Tencent shares drop nearly 11% on AI competition concerns

Tencent Holdings Ltd. has seen its stock drop nearly 11% this year, losing about US$173 billion in market value since reaching a four-year high in October, amid concerns it is falling behind in China’s competitive AI sector.

The company has adopted a cautious AI strategy, focusing on integrating AI into core services like WeChat and gaming, unlike rivals Alibaba and ByteDance, which have invested heavily in AI infrastructure and product launches.

Tencent plans to spend 1 billion yuan (US$145 million) on a Lunar New Year campaign offering red packets to promote AI tools, prompting investor concern over profit margins.

Data shows Tencent’s AI app Yuanbao has only 18.3 million daily active users, significantly fewer than ByteDance’s Doubao and Alibaba’s Qwen, which have over 78 million and 73 million users, respectively.

Some analysts believe Tencent’s large user base could help it catch up if it accelerates AI development, but market sentiment remains cautious.

🔗 Source: Bloomberg

🧠 Food for thought

Implications, context, and why it matters.

Tencent’s AI lag follows a strategic U-turn

Tencent kept spending tight as a finance choice, with capital spending down 24% year-over-year in one recent quarter 1.
Analysts backed that restraint since it helped protect profitability, while rivals like Alibaba Group Holding pledged RMB 380 billion over three years for AI and cloud infrastructure 2.
The lighter investment weakened model performance, with Tencent’s Hunyuan T1 ranking 68th on UC Berkeley’s LMArena benchmarks (a public leaderboard that compares how different AI chatbots perform in head-to-head tests) in mid-December, down from 8th in May for an earlier version 3.
That slide triggered a reset in December 2025, with Tencent moving to hire top researchers after its main AI assistant started offering the open-source DeepSeek model (a freely available AI model that companies can adapt) alongside Tencent’s own Hunyuan model 3.

Consumer app users may be a misleading metric in China’s AI race

Daily active users for apps like Yuanbao can distract from how many China AI firms plan to make money through cloud computing, where companies sell on-demand access to computing power and software over the internet 4.
Many leading tech groups give large language models away at a loss so more businesses adopt their cloud services 4.
Alibaba says 60% to 70% of new cloud demand comes from model inference (running AI models to generate answers), which puts the focus on who supplies the computing backbone 2.
Scale in that backbone can track momentum, with IDC data cited in Caixin/ThinkChina.sg saying ByteDance took nearly half of China’s public cloud model call volume in 2024 2.