The Thousand Groupon War
In the autumn of 2011, there were more than six thousand group-buying websites operating simultaneously across China. Six thousand. The number itself tells a story about a particular moment in Chinese internet history — a gold rush so frenzied, so indiscriminate, that the industry coined a term for the carnage: the "Thousand Groupon War" (qiantuandazhan). Venture capital flooded in. Salespeople fanned across cities, knocking on restaurant doors. Promotions burned cash at rates that would have made American dot-com executives blush. By year's end, 1,800 of those sites had evaporated. Within eighteen months, only a handful remained standing.
The company that emerged from this gladiatorial extinction as number one — quietly, methodically, while louder rivals like Lashou and Gaopeng (Groupon's own China joint venture) incinerated capital on billboard advertisements and hyperactive hiring — was a Beijing startup called Meituan. Its founder, a 31-year-old serial entrepreneur named Wang Xing, had already built and lost a Chinese clone of Facebook, watched a Chinese clone of Twitter get shut down by government censors, and learned a lesson that would define his next fifteen years: in a market of infinite copycats, the winner is not the first mover, the best-funded, or the most technically elegant. The winner is the most operationally efficient entity that refuses to die.
Today, Meituan is China's third most valuable internet company — a platform that processes roughly 90 million daily orders, commands nearly 70% of China's food delivery market, serves 770 million annual transacting users, and touches virtually every dimension of urban Chinese life: restaurant meals delivered in twenty minutes, hotel rooms booked on a phone, movie tickets purchased between subway stops, groceries dispatched from dark stores, bike-shares unlocked with a QR code. Its market capitalization fluctuates with the violent rhythms of Hong Kong-listed Chinese tech but has at various points exceeded $200 billion. Wang Xing has called it an "infinite game," borrowing the language of James Carse's philosophical text, and the metaphor is unusually precise. Meituan did not win a single war. It won a war, and then another, and then another — group buying, then food delivery, then hotel bookings, then in-store services — each victory funding the next campaign, each campaign expanding the battlefield until the company became something with no clean Western analogue. Not DoorDash. Not Yelp. Not Booking.com. All of them, and none of them. The Amazon of services, if Amazon had started as Groupon and fought its way into every adjacent category with an army of delivery riders on electric scooters.
The paradox at the center of Meituan's story is this: it is a technology company whose most important asset is not technology. It is operational density — the intricate, physical, unglamorous infrastructure of millions of couriers, millions of merchant relationships, billions of data points on routing and demand, layered on top of each other across 2,800 cities and counties until the network becomes nearly impossible to replicate. And yet, as of mid-2025, that network is under siege from every direction. JD.com has entered food delivery with a public spectacle — its billionaire founder personally delivering meals on a scooter. Alibaba has merged Ele.me into Taobao to create Taobao Shangou, a combined force that briefly surpassed 80 million daily orders on a single Saturday. ByteDance's Douyin has carved into Meituan's in-store business. Beijing's regulators, meanwhile, have opened an investigation into "cutthroat competition" among delivery platforms, concerned that the subsidy wars are transmitting deflation through China's urban economy.
Wang Xing, on a Q1 2025 earnings call, said what he has always said, albeit with more edge: "We are prepared to do whatever it takes to win the fight."
The fight, as it turns out, is the only constant.
By the Numbers
The Meituan Machine
~90MDaily orders across all services
770MAnnual transacting users (2024)
~70%Share of China's food delivery market
RMB 86.6BQ1 2025 revenue (~$12.1B), +18.1% YoY
5M+Active delivery couriers
~10MAnnual transacting merchants
2,800+Cities and counties served in China
The Serial Cloner
Wang Xing was born in 1979 in Longyan, a small city in Fujian Province, to a father who had been denied a college education during the Cultural Revolution on account of his family's class background. After China's reform and opening-up, Wang's father built a fortune in construction — cement factories, mostly — providing the financial cushion that would underwrite his son's remarkable tolerance for failure. Wang Xing graduated from Tsinghua University in 2001 with a degree in electronic engineering, then enrolled in a PhD program at the University of Delaware. His advisor left on sabbatical. Wang, bored, spent his time browsing the American internet and dreaming of startups. He later said that had he been admitted to a top-tier PhD program, he wouldn't have had the chutzpah to quit.
