Senior Analyst on Tencent's AI agent and China's AI landscape (with David Ingles, Yvonne Man)

Asks about the Tencent AI agent report and what fruit it will bear for the business.

David Ingles

Sean Yang

Tencent's AI move is important but they are behind in AI. They need to ramp up hardware, data centers, and hiring. The key question is what applications will merge with AI—e-commerce so far hasn't worked well for others.

We need to dig deeper into what it means for the user, ecosystem, and company.

Asks if the AI agent will pressure margins due to token costs.

Yvonne Man

Sean Yang

Yes, capex is ramping up and model running costs will be high. But at this stage, the market cares more about model capability than costs. China internet AI capex is still small vs global peers.

People should care more about the outcome rather than the costs at this point.

Asks who will be 'China's Anthropic' and what framework to use.

David Ingles

Sean Yang

The US has 3 frontier model providers; China still has 20+ players (Tencent, Alibaba, DeepSeek, etc.). Anthropic is well-positioned for enterprise, but China's enterprise market has been weak. You can't simply say one company is China's Anthropic.

Internet companies have compute advantages; independents are more focused. There will be more competition between these players.

Asks what is needed for the gap between independents and hyperscalers to narrow.

Yvonne Man

Sean Yang

Model capability is the most important thing. Monetization is less important at this stage. We need to track which model can catch up—Chinese companies will update models every month.

Asks for stock recommendations to benefit from this AI wave.

David Ingles

Sean Yang

Baidu has a self-developed GPU and will benefit from China's GPU shortage. Alibaba has clear AI revenue from cloud. Tencent has a low PE and now a free AI option, making it attractive.

These are the three companies that clearly benefit from AI.

Asks what the catalyst will be for these big three to awaken the market.

David Ingles

Sean Yang

There will be a moment—whether a DeepSeek moment or an internet company breakthrough in model capability. We just need to wait. That is the key point.

Similar to 5-7 years ago when video platforms attracted users and stocks went up daily.

SignalTube