Part IThe Story
Have You Eaten?
The poem was short and old, written more than a thousand years ago in the late Tang dynasty by a minor poet named Zhang Jie. Its title was "Book Burning Pit." Its subject was Qin Shi Huang, the first emperor, who burned the books and killed the scholars to secure a dynasty meant to last ten thousand generations, and who was brought down instead, in the poem's closing turn, by rebels who had never read anything at all. It is the sort of poem Chinese schoolchildren encounter and then forget. In the first days of May 2021, the founder of China's third-largest technology company posted it to a microblog almost nobody used anymore.
The microblog was called Fanfou, 饭否. The name translates roughly as "Dinner or not?" It echoes the old greeting that is less a question about food than an inquiry into whether a person is all right: Have you eaten? Wang Xing had built Fanfou in 2007 as China's first serious answer to Twitter. It gathered more than a million users, then millions. After the Urumqi riots in July 2009 the government took it offline, and it stayed dark for more than a year. By the time it was permitted to return, Sina Weibo and Tencent Weibo had taken the market. Fanfou survived as something closer to a private notebook kept in public, a place where its owner posted fragments, Murakami, and classical verse. Chinese media began calling him the "poet entrepreneur." The phrase was half admiring and half puzzled.
The man who owned the notebook also controlled more than 65 percent of China's food delivery market. Every day, through an army of riders numbering in the millions, his company answered the question in his old website's name in the most literal way possible: yes, you have eaten, and here it is. That spring his company was already under investigation by the State Administration for Market Regulation for alleged anticompetitive practices, a probe announced in late April. On Monday, May 10, Meituan's shares fell as much as 9.8 percent in Hong Kong, their worst single-day drop in two months. Roughly $16 billion in market value disappeared in a session. Within weeks the losses approached $39 billion. Wang personally lost more than $2.5 billion in a matter of days. He deleted the poem.
"Anything related to the Cultural Revolution or book burning is particularly sensitive" that year, the University of Hong Kong historian Xu Guoqi told the South China Morning Post, because the Communist Party's centennial was approaching. "Anything which doesn't openly praise the party can be taken to mean different things." Some weeks later, according to people familiar with the matter, Beijing officials summoned Wang to a meeting and told him to keep a low profile, at least for a while.
Nobody outside his own head knows exactly what Wang meant by the post, and he has never offered a reading anyone could hold him to. But the episode puts most of his story on one small screen. There is the habit of reading. There is the nostalgia for an earlier internet he had helped invent. There is the platform grown so large that its founder's idle sentence becomes a market event. And there is a family history in which the cost of words had once been paid in full. The man who built the most efficient system on earth for asking and answering have you eaten? is also a man who, at eighteen, sat down to hot pot with his new classmates and asked a different question entirely.
The rest of this story moves outward from that post, backward and forward, and returns to it.
By the Numbers
The Meituan Machine
CN¥364.9BMeituan revenue in 2025 (US$53.82B), up 8.1% year over year
770M+Annual transacting users at the end of 2024
14.5MAnnual active merchants at the end of 2024
5.27MDelivery riders who earned income on Meituan in 2021
~3,000Chinese Groupon clones Meituan faced within a year of launch (Forbes, 2011)
$2MPrice at which Wang sold Xiaonei in 2006, the network later rebranded Renren
110,000Meituan employees in 2026
A Family That Knew What Words Cost
Begin with the grandfather, because the profile literature on Wang usually doesn't. In He Jiayan's long 2019 profile for The House of Startups, a piece titled with a Confucian pun, Wang Xing at 40, Without Doubts, the family history arrives as a single compressed paragraph. Wang's family suffered greatly after the People's Republic took power. His grandfather was driven to suicide during the Cultural Revolution. His father, Wang Miao, spent years forced to labor in the countryside and was barred from university because of his family background. Then came reform and opening, and the father made a fortune in construction. He owned a cement factory in Longyan, a mountain city in western Fujian.
That is three generations in four sentences. One man was destroyed by the state's appetite for ideological purity. One man was denied education and then, when the rules changed, poured concrete for a country rebuilding itself and grew rich. One man was born on February 18, 1979, weeks after Deng Xiaoping's reforms formally began, into the wealth his father had made. Fujian is a province of traders and emigrants, of families who went to Southeast Asia and came back with money and opinions. Longyan sits inland, in its hills. The boy was, by every account, an obsessive reader and an excellent student.
In 1997 he was admitted to Tsinghua University's Department of Electronic Engineering, in its radio major, from Longyan No. 1 Middle School. He got in through baosong, recommended admission that bypasses the national examination and is reserved for students a school is confident about. He arrived in Beijing having just turned eighteen. More than two decades later, at a 2020 founders' summit, he did the arithmetic out loud: he had now lived in Beijing longer than in Longyan, longer than anywhere. "Beijing," he said, "is my second hometown."
There is one Tsinghua anecdote, preserved by He Jiayan, that everyone who writes about Wang eventually repeats, and for good reason. At a hot pot dinner early in freshman year, the upperclassmen offered each newcomer one question that all the seniors would have to answer. The seniors expected questions about secret crushes. Wang asked: What do you think is the meaning of life?
One version of this boy is insufferable. Another is unusually serious and unembarrassed by the size of his own questions, and that is closer to the man his later colleagues describe. He also danced. In a speech marking Meituan's first anniversary, he mentioned in passing that he had performed onstage as a dancer in high school and at university. Nobody who has watched him run an earnings call would guess it.
The hot pot question and the cement factory are worth holding together. One is metaphysics, the other is aggregate and water. Wang's adult career would be an attempt to do both at once: to build a company on philosophical first principles about networks, games, and time, whose actual product was the movement of hot food across congested cities by people on electric scooters.