He quit. In 2004, he returned to Beijing, recruited two classmates — including Wang Huiwen, a roommate from Tsinghua who had spent university gaming and arguing with strangers online rather than studying — and pooled together 300,000 RMB. They rented a three-bedroom apartment in Haifeng Garden, each took a room, and coded in the living room. Their first project, a social networking site called Duoduoyou modeled on Friendster, attracted double-digit daily users. Then single-digit. It died.
Their second project died too. Their third — Xiaonei, a pixel-for-pixel clone of Facebook launched in December 2005 and targeted at university students — gained explosive traction. Starting from Tsinghua, Peking University, and Renmin University, it spread across Chinese campuses like a wildfire. The team acquired 5,000 users by subsidizing ticket sales at a Tsinghua student festival, spending 3,000 RMB for a customer acquisition cost of roughly 0.6 yuan per person. But Wang Xing was a first-time founder with no real fundraising experience. He lost his one-page business plan on the way to a meeting with Sequoia China and tried to scribble a new one in the taxi. The investment fell through. Funding dried up. In October 2006, Wang sold Xiaonei to Oak Pacific Interactive for a reported $2 million. The site was rebranded as Renren, listed on the NYSE in 2011 at a $740 million valuation, and eventually became a cautionary tale about Chinese social networking. Wang Xing left with a
Winston Churchill quote: "This is not the end. It is not even the beginning of the end. But it is, perhaps, the end of the beginning."
He launched Fanfou, a Twitter clone, in 2007. It quickly amassed over a million users. Then the government blocked it for nearly two years. "The root cause," one observer wrote, "lies in the fact that during the site's development, Wang Xing failed to realize that at a certain size, it takes on the attributes of media." A hard lesson about operating in China.
By 2010, Wang had been through more than a dozen ventures. He was 31, had been burned by investors, censors, and his own inexperience, and had developed a worldview that was equal parts Silicon Valley optimist and Chinese internet realist. He studied American companies obsessively — not to admire them, but to understand which models could be adapted for China's specific conditions. The model he chose next was Groupon.
Survival as Strategy
Meituan launched on March 4, 2010 — the first group-buying site in China, modeled explicitly on Groupon. Within months, the clones arrived. By the end of 2010, there were close to a thousand competitors. By mid-2011, the number had swollen past five thousand. Venture capital sloshed through the sector. Lashou.com, the early market-share leader, raised $111 million and spent lavishly on offline advertising and customer acquisition. Gaopeng, a joint venture between Groupon and Tencent, expanded at hyperspeed, opening offices across China and hiring thousands of employees before anyone had figured out the unit economics.
Wang Xing did something unusual. He refused to participate in the arms race. While rivals burned cash on billboards and salesforce expansion, Meituan focused on operational efficiency, technology infrastructure, and deal quality. The company kept its average deal price lower than competitors but its revenue per deal higher — a counterintuitive result driven by tighter curation and better conversion rates. Wang invested early in building internal systems: a
CRM for managing merchant relationships, analytics tools for tracking deal performance, a data infrastructure that most of his competitors, many of whom were little more than sales organizations wrapped in websites, never bothered to build.
The bet was existential. In a market where competitors were defined by how much capital they could deploy, Meituan defined itself by how little capital it wasted. "We rejected unsustainable spending and focused on efficiency and profitability," Wang said later. The approach looked conservative, even timid, when Lashou was grabbing headlines. It looked visionary when funding dried up.
The shakeout came fast. By late 2011, 1,800 group-buying sites had closed. By 2012, the landscape had collapsed to five survivors: Meituan, Dianping, Baidu Nuomi, Lashou, and 55tuan. Meituan ended 2011 as the number-one group-buying site in China by revenue, with 12.5% market share — modest sounding, until you consider that the second-place player had 9.7% and Groupon's Chinese operation had fallen to fourteenth place at 2.9%. Wang Xing had established his playbook: survive the bloodbath through operational discipline, let competitors kill themselves, then expand from a position of strength.
The tie-up matches the expectation of our team and investors. It will afford us more time and energy to explore new businesses and make better innovative products.
The Right-Hand Man and the Multi-Front War
To understand Meituan's execution, you must understand the partnership at its core. Wang Xing was the strategist — endlessly curious, always reading, the kind of founder who would spend two or three minutes examining a journalist's backpack before sitting down for an interview. Wang Huiwen was the executor. They met on their first day at Tsinghua in 1997, became roommates, shared a computer in their dormitory, and developed a dynamic that would prove nearly unique in Chinese tech: Wang Xing chose the wars; Wang Huiwen won them.
