Hey there, bargain hunter. Wednesday evening Hong Kong time, Tencent drops its third-quarter results at the board meeting scheduled for November 18, 2026. Consensus EPS sits at about $1.07 per ADR. That alone tells you Wall Street is not expecting fireworks. What it is expecting are the two numbers management has been deliberately vague about: total AI capital expenditure for the quarter, and whether the advertising acceleration that drove Q2 can survive a Chinese consumer who just posted a 0.4% retail sales gain against forecasts of 0.8%.
Scoreboard
The first half of 2026 was not bad on paper. Interim revenue rose 10% to RMB 401.243 billion, with non-IFRS profit matching that pace. In Q2 specifically, revenue grew 11% to RMB 204.785 billion, marketing services accelerated 22% to RMB 43.6 billion, and domestic game revenue jumped 17%. The problem sat one line lower. Operating profit including new AI products rose only 9%; strip out those products and the number was 19%. The gap between those two figures is the size of Tencent’s AI bet, and it got wider in Q2 when total capital expenditure hit RMB 52.8 billion, up 176% year-over-year.
Wednesday’s Q3 figure will tell you whether that spending curve is flattening or still climbing.
What Actually Happened
China’s macro backdrop got worse between August and now. October retail sales rose just 0.4% year-on-year, missing the 0.8% forecast and down sharply from September’s 1.1%. Industrial output grew 4.8%, below the 5.3% consensus. Fixed-asset investment fell further than expected. The October data covered a month that included the week-long National Day holiday, which should have provided a lift. It did not. One ING economist described 2026 as among the weakest years for domestic demand in recent history.
That context matters for Tencent’s advertising line. The Q2 surge was real, but it was partly driven by AI-enhanced targeting lifting effective cost-per-thousand metrics on WeChat. If the consumer stays this soft, advertiser budgets get trimmed regardless of how good the algorithm is.
The Business
Tencent earns money three ways: games (domestic and international), marketing services built on WeChat and Video Accounts, and a FinTech and business services segment anchored by payments and cloud. The first half of 2026 showed games and advertising both re-accelerating while cloud got a secondary boost from enterprises buying AI capacity. R&D expenses in Q2 alone rose 35% to RMB 27.2 billion to support the Hunyuan model and AI products. Management’s own framing: excluding standalone AI products, the core business grew operating profit at 19%. That is the number to hold onto.
The tension between a booming AI infrastructure line and a softening ad engine is not unique to Tencent. Baidu’s GPU cloud revenue surged while its online marketing business fell 19% on the same weak Chinese consumer backdrop.
Is It Cheap?
At roughly 14x trailing earnings, Tencent is priced well below its long-run average, which many broker and data-provider charts put in the mid-20s. American internet peers trade in a range that starts closer to 25x and runs higher. The discount is real. The question is whether it reflects genuine undervaluation or a permanent adjustment for factors that will not go away: regulatory overhang in Beijing, geopolitical risk on the OTC listing, and an AI spending cycle that has already pushed free cash flow negative.
In Q2, free cash flow was negative RMB 13.8 billion. Exclude compute prepayments and it would have been positive RMB 37.6 billion. That distinction matters. If the prepayments are front-loaded and the capex curve bends in Q3 or Q4, the valuation case tightens quickly. If they keep accelerating, the multiple compression argument gets harder to dismiss.
Bull / Base / Bear
- Bull: Q3 AI capex shows early stabilisation, advertising holds above 18% growth, the board declares a dividend, buybacks continue. The stock re-rates toward 18x as evidence mounts that compute spending is generating measurable returns in ads and cloud.
- Base: Revenue grows roughly 10%, AI spending stays elevated but free cash flow ex-prepayments stays positive, and the dividend decision is deferred. The stock drifts sideways at current multiples.
- Bear: AI capex accelerates again in Q3, advertising softens to single digits on weak consumer data, and management gives no profitability timeline for standalone AI products. The 14x multiple compresses toward 11x.
Action Plan
Wednesday’s call is binary in one specific way: either management gives a credible capex ceiling for 2027 or it does not. If it does, the base case gets a catalyst. If it does not, the market will keep applying a discount it may deserve. For bargain hunters already holding TCEHY, the buyback provides a floor argument and there is no reason to exit before results. For those on the sideline, a quarter-position before the call and the rest contingent on the Q3 capex figure and the advertising growth rate is the cleaner approach than chasing the open Wednesday morning.
Cheap Investor Checklist
- Q3 total capex: is the RMB 52.8 billion Q2 figure rising or falling?
- Marketing services growth: does the 22% rate survive October’s 0.4% retail environment?
- Operating profit ex-AI products: holds above 15% growth or not?
- Dividend decision: declared, deferred, or increased?
- Buyback update: shares repurchased in Q3 versus Q2’s HKD 16.9 billion pace?
- Free cash flow ex-prepayments: does it stay comfortably positive?
- Management capex guidance: any ceiling for 2027 spending?
- Domestic games revenue: sustains the Q2 acceleration or fades?
Bottom Line
If Tencent’s Q3 shows AI capex flattening and advertising staying above 15% growth, the 14x multiple is a genuine bargain relative to any Western peer doing comparable things. If the spending keeps compounding and the consumer ad market softens as the macro data suggests it might, the discount is there for a reason. Wednesday morning answers that question. Watch the capex line before you watch the EPS.