Nodes and Edges
He went to America in 2001 on a full scholarship to the University of Delaware's Department of Electrical and Computer Engineering, enrolled in a PhD program. Delaware is a respectable school. It is not MIT. Wang later said, with characteristic dryness, that had he been in a top program he would not have had the nerve to quit.
His advisor worked on high-performance computing, which in 2002 meant, among other things, the human genome. The Human Genome Project was nearly complete and was producing data faster than anyone could process it, and a new field called bioinformatics needed computing power. That summer, home in China, Wang visited the Beijing Genomics Institute. It was then, as he recalled at Sequoia China's tenth-anniversary event in October 2015, "a rather shabby place" near Beigao, out by the airport. A decade later he visited again. BGI had moved to Shenzhen and become gleaming, its products abundant, and visitors could taste yogurt made with gene-modified cultures. He told the story as a parable about Moore's Law. It also works as a parable about China, and about the shabby room that becomes the glass tower, which was the only trajectory he ever cared about.
The advisor went on sabbatical, or was simply unavailable. Wang got bored, and he went online, where something was happening.
Late in November 2003, he wrote an email. He remembered the date precisely at a Tsinghua event in September 2017, partly because one of its recipients, Wang Huiwen, was sitting in the audience. Wang Huiwen was a classmate from the Tsinghua years. He would become one of the most important operators in Wang Xing's life, and two decades later he would leave to found an artificial-intelligence startup that Wang would back and then buy. The email also concerned two other Tsinghua alumni, Mu Yan and Yang Jinfang, the first and second presidents of the university's Science and Technology Innovation Association. They had built a Chinese social network called Uuzone, which Wang had read about on Shuimu Tsinghua, the campus BBS.
The email made an argument from engineering. In America, Friendster had appeared. Facebook did not yet exist; it would launch at Harvard that February. Wang had studied computer networks, and he saw social networks as a structural rhyme. A computer network, he wrote, is a web whose nodes are routers and whose edges are fiber-optic cable, and information flows along the fiber between routers, carrying entertainment, commerce, everything. A social network is the same web with the nodes replaced by people and the edges by relationships, and information would flow along relationships. This is a huge opportunity, he wrote. We should seize it.
He came home at Christmas 2003.
The next part of the story is not glamorous, and Wang has told it without much glamour. Early in 2004 he and two classmates pooled 300,000 yuan and rented a three-bedroom apartment in Haifeng Garden, on Xueqing Road in Haidian. Each partner took a bedroom and they worked in the living room. They built many products. They cloned Google Maps. They cloned Evite. In August 2004 they launched Duoduoyou, 兜兜友, "Pocket-Pocket Friends," a Friendster for Chinese university students. A year later its daily user growth had fallen to double digits, sometimes single digits, and they were begging friends and classmates to sign up. Another project, Youzitu, "Traveller's Map," aimed at Chinese students abroad. It, too, ceased operation.
What went wrong? Wang's later diagnosis, delivered in a 2012 talk on his entrepreneurial history, is worth quoting because it contains the seed of everything he did afterward. They had wanted to reach everyone, so promotion was hard and density was low. "The lack of focus stemmed from inexperience; we didn't know what was important and what wasn't." He went on to the problem with cloning: "The difficulty of cloning lies in the feeling that you can differentiate yourself slightly; when your understanding of the product is not deep enough, you don't know why it does what it does, nor do you know which differences are critical. Subtle differences can have significant implications."
In August 2005 the team made three new attempts. One launched on December 8, 2005. It was called Xiaonei, 校内网, "Inside-Campus Net," and it was, by many accounts, a pixel-for-pixel copy of Facebook. It started at Tsinghua, Peking University, and Renmin University and spread outward. The first growth campaign piggybacked on the Tsinghua Electronic Engineering Department's student festival: they subsidized ticket sales, spent 3,000 yuan, and acquired about 5,000 users, at a cost of roughly 0.6 to 1 yuan per person. Through the end of December, the team of three was still wavering between Xiaonei and another project. Dozens of competitors, at home and abroad, were also cloning Facebook, with names like 5Q and Zhanzuo. Why did Xiaonei win? "Good luck (Tsinghua) and simplicity," Wang said.
Within months Xiaonei was the most trafficked Facebook clone in China and growing toward a million users. And Wang ran out of money.
Every account agrees that this part is excruciating, though the details shift by source. On his way to pitch Sequoia China, he lost his one-page business plan and tried to rewrite it by hand in the back of a cab. He borrowed money from his parents to keep the servers running, which meant cement-factory money propping up a social network. A prospective investor signed a term sheet with a lock-up clause and then never wired the money. Wang later called the lock-up "a pitfall." He admitted he had not understood how big the thing could be: "I only wanted to raise 1 million USD, but that was far from enough." He had not met enough venture capitalists. He recommended, to anyone listening in 2012, the blog Venture Hacks and the book Founders at Work. It was the advice of a man still sore.
In October 2006 he sold Xiaonei to Chen Yizhou's Oak Pacific Interactive, also known as Qianxiang, for roughly $2 million. Chen was another serial internet entrepreneur, one who knew how to raise money. He rebranded the site Renren, 人人网, "Everyone Net," and went on to raise $430 million from SoftBank. On May 4, 2011, Renren listed on the New York Stock Exchange, raising $740 million, with a valuation that, before the offering, Forbes put at more than $4 billion.
The next afternoon, on a Sina microblog Q&A, someone asked Wang how he felt about his past setbacks.
— Wang Xing, Sina Weibo interview, May 5, 2011, the day after Renren's NYSE listingSome people can never stop thinking about what they've lost, and forget what they've gained. I'm not that kind of person. I'm full of curiosity and passion about life. "Go forward without holding back" — that's my attitude.
The phrase he used was 纵情向前. It was also the name he had given, two months earlier, to Meituan's first-anniversary launch event.
