agents would provide direct access to a massive audience. Goldman Sachs believes that only with the combined success of WorkBuddy and WeChat AI can Tencent's valuation truly rise.
- Market concerns: However, doubts remain: whether it will be launched on time (regulatory approval is required) and how privacy and security issues will be addressed.
WeChat AI represents a major opportunity for Tencent, but its actual impact is still uncertain.
4. The arrival of Yao Shunyu: Can he solve Tencent's AI challenges?
At the end of 2025, Tencent hired Yao Shunyu (formerly from OpenAI and the creator of the ReAct framework), bringing new perspectives:
- New focus on user experience: Instead of focusing solely on benchmarking, Yao shifted the development approach to prioritize user experience and eliminated unnecessary data.
- Enhanced team collaboration: He promoted cooperation between AI research and product teams, improving efficiency.
- Challenges ahead: Despite his expertise, integrating a top academic researcher into a large company like Tencent comes with its own challenges. Whether Yao can overcome these internal barriers remains to be seen.
Yao Shunyu is a promising addition to Tencent's AI team, but whether he can drive real progress will take time to determine.
5. Ongoing challenges: Insufficient computing power and unclear commercialization
Despite improvements, Tencent still faces significant hurdles:
- Computing power: The lack of GPUs persists, and even self-developed chips may not solve the issue (manufacturing requires factories).
- Commercialization: WorkBuddy is still in the investment phase without set commercial goals, while YuanBao lags behind competitors. Tang Daosheng emphasizes the need to focus on quality before considering profitability.
- Distributed efforts: Managing multiple projects simultaneously (YuanBao, WorkBuddy, WeChat AI) without clear priorities has exposed Tencent's eagerness for quick results.
These issues mean that despite progress, Tencent still has a long way to go in achieving its AI goals.
Conclusion
Tencent is finally taking more aggressive action in AI, but it must overcome challenges related to computing power, commercialization, and internal coordination if it wants to succeed in the second half of the AI revolution. The road ahead is still long and full of obstacles.