After the Groupon shakeout, Meituan did not rest. It expanded laterally — and on a timeline that remains breathtaking. In mid-2012, Meituan launched Maoyan, a movie ticketing platform. In 2013, it entered hotel bookings. In late 2013, it began food delivery. Each vertical was chosen with a specific logic: these were high-frequency, local-services transactions that shared the same fundamental infrastructure of merchant relationships, consumer demand generation, and logistics. Maoyan would eventually be spun off and go public; Meituan Hotels would surpass Ctrip in domestic hotel room nights. But food delivery — launched almost as an afterthought — would become the defining business.
Wang Huiwen ran the food delivery operation. He was, by most accounts, the person who figured out how to build a delivery network from scratch — recruiting riders, optimizing routes, developing the dispatch algorithm that matches orders to couriers in real time. The delivery network started as a marketplace model, connecting restaurants with third-party couriers, then evolved into a hybrid where Meituan operated its own fleet. By 2015, Meituan had over 530,000 daily active riders, each carrying a smartphone that transmitted real-time location data. The company's machine-learning systems used this data to optimize routing, cutting delivery times and improving rider efficiency in ways that compounded with scale.
The competitive map was staggering. In group buying, Meituan faced Dianping (backed by Tencent) and Baidu Nuomi. In movie ticketing, it competed with Alibaba's Tao Piao Piao and Tencent's Weiying. In food delivery, it battled Ele.me (backed by Tencent and Dianping), Baidu Waimai, and Alibaba-incubated Koubei. In online travel, it fought Ctrip, Baidu-backed Qunar, and Alibaba's Fliggy. Every business line was burning money. Wang Huiwen later likened it to Germany in World War II — fighting wars on multiple fronts. "But Germany only fought on two fronts," he noted. "We fought on four."
The Merger That Made the Super App
The most consequential deal in Meituan's history was not a technology breakthrough or a product launch. It was a merger forced by exhaustion.
By 2015, Meituan and Dianping — founded separately in 2010 and 2003, respectively — had been locked in a ruinous cash-burning war across group buying and local services. Dianping, originally a restaurant-review platform modeled on Yelp (though founded before Yelp), had expanded into group buying and become Meituan's most direct competitor. Both companies were hemorrhaging money. Both had powerful backers pulling strings: Alibaba stood behind Meituan, Tencent behind Dianping. The logic of consolidation was overwhelming.
In October 2015, the two companies announced a strategic merger. The new entity, Meituan-Dianping, would operate both brands in parallel. But the real story was what happened behind the deal. Alibaba, which had been a significant Meituan investor, refused to put more money into the combined company because it declined to integrate Meituan's app with Alibaba's ecosystem. In a move that would reshape Chinese tech competition for the next decade, Alibaba sold its entire stake in Meituan and invested instead in Ele.me, a food delivery startup that would become Meituan's primary rival. Wang Xing, cut loose by Alibaba, turned to Tencent. Tencent invested an additional $1 billion, merged its own delivery services into the joint venture, and — crucially — allowed Meituan-Dianping to operate independently.
How Meituan moved from Alibaba's orbit to Tencent's
2010Meituan founded; receives early investment from Alibaba
2013Dianping receives strategic investment from Tencent
Oct 2015Meituan and Dianping merge; Alibaba refuses additional investment
2016Alibaba sells Meituan stake, invests in rival Ele.me
2016–2018Tencent invests $1B+; Meituan-Dianping operates independently within Tencent ecosystem
Apr 2018Alibaba acquires Ele.me outright for $9.5B
The Alibaba-Tencent split created the defining competitive axis of Chinese local services. But it also liberated Wang Xing. Unlike Ele.me, which was absorbed into Alibaba's corporate structure and subjected to the conglomerate's strategic imperatives, Meituan-Dianping retained operational independence. Wang could allocate capital, choose battles, and build infrastructure according to his own timeline. Tencent, which had long practiced a laissez-faire investment philosophy, gave Meituan a prominent placement in WeChat's ecosystem — the most valuable piece of digital real estate in China — and otherwise stayed out of the way.
The results were decisive. By 2018, Meituan controlled approximately 64% of China's food delivery market. Ele.me had roughly 25%. The gap would only widen.
