Following the market close on June 1st, Meituan released its financial results for the first quarter of 2026.
In terms of business resilience, the company reported revenue of 91 billion yuan for the quarter, representing a year-on-year increase of 5.6%. Its operating loss decreased to 6.5 billion yuan from 16.1 billion yuan in the previous quarter, marking a sequential reduction of nearly 10 billion yuan and significantly exceeding analyst expectations.
Strategic Shifts and Technological Focus
Since its 2018 listing, the company's investment focus has shifted from broad consumer sectors to hard technology. After formally establishing its "Retail + Technology" strategy in 2019, the company accelerated this pivot, consistently increasing research and development spending. In Q1, its R&D expenditure grew by 22% year-on-year to reach 7 billion yuan.
The outcome of this sustained investment and dedication is that the company has not only navigated intense competition in food delivery but also maintained a robust moat in local life services. Furthermore, by leveraging its self-developed technology, it is systematically advancing the deployment of "Physical AI," accelerating the application of AI to meet real-world, physical demands.
In its external investments, the company has built a portfolio spanning multiple technology fields, including embodied intelligence, semiconductors, large AI models, and autonomous driving.
For instance, the prospectus for Unitree Robotics, which recently passed its IPO review for the STAR Market on June 1st, reveals that a Meituan-affiliated entity is its largest external shareholder, making the investment returns highly noteworthy. It is understood that the company invested in Unitree early on, becoming a lead investor in its Series B2 round and subsequently participating in the Series B3 financing.
From an external perspective, at a time when hard technology is a global investment trend and overseas capital is actively targeting Chinese hard tech assets, the company has positioned itself early within China's hard technology sector.
Delving into Physical AI and Realizing Tangible Value
A closer look at the Q1 report reveals the company's operational strength. Specifically, regarding loss reduction, both the Core Local Commerce and New Initiatives segments saw substantial improvements. The Core Local Commerce segment reported an operating loss of 2 billion yuan, a significant reduction from 10 billion yuan last quarter. Losses in the New Initiatives segment narrowed to 2.1 billion yuan from 4.6 billion yuan.
On the revenue front, not only did overall revenue continue steady growth, but revenue from New Initiatives surged 21% year-on-year to reach 27 billion yuan. Within this, Xiaoxiang Supermarket now covers 55 cities. By offering users high-value products, it has driven rapid revenue growth, becoming a key growth engine for New Initiatives.
Behind this performance, examining the results through the lens of Unit Economics (UE), a core metric balancing subsidy intensity and operational efficiency, provides a clearer view of the company's dominant competitiveness in food delivery. For comparison, some analysts estimate that the company's average loss per food delivery order in Q1 was approximately 1.0 yuan, while a major competitor's was around 3.6 yuan. Other analysts noted that the company's UE advantage over that competitor widened to about 3 yuan per order in Q1 (from 2 yuan in Q4 last year). This data indicates the company's superior efficiency in reducing losses and operational execution.
Drivers of Superior Efficiency
The company's higher efficiency stems from several factors. First, after over a decade of operation, it holds leading advantages in supply density, rider network scale, delivery infrastructure, and merchant partnership depth.
Building on this foundation, by sharing traffic and technology infrastructure across its platform and implementing refined operations—from front-end user matching and merchant support, mid-order dispatch and subsidy strategies, to back-end delivery fulfillment and after-sales service—the company achieves high-efficiency, end-to-end operations.
Second, the deployment and deep integration of AI applications have become key drivers of this efficiency leap. It is crucial to recognize that the company is a technology firm connecting offline services with the online world. It possesses a wealth of real-world, physical data—a competitive moat built through years of groundwork, which is also its core advantage. The comprehensive push for "Physical AI" helps fully activate and leverage this strength.
From its strategic layout and tangible outcomes, the company has progressed from developing its own foundational large model to launching AI assistant products for users and merchants, and further to achieving scaled deployment of AI hardware.
At the large model level, its next-generation model, LongCat-2.0-Preview, opened for testing in April. With total parameters exceeding one trillion, it ranks among the world's top-tier large models. Reportedly supporting a 1 million token context window for processing massive text inputs in a single inference, it is also effectively adapted for production scenarios like code generation, complex task planning, and enterprise automation, making it a potent tool for efficiency gains.