When a Website Becomes a Newspaper
After the sale he stayed briefly with the acquirer, then left in 2007 to build two more companies at once. Hainei, 海内, was another social network for adults who had graduated from Xiaonei. It never caught Renren or the newer Kaixin. Fanfou was the Twitter clone, the one with the dinner-table name.
Fanfou was, in its way, the best thing he had made. It was early: Wang's first-generation microblog predated Weibo by more than two years. It was lively, with a cult following of devoted users that included a number of China's internet intelligentsia. A 2010 alumni profile published by Tsinghua described Wang as a "shy young man" with sharp intuition, rigorous thinking, and a very fast way of talking. It concluded, in a phrase that carried a whole national mood, that he was someone who would "probably have succeeded long ago if he were abroad." Fanfou was, the profile said, "wildly popular" in 2009, with "a crowd of famous die-hard fans," until it was "abruptly stopped for policy and other reasons."
The reasons were less mysterious than that phrasing suggests. In July 2009, ethnic violence broke out in Urumqi, the capital of Xinjiang. Rumors and footage moved across Chinese social media faster than the authorities could manage them. Fanfou was taken offline and did not return for more than a year. He Jiayan's 2019 profile offers what may be the most important sentence anyone has written about Wang's education as a businessman: "The concrete reason for the ban is hard to pinpoint. But the root cause 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."
At a certain size, it takes on the attributes of media. In China, media is not a business category like retail or logistics. It is a function of the state that private companies are permitted, under supervision, to perform. A website that crosses the line between hosting conversation and shaping it has, without anyone announcing anything, changed its relationship to power. Wang had built a network in which, as his 2003 email put it, information flows along relationships. He had not fully reckoned with who else cares where information flows.
The most remarkable thing about the Fanfou years may be who stayed. Through the long purgatory of the shutdown, most of the team stuck around and waited for a reopening that kept not coming. One of the few who left was a young engineer named Zhang Yiming. He would go on to found ByteDance, a company whose core competence turned out to be a sophisticated understanding of what algorithms should and should not show to hundreds of millions of people. There is a rhyme in that. Zhang learned something from watching Fanfou get switched off. Wang learned something too, but what he learned was to get out of the information business altogether.
By January 2010, Fanfou's prospects of reopening looked hopeless. Wang began thinking about a site like Groupon.
The question of what kind of man this makes him has no tidy answer. Forbes, in April 2011, called him "The Cloner," and the label stuck: four clones, one sold, one smothered, one outcompeted, and now a fifth. The Qihoo 360 CEO Zhou Hongyi, among others, took public shots at his habit of copying. Wang's own account is subtler and less defensive than you might expect. He thought the copying was the easy part. Knowing which differences mattered was the hard part. Knowing when the copy had become something its original never had to be, a newspaper in a country without a free press, was the hardest of all.
He had now failed, depending on how one counts, nine times. A 2020 Chinese biography is subtitled, in effect, Nine Defeats, One Victory. An anonymous investor's line in He Jiayan's piece was harsher and more vivid: "From a pile of dead men, he pushed himself forward."
Three Thousand Copies of One Idea
Meituan launched on March 4, 2010. The name means "beautiful group." The night before launch, or close to it, roughly a dozen people met in a three-bedroom apartment in Huaqing Jiayuan, just outside Tsinghua's East Gate. It was the second three-bedroom apartment in this story, six years and a few hundred meters from the first. Wang recalled the meeting in a March 2015 address to the whole company. They faced many difficulties. "At that meeting, quite a few people cried," he said. "Including me."
It would be pleasing to report that Meituan's victory was a matter of a better idea. It wasn't, because there was no better idea. There was only one idea, and everyone had it. Within a year Forbes estimated that some 3,000 Groupon clones had sprung up across China, including as many as a dozen strong national contenders. Later counts ran to five thousand. The period acquired a name, 千团大战, the "War of a Thousand Groupons." The leading rival, Lashou, raised $111 million and spent $10 million on marketing in March 2011 alone. Groupon itself entered through a richly funded joint venture with Tencent called Gaopeng. Another competitor grabbed share partly by cloning the domain name Groupon.cn, which Forbes noted was done "in classic Chinese fashion."
Groupon's entry was managed by the Samwer brothers, the German operators who had built a fortune by cloning American web companies for Europe. They were, in other words, the original cloners now arriving to supervise the clones. In October 2010, Oliver Samwer met with Wang and the heads of the other leading Chinese group-buying sites. According to some of them, his message was arrogant, patronizing, and warlike: agree to be acquired, or be defeated. There is real comedy in the scene, Europe's master copyist telling China's master copyists that copying wasn't allowed. Gaopeng did not win.
The war turned Groupon's famous cash flow inside out. In America, Groupon took 30 percent or more of each coupon. In China, Meituan's cut fell from 20 percent at the start to a high single digit by spring 2011, and its rivals were all under ten. Sales teams had to be staffed in more than a hundred cities. "Victory," Forbes wrote, "may yet be defined by who can bankrupt whom first."
What did Wang do differently? The sources agree on three things, and none of them is clever. That is the clever part.
First, Meituan ran one deal per day. While competitors stuffed their pages with offers, Wang kept the original Groupon discipline so that each merchant would actually feel the effect of working with him. It was the Duoduoyou lesson about density, applied to merchants instead of users.
Second, on March 4, 2011, at the first-anniversary event in the Haidian Theater, Meituan announced unconditional refunds. A customer whose coupon expired unused would get the money back automatically. It was the first such guarantee in the industry. Wang said in 2015 that it put the "consumer first" value into "actual action," and that it built the brand, attracted merchants and consumers, and helped Meituan win the war. The invitations to that event each carried a 16-digit code redeemable for 500 yuan of Meituan credit. Wang told the assembled journalists not to throw them away: "This invitation is just like Meituan. It reminds you that there are surprises everywhere in life, in places you don't pay attention to."
