WeChat Agent AI Ecosystem Taking Shape; Tencent Valuation Undervalued

Overview

Bernstein released a flash note on Tencent Holdings, focusing on WeChat's latest developments in the Agentic AI sector. The report argues that through A2A cooperation with handset makers and opening up the Mini Program ecosystem, Tencent is positioning WeChat as the core platform layer for Chinese users to access AI services. Although increased short-term AI investment has triggered market concerns about profitability, the firm judges that the current stock price has overreacted to this negative factor and maintains an Outperform rating with a target price of HKD 780.

Core views

WeChat is building a platform-level entry point for the Agent AI era. Tencent announced opening AI ecosystem access to Mini Program developers and has signed Agent2Agent (A2A) agreements with five major handset makers: Huawei, Honor, Xiaomi, OPPO, and vivo, enabling system-level AI assistants on phones to directly recommend and operate WeChat Mini Programs. These initiatives aim to allow users to access agent services via WeChat even before unlocking their phones, thereby establishing WeChat's position as an AI service platform layer. Competitive barriers stem from irreplaceable high-frequency scenarios. While the A2A protocol is not exclusive, WeChat possesses user stickiness that competitors find hard to replicate, thanks to its average daily open rate of over 40 times and the rigid demand for social messaging. In contrast, while competitors like ByteDance are advancing AI phones, their expansion strategies conflict with the interests of other internet platforms, making it difficult to build a widely trusted multi-platform ecosystem like WeChat. Platforms such as Meituan plan to divert traffic from Tencent Yuanbao AI, and Tencent-backed portfolio companies are expected to be among the first to integrate. The pain period of AI investment is poised for a turning point. The primary bear case currently is that AI spending this year will more than double from the RMB 18 billion base in 2025, raising concerns about sustained earnings drag. However, the research note points out that the technical threshold required to support A2A orchestration is not extremely high, and Tencent's technical capabilities will soon reach a 'good enough' level. As product iterations advance and internal token usage efficiency optimizes, the negative impact of AI investment on earnings is expected to peak. Whether WeChat Agents launch by late June or in the second half of the year, product progress will support the recovery of earnings expectations. Valuation is in a significantly undervalued range. The report believes Tencent shares have been oversold. Excluding listed investment portfolios, its core operating business currently trades at only about 10x 2027E earnings. Based on a 20x FY+1 P/E multiple, the firm sets a target price of HKD 780, corresponding to approximately 75% upside, demonstrating confidence in medium-to-long-term value restoration.

Analysis framework

The report employs a four-dimensional verification framework: 'Product-Competition-Finance-Valuation.' First, it tracks specific actions regarding WeChat's A2A protocol and Mini Program openness from the product side to judge the pace of AI implementation; second, it combines user behavior data (daily active opens) with competitors' strategic contradictions to demonstrate the stability of the platform moat; third, it directly addresses the market's primary concern regarding AI costs, distinguishing between the 'cash-burning phase' and the 'efficiency optimization phase' to predict the earnings inflection point; finally, using sum-of-the-parts valuation thinking, it strips out the noise from investment assets to restore the true valuation multiple of the core business, reaching the conclusion that it is undervalued.

Methodology notes

• Competition & Strategy FrameworkMoat / competitive advantage

  High-frequency rigid-demand entry as a platform moat

  The report emphasizes that WeChat's daily open frequency of over 40 times is the core barrier to becoming an AI Agent platform. In internet analysis, user frequency often determines whether a platform can become an infrastructure-level entry point more than sheer user scale alone, as high-frequency behavior implies lower customer acquisition costs and stronger ecosystem lock-in effects.

• Valuation MethodologyPE/PEG valuation

  Core Business Stripped Valuation Method

  For diversified groups, the report excludes non-core assets such as listed investment portfolios and applies P/E multiples solely to core operating businesses. This method avoids distortion of core business valuation caused by volatility in investment assets and more accurately reflects pricing levels based on actual operating capabilities, commonly used in analyzing holding-company or investment-oriented internet firms.

• Industry/Sector Analysis FrameworkUpstream-Midstream-Downstream Transmission

  Revaluation of A2A Protocol Value as an Ecosystem Connector

  The Agent2Agent protocol is essentially a standard interface enabling interoperability between different AI systems. The report views this as a key variable in WeChat's transition from the 'application layer' down to the 'platform layer.' This analytical perspective prompts investors to focus on how technical standard-setting power reshapes value distribution across the industry chain, rather than just functional updates of a single product.

Asset mapping & comparison

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

• Tencent Holdings (0700.HK)

  Core Beneficiary: WeChat's A2A ecosystem layout directly reinforces platform value; AI product implementation is expected to reverse pessimistic earnings expectations

  Strengths

  Irreplaceable high-frequency WeChat entry; high peer trust facilitates ecosystem integration; core business valuation at historical lows

  Weaknesses

  Short-term AI investment significantly erodes profits; Agent product experience still requires iterative validation

  Comparison

  Compared to competitors like ByteDance, WeChat possesses greater neutrality and a foundation of trust in cross-platform ecosystem coordination

  Risks

  AI inference costs exceeding expectations; macroeconomic and consumption weakness; regulatory policy changes

Key data

• Target Price780.00 HKDBased on 20x FY+1 P/E, implying 75% upside
• 2025 AI Spend BaseRMB 18 Billion2026 AI spend expected to more than double from this base
• WeChat Avg. Daily Opens>40 TimesCore user behavior data supporting its role as an AI Agent platform entry
• Core Biz Valuation Multiple~10x 2027E EarningsCurrent trading level of core operating business excluding listed investments
• 2026E Adj. EPS30.00 CNY2025A was 28.09 CNY; 2027E forecast is 34.91 CNY

Impact & implications

If WeChat successfully establishes its Agent AI platform status, Tencent will upgrade from a pure traffic distributor to an infrastructure provider of intelligent services. This would not only bring new monetization funnels to the Mini Program ecosystem but also consolidate its hub position in the internet industry through a 'benevolent gatekeeper' role. For investors, the current low valuation provides a high margin of safety, and any substantive progress in AI product implementation could serve as a dual catalyst for upward earnings revisions and valuation recovery.

Risks

• Macroeconomic risks, including credit environment and retail consumption weakness
• Fluctuations or decline in platform user engagement
• Gaming and advertising businesses facing competition for demand
• Uncertainties arising from antitrust and other regulatory policies

What to watch

• Iteration speed of WeChat A2A system from v0.1 to seamless experience
• Whether agent orchestration adopts proprietary models or third-party models (e.g., DeepSeek, GLM-5.1)
• Consumer adoption level required to trigger ecosystem partner monetization
• Progress in optimizing internal employee token usage efficiency