Tencent's European Roadshow Focus: How AI Drives Growth?

At the Tencent Europe roadshow hosted by Goldman Sachs, Tencent's Chief Strategy Officer James Mitchell showcased how AI technology has become the core engine for growth across the company's various business lines, including enhancing game development efficiency driven by AI and upgrading advertising technology platforms.

On August 25, Goldman Sachs stated in its latest research report that during the Tencent Europe roadshow from August 18 to 22, Tencent's Chief Strategy Officer James Mitchell indicated that Tencent has demonstrated significant cost advantages in AI infrastructure, achieving lower AI inference costs through dynamic model routing, with the mixed Yuan 3D model leading in rankings on Hugging Face.

Goldman Sachs pointed out that key signals released by Tencent's Chief Strategy Officer during the European roadshow also included: AI technology will be deeply integrated into the WeChat ecosystem, while driving improvements in game development efficiency, advertising conversion rates, and high-quality growth in cloud services. Goldman Sachs predicts that Tencent's revenue and earnings per share will grow by 13% and 18%, respectively, by 2025.

Based on this, Goldman Sachs' analyst team emphasized that the comprehensive empowerment of AI technology provides a long-term growth runway for various business lines, reinforcing Tencent's position as a beneficiary of AI applications in China's internet sector. The firm maintains a buy rating on Tencent with a target price of HKD 701.

AI Infrastructure Investment Shows Significant Cost Efficiency Advantages

Goldman Sachs stated that Tencent has demonstrated significant cost advantages in AI infrastructure. The company employs a dynamic model routing strategy, directing queries to models at different cost levels—from deep inference models to Turbo models and then to refined models.

In the search business, most queries are now processed through the Turbo model, which can also run on the lower-cost refined model. In the long term, the shift towards a combination of lower-cost models will further support the improvement of inference cost efficiency.

In terms of GPU configuration, the company primarily uses overseas chips, with the remainder being domestic chips.

Management expressed no concerns about chip production capacity, stating that internal demand for computing power is sufficient, and the company can also utilize domestic chips for inference. This year's capital expenditure targets will depend on chip supply conditions.

Mixed Yuan Model Leads Multimodal Development

Tencent's self-developed mixed Yuan 3D model has achieved a leading ranking on the Hugging Face platform, with research and development focusing on areas ranging from text-based models to multimodal/physical world training.

AI technology has achieved efficiency improvements in several key areas: increasing click-through rates in advertising, accelerating content generation in game development, and significantly improving code generation efficiency.

The company plans to cautiously further integrate AI into the WeChat ecosystem, prioritizing user experience and security. This gradual integration strategy reflects Tencent's emphasis on the stability of its super app ecosystem.

AI-Driven Content Production Revolution

AI has brought significant improvements in content production efficiency in game development. While code can be generated automatically, manual debugging remains important, and the speed of prototype development will further accelerate in the future. For existing games, AI-enhanced virtual teammate features can improve user experience.

In the domestic gaming market, China has 400-500 million monthly active users of application games and 500 million monthly active users of mini-games, totaling 700 million users participating in application games and mini-games Management has seen the potential to convert mini-game users into application game users.

For the mobile game "Valorant," the user base and user experience are the focus in the first year of operation, with monetization being a gradual process. The company aims to extend the game's lifecycle by leveraging the strong fan base on the PC platform and dedicated mobile esports events.

The international gaming business benefits from the gradual confirmation of strong revenue in 2024 and new content updates for evergreen games. Supercell's previous restructuring has reactivated several evergreen IPs, such as "Brawl Stars" and "Clash Royale."

AI Upgrades Drive Advertising Growth

Tencent recently upgraded its advertising foundational model to drive better conversion results, mainly reflected in two aspects:

Content consumption analysis (how users browse/what articles they like) and commercial behavior analysis (the speed at which users click on ads).

The AI-driven advertising technology upgrade platform can drive advertising growth that exceeds industry averages by improving the click-through rates of Tencent's multiple advertising inventories, offsetting the depreciation effects related to AI.

In the long term, the company expects AI to create monetizable intent recommendation/agent prompt functions.

In terms of cloud business, Tencent is shifting towards value-added software/inference efficiency. The company believes that Tencent Cloud's profitability in the domestic market is healthy, and there is still room to accelerate revenue growth on a higher quality basis based on computing capacity.

Goldman Sachs' analyst team believes that Tencent's unique WeChat ecosystem and global gaming assets, combined with multiple monetization levers, will enable the company to achieve compound revenue growth throughout the macro cycle