Third, and most counterintuitive, Meituan saved money on advertising. Wang reasoned that his competitors' enormous ad budgets would educate Chinese consumers about group buying in general without building loyalty to any particular site. He let them pay for the category and spent his own money on the things that kept users coming back.
At that first-anniversary speech he called group buying "a super-perfect business model" and drew a pyramid of Chinese businesses on the screen. He was about to argue for the internet's role in the commercial life of every restaurant and cinema in the country. He told the room it would not be an exaggeration. Meituan's first full year produced 1.45 billion yuan in transactions.
Money came. Sequoia Capital put in $12 million at the end of 2010, the same firm whose meeting Wang had once arrived at without a business plan. In 2011 Alibaba led a $50 million round. Five years later, speaking at Sequoia China's tenth anniversary, Wang described his relationship with the firm as a story that went "from missing each other to finally choosing each other," and offered a line that was half compliment and half weather report.
— Wang Xing, Sequoia Capital China tenth-anniversary event, October 17, 2015As long as you're still an entrepreneur, as long as you're still in this industry, I believe everyone will run into Sequoia sooner or later — because Sequoia is always there, and always charging out in front.
The year that decided the war was 2012. At the annual company meeting, held that year in Peking University's Centennial Hall, Wang told his employees the story of the race to the South Pole. That is Amundsen's story, the explorer who covered the same distance every day in good weather and bad, while his rival pushed hard on fair days and sheltered on foul ones. Wang's instruction was to advance thirty kilometers a day. "In 2012," he recalled, "we frantically opened stations, frantically listed deals, frantically made visits." The word he used, 狂, means something like "madly." By year's end, Meituan's advantage in group buying was established. That same year the company articulated its operating philosophy, which it called the "Three Highs and Three Lows," and reached monthly breakeven. In 2013 it turned its first annual profit. In 2013, too, Wang told the company Meituan would hit 100 billion yuan in transaction volume by 2015, a number that sounded absurd. In 2014 the company did 45 billion, more than 90 percent of it on mobile.
Two paradoxes sit inside the War of a Thousand Groupons. The first is that the most famous copier in Chinese tech won by copying the original more faithfully than anyone else did, keeping one deal a day while everyone else improvised. The second is that the war was won by a man whose previous companies had been "very light." In the words of a contemporary Chinese account, nobody had expected him to be able to manage a company of four or five thousand people. The intellectual who once asked about the meaning of life had become, out of necessity, a manager of sales territories.
The Monk With No Magic
In February 2012, in an interview with the Chinese trade magazine Touzijie, Wang explained himself through a character from Journey to the West, the classical novel every Chinese child knows. Most readers love the Monkey King, Sun Wukong, the rebel with seventy-two transformations who can somersault a hundred and eight thousand li in a single leap. Wang chose the monk.
He had recently reposted someone else's microblog, he said, and it had moved him enough that he repeated it to the magazine.
— Microblog post reposted and endorsed by Wang Xing, as recounted in a February 2012 interview with TouzijieIf a few years ago I was an untamed Sun Wukong who didn't know what to do with himself, today I have become Tang Seng — without seventy-two transformations, with nothing left but a single stubborn intent. Everyone thinks Tang Seng has no abilities. Facing the eighty-one ordeals, he cried, he begged — but not once did he say, "Let's give up on the scriptures. Let's all go home."
"It's been almost eight years," he went on, "since I came back to China at Christmas 2003. A lot has changed, but much in me hasn't. I still love life, I'm still full of passion. I'm still me. In some ways I'm not that different from the Wang Xing of high school."
In the same interview he named a book that had influenced him a great deal, James Carse's Finite and Infinite Games. Carse was a religious-studies scholar at New York University, and his slim, aphoristic book divides human activity into two kinds. Finite games are played to win, and they end. Infinite games are played to keep the game going, and their players are less interested in winning than in changing the rules. A 2010 Tsinghua alumni profile of Wang carried a headline drawn from it: The infinite game is played with the rules themselves. Chinese commentators have read Wang's whole career through Carse, sometimes to the point of cliché. The influence is real, though. It shows in his insistence that a competitor today can be a partner tomorrow, in the way he treated the Dianping merger, and above all in his persistent refusal to accept that any war is over.
The Carse book sits within a larger and stranger library. Wang's reading lists circulate widely in Chinese business media, some of them assembled by fans and some cross-checked against Meituan's own recommendations. Nicholas Negroponte's Being Digital is said to have been his first internet book, with its argument that moving bits is faster and cheaper than moving atoms. Read beside Meituan, that is a richly ironic choice, since Meituan's ultimate business is moving atoms (noodles, medicine, flowers) with the information system as the scaffolding. The lists include Stavrianos's A Global History, which Wang reportedly recommended internally as a popular history with "no particular academic originality" that nonetheless "helps you understand bigger things." They include Sun Tzu, from whom commentators draw the line 胜可知，而不可为: victory can be foreseen, but it cannot be forced. They include Jack Welch's Winning, Peter Thiel's Zero to One, and The Lean Startup.
And they include the Book of Rites, about which Wang made a remark that is either profound or a very good joke. Confucius, he observed, was in a sense an entrepreneur from 2,500 years ago. He started out in the state of Lu with a set of ideas and something like a business plan. Failing to get local backing, he traveled from state to state looking for resources, hoping to turn the idea into reality and change society. "That," Wang said, "is the story of Confucius's life."