The IPO and the Food + Platform Thesis
On September 20, 2018, Meituan listed on the Main Board of the Hong Kong Stock Exchange, raising approximately $4.2 billion in what was one of the largest tech IPOs in the region that year. Shares rose more than 5% on the first day of trading. The company was valued at roughly $53 billion.
The IPO prospectus articulated what Wang Xing called the "Food + Platform" strategy. Food — delivery, restaurant reviews, reservations — was the high-frequency anchor. Everything else — hotels, travel, movie tickets, bike-sharing, ride-hailing, grocery delivery — was the cross-sell. The logic was deceptively simple: food is the most frequent transaction in human life. If you own the consumer's food relationship, you can sell them everything adjacent. A user who orders lunch three times a week is exponentially more likely to book a hotel through the same app than through a platform they use once a quarter.
We built one platform to support multiple categories, so that we can cross-sell. Food is the most important and largest category, but Meituan goes well beyond it.
The strategy rhymed with Amazon's playbook — use a high-frequency commodity category (books, in Amazon's case; food, in Meituan's) to build a customer relationship, then expand laterally into adjacent categories where the marginal cost of serving an existing customer is lower than the cost of acquiring a new one. But Meituan's version had a physical dimension that Amazon's did not. Food delivery required not just an app and a recommendation algorithm but a literal army — hundreds of thousands of riders, moving through traffic, every day, across thousands of cities. The operational complexity was staggering. The moat, if you could build it, was correspondingly deep.
By 2018, Meituan boasted 340 million annual transacting users — roughly one in four Chinese citizens had made a purchase on its platform. It worked with 4.7 million merchants across 2,800 cities and counties. The platform generated over 20 million transactions per day. Wang Xing's ambition was to make Meituan the "operating system for local life" — the default interface between Chinese consumers and the physical world around them.
The Delivery Network as Competitive Infrastructure
The beating heart of Meituan is not the app. It is the delivery network — a system so large and so intricately optimized that it functions as a near-public utility for Chinese urban life, and yet remains entirely proprietary.
By 2024, Meituan employed over 5 million delivery couriers. Combined with Ele.me, the two platforms "employ" nearly 6 million riders — a workforce larger than the population of Singapore. Each rider carries a GPS-enabled smartphone that transmits real-time location data. Meituan's dispatch system — an AI-driven algorithm that matches incoming orders to available riders, optimizes multi-stop routing, and dynamically adjusts estimated delivery times — processes millions of decisions per hour. The system has been trained on billions of historical deliveries and incorporates variables including weather, traffic patterns, restaurant preparation times, and building-access constraints (does the elevator work? is there a security gate?).
The unit economics of food delivery are brutal. The average order value in China is lower than in Western markets — a function of food prices, not consumer indifference — and the platform must extract enough margin from the commission and delivery fee to cover rider compensation, technology costs, and customer acquisition. Meituan's advantage is density. In a city where Meituan handles 70% of delivery orders, its riders spend less time idle between deliveries, its routing algorithms have richer data to optimize against, and its per-order logistics cost declines. This creates a flywheel: more orders improve rider utilization, which lowers costs, which allows the platform to offer lower prices or faster delivery, which attracts more orders.
The delivery infrastructure has also enabled Meituan's expansion into "instant retail" — the delivery of non-food goods (groceries, pharmacy items, electronics, flowers) within 30 minutes. This business, called Meituan Instashopping (闪购), leverages the same rider fleet and dispatch system but applies them to a broader set of goods sourced from local merchants and Meituan's own network of "flash warehouses" — small, strategically located fulfillment centers stocked with high-turnover items. As of 2025, Meituan's target was 18 million non-food orders daily and 100,000 flash warehouses by 2027. If Amazon is the river of everything that arrives in two days, Meituan is the geyser of everything that arrives in thirty minutes.
Wang Xing's Infinite Game
There is a book that circulates through Meituan's senior ranks like scripture. James Carse's
Finite and Infinite Games, a slim philosophical text published in 1986, distinguishes between two types of games: finite games, which are played to win, and infinite games, which are played to keep playing. Wang Xing has referenced the book repeatedly, and its influence on Meituan's strategy is unmistakable. The company has never optimized for a single decisive victory. It optimizes for survival, adaptation, and perpetual expansion of the playing field.
