Tencent Holdings (00700) Report Interpretation

Overview

This conference note summarizes Tencent management’s views on converting substantial AI investment into differentiated products, improving cost economics and broader operating benefits. Goldman Sachs remains Buy rated, with a HK$670 12-month SOTP-based target price.

Core views

Tencent management reiterated that AI is its highest capital-allocation priority. Capex plus prepayments are expected to remain meaningful in 3Q before potentially moderating in later quarters. Management characterized this as deliberate front-loading of infrastructure build-out and prepayments: it aims to lock in favorable memory prices and compute economics amid year-to-date memory-price inflation, while supporting both Hunyuan model development and commercialization of its harness products. The stated objective is long-term competitive positioning rather than a short-term reduction in spending. Management cited Hunyuan 4 preview’s performance relative to much larger models as support for three elements of its model strategy. First, it believes scaling gains increasingly depend on the complete optimization stack—including data engineering, post-training and evaluation—rather than pre-training scale alone. Second, real-world usage of WorkBuddy and CodeBuddy can create proprietary feedback loops that improve models over time. Third, Tencent can combine domain knowledge and data advantages in gaming, office productivity and financial services to develop differentiated intelligence. Management’s focus is to connect model development to real workflows; it also argued that intelligence need not face persistent price competition, noting stable user deployment of DeepSeek on Tencent’s harness products despite DeepSeek price increases. For WorkBuddy, management described the product as a harness layer with three reinforcing advantages. Greater adoption can lower incremental serving costs and broaden the skill set, which in turn can attract more users. Real coding and workflow use produces feedback for model improvement, while full-stack co-optimization across models, infrastructure and chips enables workload-specific deployment and efficiency gains. Management views these features as supporting not only eventual monetization but also model performance and long-term cost competitiveness. On Weixin Agent, management said Xiaowei addresses use cases ranging from intensive to casual usage. It expects inference costs to be manageable through a combination of cloud-based model compute and potentially smaller edge-based models, and remains confident in the return on investment as the product scales. Management acknowledged subdued consumer sentiment but highlighted operating levers, including advertising load, that could partly offset the softer macro environment. In games, it sees AI as supporting the Weixin mini-game ecosystem and making virtual-item development more efficient, allowing smaller teams to produce more. AI can also accelerate game-development processes and increase gaming margins. Goldman Sachs maintains a Buy rating on Tencent and a 12-month SOTP-based target price of HK$670. The report lists a HK$453 share price as of the 31 August 2026 close, implying 47.9% upside. Its forecasts show revenue rising from RMB751,766.0 million in 2025 to RMB828,387.4 million in 2026E, RMB902,531.1 million in 2027E and RMB977,283.7 million in 2028E; EBITDA rises from RMB350,662.0 million to RMB367,319.8 million, RMB424,300.0 million and RMB493,105.3 million over the same periods. EPS is forecast at RMB29.73 in 2026E, RMB31.19 in 2027E and RMB34.42 in 2028E, after RMB28.09 in 2025.

Analysis framework

The note relays management commentary from the Asia Leaders Conference and connects investment priorities to model differentiation, product-scale economics and operating applications. Goldman Sachs values Tencent using a sum-of-the-parts approach and presents forward financial estimates alongside the rating and target price.

Methodology notes

• Valuation methodsSOTP (Sum-of-the-Parts) Valuation

  12-month sum-of-the-parts target-price valuation

  A sum-of-the-parts valuation values distinct parts of a business separately and combines them. Goldman Sachs uses it to set Tencent’s HK$670 12-month target price.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Tencent Holdings (0700.HK / 00700.HK)

  Primary covered company; the report links AI investment and product deployment to long-term competitiveness, cost efficiency and operating opportunities.

  Strengths

  AI is the highest capital-allocation priority; Hunyuan, WorkBuddy and domain-specific data and expertise are presented as differentiated capabilities.

  Comparison

  Management said Hunyuan 4 preview achieved comparable performance versus larger models.

  Risks

  AI initiatives may progress below expectations; AI investment may exceed expectations; advertising competition, game-launch or approval delays, slower FinTech and Cloud growth, slower AI development and reinvestment risk could weigh on outcomes.

Key data

• 12-month price targetHK$670.00SOTP-based Goldman Sachs target price.
• Share priceHK$453.00Price as of 31 August 2026 close.
• Implied upside47.9%Reported upside to the target price.
• RevenueRMB828,387.4 million / RMB902,531.1 million / RMB977,283.7 millionGoldman Sachs forecasts for 2026E / 2027E / 2028E; 2025 revenue was RMB751,766.0 million.
• EBITDARMB367,319.8 million / RMB424,300.0 million / RMB493,105.3 millionGoldman Sachs forecasts for 2026E / 2027E / 2028E; 2025 EBITDA was RMB350,662.0 million.
• EPSRMB29.73 / RMB31.19 / RMB34.42Goldman Sachs forecasts for 2026E / 2027E / 2028E; 2025 EPS was RMB28.09.

Impact & implications

The report frames Tencent’s elevated near-term AI spending as an investment intended to secure compute economics and build differentiated model and harness capabilities. Management expects these capabilities to support product monetization, lower long-term costs and operating gains in games and advertising despite a subdued consumer backdrop.

Risks

• AI initiatives may progress below expectations.
• AI-related investment may be higher than expected.
• Competition in performance-based advertising may intensify.
• Game launches or Banhao approvals may be delayed unexpectedly.
• FinTech and Cloud business growth may moderate.
• AI development may progress more slowly than expected.
• Reinvestment risk may materialize.