The investor Xu Xin of Today Capital has called him "a deep learning machine." Shen Nanpeng of Sequoia China said Wang was one of the few thinkers with a clear understanding of the wild, overgrown terrain of the Chinese internet, and one of the best at applying philosophical rigor to management. A saying attributed to him, repeated so often in Chinese media that it has become a kind of signature, holds that most people will do anything to avoid real thinking. The other line he returns to, in many forms, is about time: the more confident you are about the future, the more patient you are with the present.
It would be easy to stop there and leave him a philosopher-king. The texture of his speeches suggests something more nervous. In December 2013, onstage at a Chinese business leaders' conference on the theme of "overcoming internet anxiety," Wang delivered one of his most quoted lines. Mobile had overturned PC internet in a very short time, he said, but "mobile internet is only in the pocket-watch era. It hasn't reached the wristwatch era yet." Then he admitted that he himself was anxious. Sixty percent of Meituan's transactions were already on mobile. But the app had 60 million installs in a country of perhaps 600 million mobile internet users, so they served only a tenth. They were the biggest in group buying, with 50 percent share and 10 billion yuan, yet "we still have too much to do." Companies, he said, should prepare for a future of 永无宁日: never a day of peace.
The monk with no magic does not rest, because there is always another ordeal.
Food Is the Front Door
In late 2013, on a date the company records as December 19, Meituan Waimai, its food delivery service, began trial operations in Beijing. That year Meituan formally declared what it called its O2O strategy, online-to-offline, which had become the Chinese internet's favorite acronym. Group buying had taught the company how to sign up restaurants by the hundreds of thousands. Delivery would teach it something else: how to own the most frequent transaction in a city-dweller's day.
A hotel is booked a few times a year. A movie ticket is bought a few times a month. Lunch happens every day, and dinner too. Meituan's food delivery business became, in the phrase of a later analysis, "a key driver of user engagement," pushing traffic toward hotels, travel, and everything else. In Wang's industrial philosophy, food was the front door. Once you had walked through it twice a day for a year, you might as well book your hotel and your haircut from the same app.
That bet put him on a collision course with almost everyone. The first collision, oddly enough, ended in marriage.
Dianping, 大众点评, was China's Yelp: a Shanghai-born review platform that had accumulated a decade of user-generated content on restaurants, with photographs, ratings, and recommendations. It was the deepest database of opinion about where to eat in the country, and it was also in group buying, and therefore at war with Meituan. In March 2015 Wang told his staff that this would be "the year of the decisive battle for O2O." On October 8, 2015, the two companies announced that they would become one, in a deal Bloomberg valued at $15 billion. Wang became co-CEO of the merged company. Days later, at Sequoia's anniversary, he introduced himself with mild deadpan: "About a week ago Meituan and Dianping became one family."
The merger is a hinge in the story, and its geopolitics are worth spelling out. China's internet in those years was a system of allegiances to its two great patrons, Alibaba and Tencent. Meituan had taken Alibaba-led money in 2011. After the merger, Tencent became the combined company's main backer. Alibaba cut its stake and in May 2017 invested $1 billion in Ele.me, Meituan's chief delivery rival, which it would later absorb. In January 2016 the merged company raised $3.3 billion at a valuation above $18 billion. In 2017 it raised another $4 billion at $30 billion. Wang told Chinese reporters that the competition with Ele.me would last a long time and that neither side should expect to dominate soon. He was half right: the competition lasted, and his side came to hold more than 65 percent of the market.
The company kept expanding. Hotels and travel arrived in 2015. Ride-hailing launched in Nanjing in 2017, a frontal move into a market Didi effectively monopolized. Through the merger Meituan also became one of China's largest online hotel platforms. Asked about limits, Wang told Chinese media, "I don't think I have to set limits on myself." Fortune, in 2018, reported that he had begun calling his creation an "Amazon for services."
In 2018 he bought Mobike, the bike-sharing company whose orange bicycles had flooded Chinese sidewalks, for $2.7 billion. The rationale was another front door: short trips, high frequency, a map of how people move through a city. The cost was visible in the IPO filings. In the four months ended April 2018, revenue grew 95 percent to 15.8 billion yuan while losses nearly tripled to 22.8 billion, including the impact of the money-losing bike company.
On September 20, 2018, Meituan Dianping listed on the Main Board of the Hong Kong Stock Exchange at HK$69 a share. Wang struck the ceremonial gong. The shares closed the first day at HK$72.35, up 5.3 percent, giving the unprofitable company a market value of $50.9 billion, more than General Motors. It raised $4.2 billion, Hong Kong's second-biggest tech listing of the year. Wang's roughly 11 percent stake was worth $5.3 billion. He was thirty-nine. That week he told Bloomberg TV: "We have our largest core business, food delivery, that's growing super fast and it's also very close to break even." The pre-IPO investor Ben Harburg, of MSA Capital, explained his patience in an analogy that no one in the room needed explained. "We all have to approach them like an Amazon. I don't believe they will be profitable in the short term."
The kid who lost his business plan in a cab had become, fifteen years after the November 2003 email, one of the most powerful people in Chinese commerce. In his own frame, he had built a network whose nodes were restaurants, riders, and hungry people and whose edges were orders. In 2019 Meituan changed the main color of its logo from blue to yellow, the color of the riders' jackets and helmets that had become part of the visual texture of every Chinese city. That same year Fast Company named it the world's most innovative company.
The Temperature of the Room
Every serious founder eventually develops a theory of history. Wang's arrived early and was delivered, as was his habit, in long internal speeches that leaked.
In April 2017 he told an audience that the mobile internet's free ride was over. Smartphone shipments had stopped growing. Consumers were uninstalling apps. Preinstall fees and app-store promotion costs would rise 30 percent over 2016. "The entire mobile internet will be extremely brutal," he said. "Many companies you can't imagine will die." Hot money had inflated a bubble, the dividend period had ended, and capital for startups would grow scarcer. The next opportunity, he argued, lay in serving businesses rather than consumers. He drew the parallel to America, where companies that hit the ceiling of market growth earlier had turned to tools for efficiency and cost, and enterprise software had flourished as a result.
