Meituan's Wang Xing Talks About AI Investment Again, Says No Blind Investment Beyond Financial Capability

```

Author | Wang Xiaojuan

At a recent Meituan shareholders’ meeting, Meituan CEO Wang Xing clearly outlined the company's next-phase AI development strategy. Wang Xing stated that Meituan will consider its cash flow situation and will not blindly invest beyond its capital capabilities.

He emphasized that AI is a proactive productivity tool. Meituan will invest within its capabilities, but will not make irrational investments beyond its financial capacity.

This statement to a certain extent indicates that Meituan’s strategy in the AI field is shifting from early-stage capital deployment and technology trials to a phase guided by capital efficiency, business synergy, and ROI.

Over the past two years, Meituan’s external investments in AI have mainly focused on two core tracks: foundational large model infrastructure and embodied intelligence. As relevant portfolio companies are entering IPO processes, Meituan’s early-stage strategic investments are gradually transforming into financial assets with monetization dividends.

In the embodied intelligence domain, Meituan participated in multiple rounds of financing for Unitree Robotics.

On June 1 this year, Unitree Robotics passed the IPO review on the STAR Market, with a post-issue valuation of about 42 billion yuan. Public data shows that Meituan-related funds hold a 9.65% stake in Unitree Robotics, making it the largest external institutional shareholder. In addition to financial returns, this investment mainly points to Meituan’s long-term reserves for future automated delivery networks at the technical level.

In the large model track, Meituan holds a 3.86% stake in Zhipu AI. After Zhipu AI was listed on the Hong Kong stock exchange in January this year, its market value peaked at 650 billion HKD. Against the backdrop of Meituan management emphasizing improved fund usage efficiency, the book appreciation of these equity assets is becoming a potential incremental non-recurring income for the company.

In terms of self-research and product application, Meituan’s investments show clear boundaries: avoiding participation in high-consumption foundational large language model arms races, and focusing R&D resources on the application layer and automated toolchains.

On June 9 this year, the AI-native browser "Tabbit 1.0", developed by Meituan’s Beyond Lightyear team, was officially launched. The product did not take a purely self-developed single-model path, but instead integrated Meituan's self-developed LongCat model, along with DeepSeek, Zhipu GLM, Kimi, and other major third-party models, positioning itself as a cross-software, cross-web task execution entry point.

Data shows that Tabbit’s Agent task execution success rate has improved from 53.1% during public beta in March this year to 91.8% currently. In terms of controlling operating costs, the product adopts a tiered charging model: the standard version is free, and advanced automation features and scene customization are monetized through a professional version (9.9 yuan/week), easing computing cost pressures.

Additionally, in its main local life business, Meituan is advancing low-cost AI gray-scale tests, including launching errand Skills that connect various third-party AI assistants, and deploying the “Xiaomei” agent on platforms like Tencent Yuanbao.

Moreover, Wang Xing introduced that Meituan has already conducted gray-scale tests for the "Xiaotuan" function within the main app, serving as a direct entry point for AI in local lifestyle scenarios.

Regarding preliminary data for this function, Wang Xing gave a rational assessment, saying, “No explosive results have been seen so far.” But he also predicted the evolution of terminal interaction methods: “I believe typing will decrease, and voice will be used more and more.”

This judgment suggests that Meituan’s future iteration focus for AI products directed at consumers will likely tilt towards a deep integration of voice commands with LBS scenarios, seeking the most natural entry point within the existing order transaction chain.

For Meituan, its core business model is fundamentally built on high-frequency, low-margin offline transactions and fulfillment networks. Unlike purely online content distribution platforms, large AI models are unlikely to generate disruptive revenue increments to its main businesses such as food delivery and in-store services in the short term.

As the local lifestyle track faces multi-line competition and core businesses enter a stock game, Wang Xing’s restraint in AI investment is essentially a clear focus on company strategy. This capital discipline, which strictly limits AI to efficiency tools and sets cash flow safety as the bottom line, also somewhat reflects the current practical choices for domestic Internet giants facing technology cycles.

Risk Warning and DisclaimerThe market has risks and investment must be cautious. This article does not constitute personal investment advice and has not considered the special investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are applicable to their specific circumstances. Invest accordingly at your own risk. ```