Tencent 2Q26: Core strength provides capacity for AI investment

• Tencent’s core businesses remain resilient. Revenue rose 11% and adjusted net profit increased 9%, supported by 17% growth in domestic games and 22% growth in Marketing Services. Strong underlying cash generation gives Tencent the capacity to fund its AI investment cycle.

• AI investment drove the negative surprise. Quarterly capex reached RMB52.8 billion, around 64% above consensus, while the operating-profit drag from new AI products widened to RMB10.5 billion.

• Tencent's AI strategy is becoming more defined. The higher spending partly reflects catch-up after a cautious start, but resources are now being directed more clearly towards models, enterprise agents, Weixin initiatives and AI-cloud capacity. Management expects further investment to be tied increasingly to demand and returns.

• AI monetisation is beginning to take shape, with uneven progress across different products. Standalone financial disclosure remains limited, so the scale and pace of returns cannot yet be fully assessed.

• A re-rating does not require an immediate AI earnings inflection. Clearer monetisation evidence or greater confidence that capex growth is approaching a peak could improve expectations before the reported profit drag from new AI products begins to narrow.

Tencent's second-quarter results confirmed that its core businesses remain healthy. The share-price correction of 4.5% on the first trading day after results reflected a widening gap between rapidly rising AI investment and the commercial returns investors can currently verify.

Core operations delivered; AI spending drove the concern

Tencent delivered another solid quarter. Revenue rose 11% year on year to RMB204.8 billion, slightly ahead of expectations, while non-IFRS net profit increased 9% to RMB68.4 billion, broadly in line.

Domestic games grew 17% and Marketing Services revenue increased 22%. Value-Added Services, Online Advertising and FinTech & Business Services also beat consensus, showing that the strength was broad-based rather than dependent on one business line.

Table 1: Tencent 2Q financial highlights

Source: Bloomberg Finance L.P., iFAST Compilations.

Data as of 30 June 2026

Capex surprise reflects a clearer—but costlier—AI roadmap

The clearest negative surprise was the scale of AI spending. Capital expenditure reached RMB52.8 billion, around 64% above the pre-results consensus used in our model. Together with compute prepayments, this pushed reported free cash flow to negative RMB13.8 billion. The operating-profit drag from new AI products also widened from RMB8.8 billion in the first quarter to RMB10.5 billion.

Chart 1: Tencent quarterly capital expenditure, 1Q25–2Q26.

Source: Bloomberg Finance L.P., iFAST Compilations.

Data as of 30 June 2026

AI-native businesses are not a new direction for Tencent. What changed in the second quarter was the scale of commitment and the clarity of resource allocation. Tencent is moving from selective experimentation towards scaling the infrastructure behind specific priorities: training larger Hunyuan models, supporting WorkBuddy and CodeBuddy inference, expanding Weixin AI initiatives and meeting external cloud demand. The increase is therefore both catch-up and commitment. It partly addresses gaps left by Tencent's more measured start and earlier chip constraints, while showing that management now has a more defined AI roadmap.

Management describes 2026–2027 as a concentrated build-out rather than a permanently higher annual run rate. Further expansion in inference and cloud capacity should increasingly depend on demand and returns. Tencent may direct more cash towards capital expenditure—and potentially less towards buybacks—if it sees superior returns; otherwise, incremental spending should become more constrained. For investors, the positive development is greater strategic clarity, but the higher spending also raises the burden of proving that the roadmap can generate measurable commercial returns.

Strong core businesses can fund the AI build-out

Tencent’s underlying profitability and cash generation remain strong enough to support the current investment cycle. On the company’s newly disclosed basis, non-IFRS operating margin excluding new AI products increased to 42% in 2Q26 from 39% a year earlier. Operating profit on the same basis rose 19% year on year to RMB86.1 billion. Free cash flow remained positive at RMB37.6 billion after excluding compute-capacity prepayments. These figures show that Tencent's core businesses can finance the AI build-out through internal cash generation.

The widening margin gap nevertheless highlights the opportunity cost: a growing share of core-business profit is being reinvested in new AI products rather than flowing through to reported earnings or shareholder returns.

Excess compute could also be rented or resold through Tencent Cloud if internal demand falls short. This reduces the risk of underused infrastructure, similar to a downside protection.

AI monetisation is visible, but still uneven

Tencent is not starting from zero. AI is already supporting advertising and cloud growth, while standalone products are showing early commercial signals.

Advertising and cloud: the clearest current returns

AI recommendation models and automated advertising tools have improved conversion rates and eCPM (the effective cost per thousand ad impressions, a standard advertising-revenue efficiency metric). Together with higher Video Accounts traffic and advertising inventory, this helped Marketing Services revenue grow 22%.

Cloud revenue growth also accelerated from the high teens in 1Q26 to the low 20s in 2Q26, supported by AI demand, international expansion and improved general-purpose cloud pricing.

These are the strongest signs that Tencent's AI investment is already producing returns.

