Meituan CEO Wang Xing has detailed the company's upcoming artificial intelligence strategy, stressing that investments will be made within its financial means and not exceed its capital capabilities.

During a recent shareholder meeting, Wang described AI as a positive productivity tool. He stated that Meituan will invest in the technology to the extent it can, but will avoid irrational investments that strain its finances.

This stance signals a shift in the company's AI approach, moving from early-stage capital deployment and technical exploration to a phase focused on capital efficiency, business synergy, and return on investment.

Over the past two years, Meituan's external AI investments have concentrated on two core areas: foundational large language model architecture and embodied intelligence. As related portfolio companies progress toward public listings, these early strategic investments are gradually transforming into financial assets with monetization potential.

In the embodied intelligence field, Meituan participated in multiple funding rounds for Unitree Robotics. Following its IPO approval on China's STAR Market in June, Unitree's post-issue valuation reached approximately 42 billion yuan. Public data shows that Meituan-affiliated capital holds a 9.65% stake, making it the largest external institutional shareholder. Beyond financial returns, this investment is strategically aimed at building a future automated delivery network.

In the large model sector, Meituan holds a 3.86% stake in Zhipu AI. After Zhipu AI's listing in Hong Kong in January, its market capitalization once reached a high of 650 billion Hong Kong dollars. As Meituan's management emphasizes improving capital efficiency, the appreciation of this equity stake represents a potential source of non-recurring gains.

Regarding in-house R&D and product application, Meituan's efforts show clear boundaries. The company is avoiding the high-energy-consumption race to develop foundational large language models and is instead concentrating R&D resources on the application layer and automated toolchains.

On June 9, the AI-native browser "Tabbit 1.0," developed by Meituan's Guangnian Zhiwai team, was officially launched. Instead of relying solely on a proprietary model, the product integrates multiple models, including Meituan's own LongCat model and third-party mainstream models like DeepSeek, Zhipu GLM, and Kimi. It is positioned as a cross-software, cross-webpage task execution portal.

Data indicates that Tabbit's agent task execution success rate has improved from 53.1% during its public beta in March to 91.8% currently. To control operational costs, the product uses a tiered pricing model: a standard free version, with advanced automation features and scenario customization offered through a professional version priced at 9.9 yuan per week, helping to alleviate computing power cost pressures.

Furthermore, within its core local services business, Meituan is advancing low-cost AI pilot tests. This includes launching delivery skills that connect to various third-party AI assistants and deploying its "Xiaomei" AI agent on platforms like Tencent's Yuanbao.

Wang also introduced that Meituan is piloting the "Xiaotuan" feature within its main app, serving as a direct AI entry point for local services scenarios.

When assessing initial data for this feature, Wang offered a measured evaluation, acknowledging that "explosive results have not been seen yet." However, he also predicted an evolution in user interaction methods, stating, "I believe typing will become less common, and voice usage will increase."

This outlook suggests that the focus of future iterations for Meituan's consumer-facing AI products will likely tilt toward the deep integration of voice commands and location-based service scenarios, aiming to find the most natural entry point within the existing order and transaction process.

For Meituan, its core business model is built on a high-frequency, low-margin offline transaction and fulfillment network. Unlike purely online content distribution platforms, large AI models are unlikely to generate disruptive revenue growth for its main businesses like food delivery and in-store services in the short term.

At a time when the local services sector faces multi-front competition and core businesses are in a battle for market share, Wang's restrained approach to AI investment fundamentally clarifies the company's strategic priorities. This capital discipline, which strictly confines AI to the role of an efficiency tool with cash flow safety as the bottom line, also reflects the practical choices major Chinese internet giants are making in the face of new technology cycles.