Tencent breaks records: AI offensive and jump in sales in gaming!

Tencent breaks records: AI offensive and jump in sales in gaming!

Tencent exceeded analysts' expectations in the second quarter of 2025, as boerse-express.com reported. The company's revenue rose 15 percent to 184.5 billion yuan, while net profit rose 17 percent to 55.6 billion yuan. These impressive results were driven primarily by growth in domestic gaming business, which increased by 17 percent, as well as international growth of 35 percent in this segment.

Another key to Tencent's success is a massive increase in AI investments, which have increased by 119 percent. By integrating AI features into the popular messaging app WeChat, powered by the popular chatbot Yuanbao, the company plans to further revolutionize its services. Tencent also has sufficient GPU inventory for further AI development, and advances in software optimization are significantly reducing hardware requirements. The company's current AI initiatives include the introduction of the multimodal Hunyuan Large Vision model and massive investments in server infrastructure.

Challenges and risks

Despite its current success, Tencent faces various challenges. These include a $76.9 million fine for distributing illegal content and the impact of US sanctions. The company also faces strict gaming restrictions in China. These factors could potentially affect the Company's future performance.

In addition, Tencent shares are currently trading just below their 52-week high of EUR 65.54. The Relative Strength Index (RSI) is at 74.9, indicating overbought territory. This suggests that the sustainability of growth will be crucial in the coming quarters.

Global perspective and data issues

In the context of the growing AI developments worldwide, the scaling laws of American tech companies are also being discussed. In an article by theregister.com addresses the inefficiency of using larger computers to solve mathematical problems compared to more intelligent approaches. These smarter methods lead to more insights and solutions, while the effects of old Soviet mathematics are still visible today.

The analysis indicates that the latest AI models often require more data than is available. Much of this data is already generated by AI models, which could affect the quality of the training data. Comparing AI-generated data to fast food and hard alcohol highlights concerns about data quality. The limitations highlighted there could in turn promote the development of its own hardware and more efficient software in China.