For users, the company upgraded its consumer-facing AI assistants, "Xiao Tuan" and "Xiao Mei," helping users make complex, cross-scenario decisions. A recent collaboration between "Xiao Mei" and Tencent's Yuanbao will soon launch, allowing users to complete one-stop local service transactions more conveniently within Yuanbao, enhancing both service efficiency and user experience.
For merchants, the "Smart Shopkeeper" tool for in-store dining services has cumulatively served over 700,000 businesses, while the "Digital Employee" for service retail covers more than 300,000 merchants. An increasing number of businesses are utilizing their own AI assistants to reduce costs and improve efficiency.
In AI hardware, the "Urban Low-Altitude Network" autonomously built by the company's drone unit entered regular operations in May and is now open to the entire industry. Currently, its drone delivery services are operational in multiple cities including Beijing, Shanghai, Shenzhen, Hong Kong, and Dubai, having cumulatively completed over 900,000 commercial orders, ranking second globally. Among these, emergency delivery of medical supplies has become a regular service in several major cities.
The comprehensive advancement of "Physical AI" not only drives operational efficiency but also positions the company to tap into the next wave of AI's "super dividends." This is because the global AI industry is at a critical juncture, transitioning from "infrastructure explosion" to "scenario implementation." The company's deep focus on Physical AI, with both B2B and B2C products, places it in a rare position to capture substantial commercial AI value.
In essence, this strategy involves deeply integrating AI with high-frequency, essential, and strongly-fulfilled life services, maximizing the real-world, industrial value of AI.
Eight Years of Committed Investment in Chinese Hard Tech
With Unitree's IPO progress, the scope of the company's technology investments is becoming clearer. Beyond Unitree, early-stage support from the company stands behind many of China's prominent AI tech firms, including Zhipu AI, Galaxy General AI, Moonshot AI, and Xinghai Tu, with the company also being the largest external shareholder in Unitree.
The characteristics of the company's tech investments can be summarized in three main aspects.
First, it employs a concentrated, high-pressure investment approach in hard tech, primarily focusing on embodied intelligence, semiconductors, large AI models, autonomous driving, and AI hardware. It was not only one of the earliest major internet firms to invest in hard tech but has also done so extensively. From 2020 to May 2025, the company participated in 89 investment deals, with 47 targeting hard tech, accounting for over half.
Notably, even when its core business turned from profit to loss in 2025 due to intense competition in food delivery, its investment in hard tech increased rather than decreased, demonstrating remarkable strategic resolve and a long-term vision.
Second, the company goes beyond mere "financial investment" by providing substantial developmental resources, seeking to align innovative technology with consumer scenarios, fostering collaborative creation and ecosystem synergy.
This strategy is viable because of the company's powerful scenario advantage. Its instant delivery network covers over 2,800 cities and counties nationwide, coupled with massive offline service scenarios, providing an ideal environment for portfolio companies to test technologies and gather data.
This addresses a critical need for hard tech companies, which require vast amounts of real, complex physical interaction data to move "out of the lab" and integrate into real-world scenarios like services, retail, smart homes, and industrial production.
Taking the robotics field as an example, in March of this year, Unitree's founder Wang Xingxing publicly stated that the biggest current "pain point" for robots is insufficient generalization capability, which requires support from "video generation-based world models" and large-scale real-world physical scene data.
Clearly, the alignment and synergy between innovative technology and consumer scenarios not only help portfolio companies grow rapidly but also provide strong support for the company's own "Physical AI" initiatives, redefining the potential of its AI ambitions. This mutually beneficial relationship amplifies industrial collaborative value, representing true "co-creation and shared success."
Third, these technology investments have generated substantial returns for the company. Data shows that just from its holdings in listed companies, the market value exceeds 35 billion yuan. For instance, its approximately 4% stake in Zhipu AI, dubbed the "world's first large model stock," has generated unrealized gains exceeding 20 billion Hong Kong dollars. Furthermore, companies like Unitree, Unisoc, and Xinghai Tu are progressing with IPO processes, and 11 other unlisted portfolio companies are valued at over 10 billion yuan each.
However, the value of these technology investments is not reflected in the company's profit and loss statement nor fully captured in its current market valuation. While these are strategic investments that may not be realized in the short term, in the long run, they represent assets whose value far exceeds their book value.
Particularly regarding scenario collaboration, the combination of tech companies' technical prowess with the company's own scenario advantages suggests that the future potential of its AI initiatives may reach even greater heights.
Comments