That same year he gave a keynote on what he called the "second half" of the internet economy. The phrase 下半场 spread through the industry and, as he later noted with some pride, through the state's own vocabulary. He proposed three directions. Up to the sky meant real high technology: AI, big data, cloud. Into the ground meant not merely staying grounded but going underground, because connecting consumers was no longer enough. WeChat had solved connection, the barriers to it were low, and a company had to dig into every link of an industry's supply chain. Globalization meant recognizing that countries are not boundaries; they are sources, currencies, and cultural habits. For Toutiao, he observed, the boundary was language.
In December 2018, three months after the IPO, he gave an internal speech of some ten thousand characters titled, roughly, Twenty Years of To C, Twenty Years of To B. It is the most revealing document in the record because of how far from food it travels. He began with the planet. "Although we are a company of only 35,000 people, with a market value of only about 100 billion yuan," he said, "we pay extreme attention to macro. Macro is extremely important." To show how important, he talked about temperature. Humans had lived on earth for a very long time, he said, and certain things in their environment were so fundamental that, because they rarely changed, nobody noticed them: air, water, temperature. He asked how sensitive human survival was to a few degrees of change. A small fluctuation in the macro environment, the speech warned, can be fatal to a company.
Read now, this is a man warning himself.
On March 4, 2020, Meituan's tenth birthday, the country was locked down. In an internal letter Wang thanked his colleagues for their work during the pandemic, "especially those on the front line in Wuhan: thank you for holding on." He placed the company's decade inside the macro frame he loved. In those ten years China's urbanization rate had passed one half, the service sector had passed half of GDP, and mobile internet users had passed half the population. "This is the backdrop to the booming of China's life-services industry," he wrote, "and it is what has allowed us to digitize the life-services industry in this way." In the same year Meituan launched Meituan Select, 美团优选, a community group-buying service. Neighborhood organizers aggregated orders for fresh produce and daily necessities, with next-day delivery to pickup points in lower-tier cities. Pinduoduo, Alibaba, and Didi all piled in. Subsidies ran into billions of yuan.
On September 30, 2020, the company dropped "Dianping" from its corporate name and became simply Meituan. In April 2021 it raised nearly $10 billion, about $7 billion in equity and $3 billion in convertible bonds, to fund groceries, drones, and autonomous delivery. It was a war chest at the top of a cycle. Later that same month the State Administration for Market Regulation opened its antitrust investigation. Ant Group's IPO had already been halted. The temperature in the room was changing, by a great deal more than a few degrees.
Then came the poem.
It would be wrong to say the poem caused the crackdown. The investigation predated it by about two weeks. Bloomberg reported that the subsequent scrutiny was intensified as political reaction to the post, and the Beijing Municipal Human Resources and Social Security Bureau also took an interest in the company. What the poem did was turn Wang personally into a variable. A man whose speeches had warned about the macro environment for years had, in a single post on his own private notebook, become part of the macro environment himself.
Meituan accepted a fine without appeal and said it would comply with regulatory requirements in the future. It was fined for predatory pricing and false advertising in community group buying. Wang went quiet. He kept, by every visible indication, the low profile officials had requested. In September 2025 he attended the commemoration of the 80th anniversary of the victory in the War of Resistance Against Japanese Aggression. He sits on the 13th Executive Committee of the All-China Federation of Industry and Commerce. The granddaughter generation of the families Mao destroyed is not his to speak for, and he does not try.
Meituan Select closed in 2025. At an annual general meeting on June 26, with management presenting a review of strategic mistakes across the previous five years, Wang named two. One was Youxuan, the community group-buying venture. The other was not going international fast enough after the IPO, missing a window to build early presence abroad. Public introspection of this kind is rare among Chinese platform CEOs. It was rare for Wang too. The monk who never said let's go home had finally admitted, out loud, that one of the ordeals had not been worth it.
Keeta, and the Owl
The second regret had already been addressed by the time he named it.
On February 2, 2024, Wang sent an internal email announcing the first large-scale reorganization since the "Retail + Technology" strategy of 2021. Core local businesses were consolidated. On-demand delivery, the in-store business, the Meituan platform, and basic R&D would all report to senior vice president Wang Puzhong. Dianping, SaaS, bike-sharing, and power-bank sharing went to senior vice president Zhang Chuan. Two units would report directly to Wang Xing: drones and overseas. Of all the businesses in a company of more than 100,000 people, the founder kept the two that were pure frontier, the one that went up into the sky and the one that went out into the world. "Currently, both internally and externally, we face various challenges," he wrote, "but at the same time, these are opportunities for the company to further grow."
The overseas business had a name by then. In May 2023 Meituan launched a food delivery app called Keeta in Hong Kong, starting in two densely packed Kowloon districts, Mong Kok and Tai Kok Tsui. It was the first time the company had offered its services outside mainland China. Analysts were skeptical. Keeta covered the whole territory by the end of October, ahead of schedule. By early 2024 it had 1.3 million registered users and roughly 37 percent of Hong Kong's market.
In September 2024 Keeta launched in Riyadh with free delivery and discounts, backed by a stated investment of SR1 billion, about $266.6 million. It aligned its pitch with Saudi Arabia's Vision 2030, and the company spoke openly about capturing a very large share of the market. In August 2025 Keeta entered Qatar. On September 27, 2025, it launched in Dubai; a Chinese commercial office in the UAE marked the occasion with a headline about Meituan's internationalization moving "at full speed." In October 2025 Keeta entered Brazil.
