Morgan Stanley Maintains Overweight Rating on Tencent with HK$550 Target Price
nashnova research
Morgan Stanley kept its Overweight rating and HK$550 target on Tencent, but trimmed 2026–2027 operating-profit forecasts by 1%–2% — accelerating AI spending and slowing core revenue are squeezing near-term margins simultaneously.
Why hold the rating but cut the profit forecast?
The rating stays Overweight at HK$550, implying a 2027 P/E of 16.4× versus the current 12.7×.
Yet 2026–2027 non-IFRS operating-profit estimates were lowered by 1%–2%, caught between a softer macro and rising AI costs.
This means → Morgan Stanley's message is "buy for the long run, but brace for a rough patch on profits."
What is dragging Q3 revenue?
Q3 2026 revenue is forecast to grow 8.9% year-on-year, down from Q2's 11.0%.
Two forces are at work: a weaker macro dampening overall spending, and a tough year-ago comp in gaming.
Strip out AI, and operating profit still grows 10% YoY; but AI's projected RMB 15 billion quarterly loss pulls the headline number to a 3.1% decline.
In plain terms = the legacy business keeps minting money — AI is the new project burning through it fast enough to flip the combined profit line negative.
Games, ads, cloud — which line hits hardest?
Advertising leads: marketing-services revenue is forecast up 18.0% YoY, driven by better ad tech, more Video Account ad slots, and search ads.
Cloud is close behind: enterprise-services revenue is expected to jump 27% YoY — the fastest-growing segment.
Gaming is steady but slower: online-game revenue up 7.8% YoY, with domestic up 13.0% but international down 3%. *Honor of Kings* topped global mobile-game revenue in Q3, yet high-base effects are setting in.
Fintech growth decelerates to 4.5%, the laggard of the group.
What has all that AI spending produced?
The Hunyuan large model has moved from model layer to platform layer: the Hy4 Preview launched ahead of schedule and beat expectations.
WorkBuddy — an office AI agent — opened its platform in September, pivoting from a standalone tool to an ecosystem for hardware, apps, and developers.
WeChat's AI agent "Xiaowei" is expanding its test base to tens of millions of users. Morgan Stanley sees this as a potential key catalyst.
This means → Tencent's AI play is not just "training models" — it is embedding AI into WeChat's massive user base. If it works, the path to monetization is far shorter than selling cloud compute alone.
Will short-term cash flow turn negative?
Morgan Stanley raised 2026–2027 capex forecasts by 3% and 12% respectively, to RMB 206 billion and RMB 224 billion — mostly for compute procurement and AI prepayments.
The bank warns that near-term free cash flow could turn negative.
But it expects any shortfall to be covered by debt financing, not equity dilution; core-business cash flow plus asset disposals should support investment needs.
In plain terms = the wallet will be tight for a while, but Tencent will not need to ask shareholders for money — the legacy "cash machine" is still running.
What does HK$550 actually depend on?
The target is derived from a DCF-based sum-of-the-parts (SoTP) valuation, arriving at HK$550.
The current share price implies a 2027 P/E of 12.7×; the target implies 16.4×. The gap is the "AI premium" the market has not yet awarded.
This reflects Morgan Stanley's central bet: whether AI spending can translate into incremental earnings beyond 2027 is the single verification point for that premium.
市场有风险，内容仅供研究参考，不构成投资建议。