WorkBuddy: promising validation, but not yet an earnings driver

WorkBuddy matters less for its current earnings contribution than for what it suggests about Tencent’s ability to turn AI technology into a product that users adopt, retain and pay for. Since its launch in March 2026, PC monthly visits have exceeded 20 million, placing it first in China’s desktop AI-native office-agent market. Active-user retention exceeded 60% and paying-user retention surpassed 80%, although Tencent reported these figures jointly with CodeBuddy. Subscription cash collections are accelerating, while gross margin from paying users is approaching Tencent Cloud’s overall level. Together, these indicators provide early evidence that Tencent can simplify complex AI capabilities and translate them into sustained usage.

The WorkBuddy Open Platform, launched on 2 September, seeks to compound this product advantage. By allowing third parties to contribute specialised Skills, Experts and Connectors, Tencent can broaden the range of tasks available without developing every capability internally. More use cases could increase engagement and generate richer insight into customer needs, while accumulated workflows and organisational knowledge could embed WorkBuddy more deeply in daily operations. Combined with Tencent’s distribution across WeCom, Tencent Meeting, Tencent Docs and Tencent Cloud, this creates a potential path from product execution to stronger retention and higher switching costs.

Xiaowei and Hunyuan 4: potential catalysts for expectations

Xiaowei, Tencent's Weixin AI agent, remains in small-scale prototype testing. The company plans a phased rollout as its conversational capability, memory, service integration and agent framework mature.

Current forecasts assign little economic value to Xiaowei. Its first disclosed user, task-volume, transaction or monetisation metrics could therefore have a disproportionate effect on market expectations.

Hunyuan 4 offers another proof point. A competitive and on-schedule launch would strengthen confidence in Tencent's model capabilities. A delay or an uncompetitive release would weaken the broader AI-product thesis.

Crucially, the market may not need to wait for the reported AI profit drag to narrow before sentiment improves.

Key risks

• A slower-than-expected recovery in Chinese consumer spending could constrain advertising budgets and in-game spending, causing advertising and gaming revenue to fall short of the growth assumed in this thesis.
• If AI monetisation—whether through advertising-efficiency gains or standalone products such as WorkBuddy—develops more slowly than expected, elevated investment in AI infrastructure and new products could weigh on margins and cash generation for longer than anticipated.
• Renewed regulatory tightening in China’s gaming sector, including restrictions on game approvals, playing time or in-game spending, could disrupt monetisation across Tencent’s gaming portfolio regardless of underlying demand.
• Competition in AI is intensifying as Alibaba, ByteDance and global developers invest heavily in agentic and enterprise products. If competitors narrow the product gap before WorkBuddy establishes deeper ecosystem integration and monetisation, its early lead could erode and weaken the broader AI re-rating thesis.

Investment view: a timing mismatch, not a broken thesis

Tencent's challenge is not the absence of AI returns, but the gap between rapidly rising investment and the revenue investors can currently verify. We forecast adjusted earnings growth of around 6% in both 2026 and 2027, after incorporating higher capex and the larger drag from new AI products.

We reiterate our constructive view. Tencent's 2026 investment surge partly reflects catch-up after a cautious start, but it also marks a more defined allocation of resources across models, AI-native products and cloud infrastructure. Its core businesses remain strong enough to fund this build-out, while early monetisation evidence suggests the strategy has credible commercial potential.

We therefore view Tencent as a medium- to long-term opportunity, rather than a short-term AI earnings-inflection trade. Its ecosystem — spanning Weixin, WeCom, Tencent Meeting, Tencent Docs and Tencent Cloud — provides a distribution advantage that is difficult for standalone AI challengers to replicate. WorkBuddy's early traction demonstrates a proven ability to package complex AI capability into accessible, widely adopted products. A re-rating could begin before reported AI losses narrow if monetisation becomes more measurable and as capex growth approaches a peak.

Based on our updated earnings estimates, our target price is SGD 11, implying an upside potential of around 54% from current levels.

Potential catalysts include:

• Advertising and cloud maintaining strong growth;
• WorkBuddy disclosing paying-user numbers, deferred revenue or standalone revenue;
• Xiaowei disclosing user adoption, task volumes, GMV or monetisation metrics;
• The net operating-profit drag from new AI products stabilising or narrowing.

Table 2: projections for Tencent (SGX: HTCD)

	2025	2026E	2027E	2028E
EPS (CNY)	27.88	29.57	31.37	34.28
EPS growth	18.60%	6.07%	6.09%	9.28%
PE Ratio	13.54	12.76	12.03	11.01
Upside Potential (Fair PE of 17x)	-	-	-	54%
Target Price (HKD)	-	-	-	670
Target Price (SGD)	-	-	-	11

Source: Bloomberg Finance L.P., iFAST Compilations.

Data as of 9 September 2026.

Declaration

For specific disclosure, at the time of publication of this report, IFPL (via its connected and associated entities) and the analyst who produced this report hold a NIL position in the abovementioned securities.

This research report was prepared with the assistance of artificial intelligence (AI) tools. iFAST Financial Pte Ltd does not rely exclusively on AI for content generation; the content of this report — including all investment theses, ratings, price targets and conclusions — has been independently reviewed and verified by the research analyst(s) to ensure accuracy and professional integrity.