In the 2017 speech, globalization had been the third of three directions. Wang's line then was that countries are not boundaries but sources, currencies, and cultural habits. Keeta is that line made literal: the same logistics brain, the same subsidy playbook, applied to Kowloon tenements and Gulf suburbs and São Paulo traffic. The playbook is recognizably the one from 2011, minus the one-deal-a-day restraint: free delivery, steep discounts, density first.
The sky came next. In February 2023, Wang personally put money into a new AI startup called Light Year as an angel, and in March he joined its board. Light Year had been founded by Wang Huiwen, the classmate who had received the November 2003 email about nodes and edges and sat in the audience at Tsinghua in 2017 when Wang told the story. In June 2023 Meituan bought Light Year outright for $281 million. The man who had once sold his own company too cheaply for lack of a check was now writing the check for his oldest collaborator's.
In September 2025 Meituan released its first open-source large language model, LongCat-Flash-Chat. Variants followed in quick succession: a reasoning model, a multimodal model, a cheap mixture-of-experts model, a long-video generator, and a small image model that, by one account, beat larger rivals through cleaner data. In June 2026 the company released LongCat-2.0, a 1.6-trillion-parameter model trained on 50,000 Chinese-made chips and published under the MIT license. Meituan did not name the chips' manufacturer, but in a WeChat post it said it had used a Huawei software library for inter-chip communication. It also said it had needed to compensate for the domestic chips' lower memory capacity and less mature software ecosystem compared with Nvidia's H800s. Before the official launch, a pre-release version had been quietly circulating on the OpenRouter inference service under a code name, "Owl Alpha."
An owl is a fitting mascot for a man who reads at night. The practical application of all this intelligence is an agent called Xiaomei, which helps users order dinner and book hotels. Two decades of the most advanced information technology in China have converged, in the end, on the same question.
There is one more detail from the sky. Meituan began using drones for urban deliveries in 2023, with a maximum range of five kilometers and the ability to fly in bad weather. For its autonomous delivery vehicles it licensed the complete vehicle design, molds, and tooling for a model from Li Auto, the electric-car maker on whose board Wang sits as a non-executive director. Between September 12 and 15 of one year, he reduced his holding of Li Auto's Hong Kong shares by 1,947,900 shares, cashing out about HK$311 million. The founder who once borrowed from his parents to pay for servers now trims positions in other people's car companies over a long weekend.
The macro, meanwhile, has not stopped moving. In September 2026, regulators were again reported to be investigating technology companies, Meituan and Alibaba units among them, for suspected violations of competition law.
The Riders, and the Question at the Table
There is a part of the Meituan story that Wang's philosophy has never quite metabolized, and it wears a yellow jacket.
In 2021, by the company's own count, 5.27 million people earned income delivering for Meituan, more than a million of them on any given day, across 2,800 counties and cities. They are the edges of the network, in Wang's 2003 vocabulary: the things along which the information flows, except that in this case the information is soup. The sociologist Ya-Wen Lei, in a 2021 paper in the American Sociological Review titled "Delivering Solidarity," studied how the architecture of platforms like Meituan shapes the ways their workers come together to contest it. The riders have been the subject of Chinese documentaries, some pulled from circulation, about the pressure of algorithmic timing and the price of a three-yuan delivery. In 2025 Meituan piloted features allowing riders to rate and block customers. In November of that year it launched what it called a Rider Grievance Care Program, along with awards for riders' acts of bravery and good deeds.
None of this resolves neatly, and it would be dishonest to pretend otherwise. Wang's company has made Chinese urban life measurably easier for hundreds of millions of people. Chinese millennials joke that Meituan is more convenient than their mother's kitchen. It has done so partly by organizing an enormous number of other people's hours with extraordinary precision. The mission statement Wang wrote, 帮大家吃得更好，生活更好, "We help people eat better, live better," is sincere. Whether 大家, "everyone," includes the person carrying the bag is a question the company now answers with programs and awards, and the critics answer differently.
In his 2017 talk at Tsinghua, standing in the Mong Man Wai Concert Hall twenty years after he had arrived as a freshman, Wang gave the students three pieces of advice. Don't take yourself too lightly: Tsinghua would batter them for four years and convince them they were ordinary, but they were not. Don't take yourself too seriously: much of what they had learned would not apply, and a graduate who refused to start from the most basic work would miss many opportunities. And don't take other people too seriously, he said, with one exception: but treat the customer as the most important thing of all. A university vice president, Yang Bin, introduced him that day. Yang noted that Tsinghua had historically produced few alumni in the industries closest to people's daily lives, which he called the "soft color" of livelihoods, and hoped companies like Meituan would add it.
It is possible to read the whole career as a long answer to a question asked at a dinner table in 1997. The answer arrived in stages. First came the network as a theory of meaning, people as nodes and relationships as edges. Then came the network as a theory of speech, which the state ended in July 2009. Then came the network as a theory of survival, thirty kilometers a day through the War of a Thousand Groupons. Then the network as a theory of appetite, twice a day, in a yellow helmet. Then, in May 2021, it briefly became a theory of history again, on a site named after the oldest question at the Chinese table, and the market answered within hours.
Fanfou is still there. He still owns it.
What remains is the image before all of it: a long table in Beijing in the autumn of 1997, an eighteen-year-old from the Fujian hills who has just turned down the chance to ask about anyone's secret crush, the seniors looking at one another, and between them all, unattended, the hot pot still boiling.
Part IIThe Playbook
What follows is an attempt to extract operating principles from Wang Xing's career that are specific enough to be useful and uncomfortable enough to be true. Wang is often summarized as a "relentless" founder, which is accurate and useless. The more instructive record lies in the particular decisions behind the relentlessness: what he refused to spend money on, what he copied faithfully and what he copied wrong, when he merged instead of fought, and the two strategic errors he eventually named in public. Each principle below is grounded in the narrative above and ends with a single tactic.
From Zero to Meituan: How Wang Xing Built China's Leading Tech Empire offers a longer treatment for readers who want the company history in more detail.