This is the through-line connecting every phase of Meituan's history. Wang Xing did not try to win the Thousand Groupon War by outspending Lashou. He survived it. He did not try to crush Dianping; he merged with it. He did not try to beat Alibaba at e-commerce; he built an adjacent empire in local services that Alibaba could not easily attack. When Douyin began encroaching on Meituan's in-store business in 2023, Wang did not panic and launch a content platform. He reorganized the company — the largest structural overhaul in six years, integrating the previously separate "to-home" (food delivery, flash sales) and "in-store" (group buying, reviews) business groups under a single leader, Senior Vice President Wang Puzhong — and refocused on operational synergies.
Wang Xing is endlessly curious. He reads voraciously. He once spent several minutes examining a journalist's backpack before consenting to an interview. He founded Fanfou, China's first Twitter clone, not because he thought microblogging was a business but because he was interested in information flow. His personal social media posts have included ancient Chinese poetry — on one occasion, a Tang dynasty poem whose veiled critique of imperial power was interpreted as a shot at Xi Jinping's government, wiping $2.5 billion from his net worth in 48 hours and sending Meituan shares down 14%. The episode revealed something about Wang Xing that is unusual among Chinese tech founders: a streak of intellectual recklessness that coexists with — and sometimes undermines — extraordinary strategic discipline.
RMB 10 billion here, RMB 10 billion there. It seems every internet player wants to chip in.
The Regulatory Vise
In April 2021, China's State Administration of Market Regulation (SAMR) announced an antitrust investigation into Meituan. The probe followed a record $2.8 billion fine levied on Alibaba just weeks earlier and signaled that Beijing's tech crackdown was widening beyond
Jack Ma's empire. At issue were "pick one from two" (
er xuan yi) practices — forced exclusivity arrangements in which Meituan allegedly pressured merchants to sign agreements preventing them from listing on rival platforms. The company was accused of using differential commission rates, security deposits, and algorithmic manipulation to enforce these arrangements.
Meituan released a contrite statement promising to "actively cooperate with the investigation" and "fulfill its social responsibilities." The fine, when it came, amounted to approximately $533 million — 3% of the company's 2020 revenue. It was, by Alibaba's standards, modest. But the investigation's real impact was behavioral, not financial. Meituan was forced to relax exclusivity requirements, reduce commission pressure on merchants, and — perhaps most significantly — operate with a level of public visibility that it had previously avoided. The delivery riders, in particular, became a political flashpoint. A viral 2020 investigative piece, "Delivery Drivers, Trapped in the System," exposed the algorithmic pressure systems that pushed riders to violate traffic laws to meet shrinking delivery windows, sparking a national debate about gig labor, platform accountability, and the human cost of thirty-minute convenience.
Meituan now employs over 5 million couriers, many of them rural migrants working as independent contractors without comprehensive social insurance. The political sensitivity of this workforce — young, urban, visible, essential — gives Beijing a permanent lever over Meituan's business. When JD.com entered food delivery in February 2025, its founder Liu Qiangdong made a point of offering riders comprehensive social insurance and housing fund contributions (the wuxian yijin), turning labor welfare into a competitive weapon. Meituan scrambled to match the offer.
The Three Kingdoms of 2025
The food delivery war of 2025 is not a mere subsidy skirmish. It is a structural collision between three of China's largest internet companies, each pursuing food delivery for fundamentally different strategic reasons.
Meituan is the incumbent, defending the franchise that generates the majority of its revenue and profit. Its ~70% market share and 90 million daily orders represent an enormous base — but also an enormous target. CEO Wang Xing has described the company's stance as one of total commitment: "We are prepared to do whatever it takes to win the fight."
JD.com entered in February 2025 with a publicity offensive that included founder Liu Qiangdong personally delivering meals on a scooter. JD's logic is not that food delivery will be profitable on its own — it almost certainly won't be, at least for years — but that high-frequency delivery drives daily engagement with the JD app, which in turn supports the e-commerce business. As Liu himself said: "Even if we lose money on delivery, it's still cheaper than buying traffic from Douyin or Tencent." JD claimed 20 million daily orders by May 2025, barely three months after launch.
Alibaba merged Ele.me into its core commerce platform, rebranding the combined offering as Taobao Shangou. On July 5, 2025 — dubbed "Super Saturday" — Taobao Shangou hit 80 million daily orders, approaching Meituan's territory. Alibaba secured a RMB 50 billion ($7 billion) internal investment for the business, signaling a willingness to sustain losses that Goldman Sachs estimated could reach $5.7 billion over twelve months.
