Goldman Sachs believes "Tencent's valuation recovery depends on progress in the AI narrative," and the internal testing of WeChat's AI is a key step.

```

WeChat’s super app ecosystem has taken a key step toward the commercialization of intelligent agents, but Goldman Sachs believes that the short-term asymmetry of costs and returns restricts stock performance.

According to Wind Chasing Trading Desk news, on June 23, the Goldman Sachs Ronald Keung team released a research report: WeChat, under Tencent, launched internal testing of its built-in AI assistant "Xiaowei" last weekend, which marks that this super app with the largest user base has officially entered the intelligent agent AI track.

However, Tencent’s stock price fell 1.6% on Monday, underperforming the Hang Seng Index by 0.7 percentage points, with a lukewarm market response.

Goldman Sachs believes that the news of the internal test failed to boost the stock price due to three major core investor concerns: controversy over model selection, potential impact of inference costs cannot be ignored, and unclear path to short-term monetization.

The report clearly points out that the restoration of Tencent’s valuation multiples in the coming quarters will mainly depend on the effectiveness of its AI narrative, with the user penetration of the WeChat AI Assistant being a key variable.

Three Core Concerns Suppress Stock Price

On the first trading day after the news of the internal test (Monday), Tencent stock fell 1.6%, while the Hang Seng Index only dropped 0.7%. Tencent clearly underperformed the market.

This "good news but no rise" phenomenon reflects investors’ deep concerns about three core issues, according to Goldman Sachs.

First, doubts about model selection. Xiaowei uses WeChat’s self-developed WeLM large language model, rather than Tencent Group’s more well-known Hunyuan model.

This means that Tencent is maintaining two independent large model development systems within the group, operating in parallel. Whether the two can coordinate, and whether training costs can eventually be integrated, remains unknown. The market worries this is a duplication of resources.

Second, the potential pressure of inference costs on profits.

Goldman Sachs constructed a bottom-up estimation framework. By evaluating the daily average token consumption of the WeChat AI agent, it estimates that, if fully promoted, the incremental inference cost would be equivalent to about 5% to 17% of Tencent’s projected adjusted operating profit in Q4 2026.

In other words, with very limited incremental revenues in the initial phase, cost pressures may emerge first, materially dragging down WeChat’s profit margins.

Third, short-term monetization paths remain unclear.

Goldman Sachs admits that currently, the main source of incremental long-term revenue will still rely on the existing online advertising market, particularly advertising monetization around transaction scenarios, which requires deep penetration of the WeChat AI agent in local lifestyle services, content discovery, and shopping scenarios.

This process cannot happen overnight, so the market remains cautious about the strength of short-term catalysts.

What Can Xiaowei Actually Do? Its Feature Coverage Is Already Quite Extensive

Despite a lukewarm market response, Xiaowei’s feature completeness is already notable.

According to early feedback from internal test users, Xiaowei can perform all WeChat native functions via voice or text commands: send messages, post to Moments, make calls, and—by invoking mini-programs—complete day-to-day tasks like appointment registration and ordering takeout (users still need to complete payment themselves).

On the content side, Xiaowei combines chatbot knowledge bases with the WeChat content ecosystem, supporting efficient information retrieval (related videos and articles), content generation (note-taking, answering questions), as well as multimodal understanding and generation, such as creating images/posters and generating music playlists.

Notably, Xiaowei also supports natural language generation of mini-programs. Users can describe their requirements in plain text, and Xiaowei will return a prototype for further layout and function adjustments.

Although the generated mini-program can currently be viewed only by the creator, and not shared externally, this capability—once opened up—may activate the potential of AI-powered user-generated content (UGC) within the WeChat ecosystem.

Goldman Sachs’ previous key China AI issues report noted that the current market focus remains on enterprise AI (due to clearer monetization paths), but consumer-side AI agents likewise have the potential to reshape traditional application structures. By occupying the traffic gateway, they trigger strategic competition and collaborative restructuring between OS-level ecosystems and in-app ecosystems.

Within this competitive landscape, Tencent’s WeChat has unique advantages: it has the largest user base in China, rich social scenario data, and a mature mini-program ecosystem. The combination of these three makes it one of the platforms with the most potential for consumer-facing AI agents to land.

The Key to Valuation Recovery: The AI Narrative Must Continue to Materialize

Goldman Sachs points out that the restoration of Tencent’s current valuation multiples depends on whether its AI narrative can continue to be reinforced over the next few quarters. Specific dimensions to watch include:

Whether Tencent can transform from a "latecomer" in foundational models to a "proactive player" across models and capital expenditures;Whether advertising growth can reaccelerate;Whether Tencent Cloud can consolidate its position as one of the top three hyperscale cloud service providers in China;And user adoption progress for WeChat AI Assistant and productivity tools (such as WorkBuddy).

From a fundamentals perspective, the foundation for Q2 performance remains solid. Advertising business continues to grow, and game revenue’s quarter-on-quarter growth rate is expected to be faster than transaction platforms. This is an important basis for Goldman Sachs’ recent upgrade of the games and entertainment sub-sector to the second preferred sub-industry.

Goldman Sachs uses a sum-of-the-parts (SOTP) valuation method to set Tencent’s 12-month target price at HK$700, representing a 61.7% upside from the current price of HK$433 (as of the market close on June 22, 2026).

Of course, Goldman Sachs also lists several potential risks: AI business progress falling short of expectations; AI-related investments exceeding expectations; intensifying competition in performance-based advertising; unexpected delays in new game launches or approval processes; slowdown in fintech and cloud business growth; and overall reinvestment risk.

For investors, the core logic at this point is that Tencent’s stock price has already reflected a pessimistic expectation of AI progress to some extent, while the launch of Xiaowei’s internal testing is itself the "important step in Tencent’s AI roadmap" described by Goldman Sachs. The valuation recovery window exists, but its pace will be determined by the speed at which the AI narrative is substantively realized.

~~~~~~~~~~~~~~~~~~~~~~~~

The above exciting content is from Wind Chasing Trading Desk.

For more detailed interpretations, including real-time analysis and frontline research, please join [Wind Chasing Trading Desk ▪ Annual Membership]

Risk Warning and DisclaimerThe market has risks, investment requires caution. This article does not constitute personal investment advice, nor does it consider the unique investment goals, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article suit their specific situation. Invest accordingly, and you bear your own responsibility. ```