The combined subsidy burn is extraordinary. Nomura estimated that Meituan, Alibaba, and JD collectively spent roughly $4 billion on discounts in Q2 2025 alone. The effect on profitability has been savage: Meituan reported an 89% drop in Q2 2025 adjusted net profit. Its stock fell more than 20% year-to-date by mid-year. Combined, Meituan and JD shed approximately $70 billion in market capitalization from their March highs.
Key combatants and their strategic logic
| Platform | Market Share (est.) | Daily Orders | Strategic Rationale |
|---|
| Meituan | ~65–70% | ~90M | Core business defense |
| Taobao Shangou (Alibaba/Ele.me) | ~25–33% | ~80M (peak) | High-frequency driver for Taobao ecosystem |
| JD Takeaway | <10% | ~20M | App engagement; cheaper than buying traffic |
Beijing is watching nervously. In January 2026, SAMR opened an investigation into "cutthroat competition" among delivery platforms — not, this time, to punish a monopolist, but to restrain all three combatants. The concern is that subsidized meal prices anchor consumer expectations, compress restaurant margins, transmit deflationary pressure through the urban economy, and exploit a delivery workforce whose conditions are already politically sensitive. The platforms have publicly pledged to curb price wars. The pledges, so far, have been largely performative.
Beyond the Border: Keeta and the Overseas Gambit
While the domestic war rages, Meituan has quietly opened a second front. In 2023, the company launched Keeta, an international food delivery brand, in Hong Kong. Within a year, Keeta became the market leader in the territory — a remarkable result given that Hong Kong already had established players including Deliveroo and Foodpanda. Meituan's formula was familiar: aggressive subsidies to acquire users and merchants, superior logistics technology, and a willingness to absorb losses in pursuit of density.
From Hong Kong, Meituan expanded Keeta to Saudi Arabia in 2024, reaching second place in the market within approximately a year. It launched Keemart, a grocery delivery service operating from front-end warehouses in Riyadh. Brazil, chosen as the beachhead for Latin America, received a $1 billion investment commitment. The market selection was deliberate: Meituan avoided Southeast Asia (where Grab and GoTo dominated, and where it had previously invested in Gojek) and chose markets where delivery infrastructure was less mature and incumbent competition thinner.
The overseas expansion represents a strategic hedge. If the domestic market is approaching saturation — 770 million annual users covers more than half of China's population — and if regulatory and competitive pressures are compressing domestic margins, international growth offers a new vector. The global online food delivery market is valued at approximately $1.4 trillion. But the challenges are formidable: Meituan must replicate its operational density in unfamiliar regulatory environments, with local labor forces, against competitors who have their own advantages in cultural familiarity and supply-chain relationships. The early results in Hong Kong and Saudi Arabia are encouraging. Whether they can be sustained in markets as complex as Brazil remains to be seen.
The Co-Founder's Return and the AI Pivot
In January 2020, Wang Huiwen announced his retirement from Meituan at age 40. The co-founder who had built the delivery network, who had fought the multi-front war, who had been Wang Xing's indispensable right hand for more than a decade, stepped away from daily management. The loss was, by any measure, significant.
Then came ChatGPT.
In early 2023, Wang Huiwen founded Light Year (Guangnian Zhiwai), an AI startup focused on large language models. Wang Xing invested personally in the Series A. Sequoia Capital and ZhenFund piled in. Light Year raised $230 million in two months and acquired OneFlow Technology. Then Wang Huiwen fell ill and stepped back. Meituan acquired Light Year for RMB 2.065 billion ($290 million) in June 2023, absorbing the team into the company.
By 2025, Wang Huiwen had recovered and returned to Meituan, leading an autonomous AI team internally known as GN06, focused on AI companions and chatbots. Meanwhile, Meituan itself declared food delivery and AI as its two strategic priorities. On the Q1 2025 earnings call, Wang Xing outlined the company's AI roadmap in three layers: applying AI to internal operations for productivity; using AI to upgrade existing B2B and B2C products; and building proprietary foundation models. The company's in-house model, LongCat, was publicly discussed for the first time. Meituan hired Pan Xin, former CTO of Sharge and former head of vision large-model development at ByteDance, to lead multimodal AI innovation. AI-powered tools for restaurant merchants were launched in October 2025, offered free across the industry.
The AI bet is embryonic but strategically coherent. Meituan's operational data — billions of delivery records, merchant transactions, consumer behavior patterns — constitutes an enormous training corpus for AI systems optimized for local commerce. If the company can build AI that genuinely improves routing efficiency, demand forecasting, merchant productivity, or consumer discovery, the compounding effects on the core business could be substantial. The question is whether a company fighting an existential war on three fronts in food delivery can simultaneously fund and execute a meaningful AI transformation.
The Membership Gambit
On March 31, 2025, Meituan launched Meituan Membership — a unified loyalty program spanning food delivery, in-store services, hotels, transportation, and entertainment. The timing was not coincidental. With 770 million annual users, the company had reached a ceiling on new customer acquisition. "With a large enough user base, the growth space for new customers is bound to shrink," said Zhou Mo, the executive who led the initiative. "We need to increase the consumption frequency of existing customers."
The membership system — built on a foundation of "Shen Coupons" (深券) that offer universal discounts across Meituan's various service categories — represents a strategic pivot from acquisition to retention. It is Meituan's version of Amazon Prime: not a premium subscription in the traditional sense, but an incentive architecture designed to collapse the barriers between Meituan's various business lines. A user who orders food delivery might be nudged, via membership benefits, to book a hotel through Meituan rather than Ctrip. A user who buys movie tickets might be offered a discount on a massage at a nearby spa. The goal is to make Meituan the default interface for all local consumption — to turn occasional users into habitual ones, and habitual users into ones who cannot imagine using anything else.
When a user is just one cup of coffee, one night in a hotel, one movie ticket away from the next membership tier, they'll choose Meituan.
The Weight of the Network
There is a particular detail that captures something essential about what Meituan has become.
In December 2025, the company released a three-minute promotional film about a delivery rider named A Lan in Dali, Yunnan Province. In the film, A Lan dreams of becoming a photographer. She rides through mountains in a spotless yellow jacket with bunny ears on her helmet. After three months of deliveries, she saves enough to buy a 15,000-yuan camera and holds a small exhibition of the photographs she takes between orders. The film was removed from the internet within 24 hours, after viewers excoriated it as jitang — "chicken soup," toxic positivity. "It's clear from the video that not only has the director never been a food delivery driver," one user wrote, "but also that he has never ordered take-out."
The backlash was not about a bad advertisement. It was about the growing distance between the machine and the humans who power it. The average monthly income of high-frequency Meituan riders — those working 26 or more days a month — ranges from 7,230 to 10,100 yuan. The sector employs roughly 10 million drivers across platforms. A 26-year-old rider named He Xu, who juggles delivery and photography in Yunnan much like the fictional A Lan, told a reporter that her first day earned her 87 yuan — about $12 — for nearly twelve hours of work. "Being a food delivery rider is exhausting and anxiety-inducing for me," she said. "All my focus has to stay on one thing: balancing my own safety on the road with delivering the food as fast as possible."
This is the tension that Meituan cannot resolve. The company's competitive advantage — its 5 million riders, its 90 million daily orders, its 30-minute delivery promise — is built on human infrastructure that operates at the edge of sustainability. Every efficiency gain the algorithm produces, every minute shaved from a delivery window, every incentive to keep riders moving faster, creates compounding pressure on the bodies doing the work. The delivery riders are both the engine of Meituan's value and the political vulnerability that regulators, competitors, and the public can exploit. Wang Xing's infinite game continues. But the riders know almost nothing about their role in it. They are still sprinting ahead, as one journalist wrote, "for the possibility of a better life."
Meituan's market capitalization, as of late 2025, sits at roughly $100 billion — down from a peak above $200 billion, compressed by the delivery war, regulatory pressure, and the general malaise hanging over Chinese tech. The company processes 90 million orders a day across a network spanning 2,800 cities. In the Q2 2025 earnings call, Wang Xing said: "No matter what happens in the market we will focus on doing the right things, going back to basics — selection, price, service, and delivery." Then he pivoted to a discussion of AI.
Somewhere in Dali, a delivery rider with bunny ears on her helmet is weaving through traffic, racing against an algorithm she cannot see, carrying someone's lunch. The camera around her neck is imaginary. The 87 yuan is real.