The principles
- 1.Copy the form; study the differences.
- 2.Raise for the company you'll become, not the one you are.
- 3.Notice when your product changes category.
- 4.Let competitors pay to educate the market.
- 5.Spend trust before you spend reach.
- 6.March thirty kilometers a day.
- 7.Win the highest-frequency transaction.
- 8.Merge with your mirror when the war stops paying.
- 9.Treat the macro as temperature.
- 10.Keep the founding table intact for decades.
- 11.Put the frontier on the founder's desk.
- 12.At scale, the founder's voice is a market event.
九
Nine Defeats, One Victory
The failures that taught the lessons Meituan applied.
2003
Late November: Wang emails Tsinghua classmates proposing a social network built on the logic of computer networks. Returns to China at Christmas.
2004
Three partners pool 300,000 yuan and rent an apartment in Haifeng Garden. Launch Duoduoyou (a Friendster clone) in August; it stalls.
2005
Youzitu and clones of Google Maps and Evite fail. Xiaonei launches December 8, acquiring ~5,000 users for 3,000 yuan.
2006
Out of cash after an investor signs but never funds, Wang sells Xiaonei to Oak Pacific Interactive for ~$2M in October.
2007
Launches Fanfou (a Twitter clone) and Hainei (social network). Hainei fails to catch Renren and Kaixin.
2009
Fanfou is taken offline after the Urumqi riots in July; it stays dark for more than a year.
2010
With Fanfou's reopening hopeless, Wang launches Meituan on March 4.
2011
Renren lists on the NYSE on May 4. Meituan announces unconditional refunds and does 1.45B yuan in first-year transactions.
Principle 1
Copy the form; study the differences
The "copycat" label obscures what Wang actually learned from his clones. The lesson was not that copying works; three thousand Groupon clones proved it mostly doesn't. The lesson was that copying is a test of understanding. His 2012 post-mortem on Duoduoyou is a manifesto for this: founders who copy are tempted to add small differentiations, but "when your understanding of the product is not deep enough, you don't know why it does what it does, nor do you know which differences are critical."
Meituan's success came from copying the non-obvious parts of Groupon faithfully, above all the one-deal-per-day constraint, while competitors copied the surface and improvised everywhere else. Xiaonei succeeded partly because it was, in Wang's word, simple. Both times the faithful copy beat the clever one, because the original's constraints existed for reasons the copiers hadn't bothered to understand.
The deeper version of this principle concerns which differences matter when you transplant a model into a new environment. Groupon in America took 30 percent; in China the war drove take rates to single digits. Twitter in America was a media company that could be a media company; in China that turned out to be the fatal difference. Wang's skill, eventually, was to tell apart the differences you can ignore and the ones that kill you.
Tactic: Before modifying any model you're borrowing, write down the reason behind each of its constraints, and change only the ones you can explain.
Principle 2
Raise for the company you'll become, not the one you are
Xiaonei did not die of bad product. It died of underestimation. Wang only wanted to raise $1 million, "but that was far from enough." He hadn't met enough investors. He accepted a term sheet whose lock-up clause bound him to a partner who never funded. He lost his business plan on the way to Sequoia. The man who bought Xiaonei for about $2 million raised $430 million from SoftBank on the same asset and took it public at a valuation Forbes put above $4 billion.
The lesson stuck. Accounts of Meituan's founding emphasize that Wang, drawing on his earlier ventures, began to prioritize cash flow and team management and to "actively integrate into the operational rules of the business world." Meituan reached monthly breakeven in 2012 and annual profit in 2013, in the middle of a subsidy war, while competitors flamed out. When he did raise, he raised very large sums: $3.3 billion in 2016, $4 billion in 2017, $4.2 billion at the IPO, nearly $10 billion in April 2021.
The principle is not "raise a lot." Wang's 2021 raise preceded a regulatory winter and the Youxuan losses. The principle is to size capital to the company's true potential scale and to protect your negotiating position from the kind of single-counterparty dependency that sank Xiaonei.
Tactic: Never sign exclusivity or lock-up terms with an investor until the money has actually arrived in your account.
Principle 3
Notice when your product changes category
He Jiayan's diagnosis of Fanfou is the most important strategic insight in Wang's history: "at a certain size, it takes on the attributes of media." Fanfou had not changed its features. It had changed its nature simply by growing, and the regulatory, political, and social rules that applied to it changed accordingly.
This happens in every market, not just in authoritarian ones. A marketplace becomes a labor employer when enough people depend on it for income; Meituan's 5.27 million riders made it a de facto labor institution with all the scrutiny that implies. A delivery app becomes a utility when enough cities depend on it, which partly explains why antitrust regulators arrived in 2021. A founder's private microblog becomes a market-moving publication when the founder controls a company worth hundreds of billions.
Wang's response to Fanfou was to leave the information business for commerce, a category the Chinese state was far more comfortable letting private companies run. Zhang Yiming, who left Fanfou during the shutdown, drew a different conclusion and built ByteDance, a media company engineered from the start to handle what Fanfou could not.
Tactic: At every order-of-magnitude increase in users, ask in writing what kind of institution your product now resembles, and who regulates institutions like that.
Principle 4
Let competitors pay to educate the market
In the War of a Thousand Groupons, Lashou spent $10 million on marketing in a single month. Meituan deliberately underspent on advertising. Wang reasoned that rivals' campaigns would teach Chinese consumers what group buying was without creating loyalty to any particular site. Category education is a public good. Whoever pays for it subsidizes everyone else.
This is a counterintuitive position for a founder in a land grab, where the instinct is to shout loudest. But in a market with near-identical products and low switching costs, awareness is the cheapest thing to borrow and retention is the most expensive thing to build. Meituan put its money into merchant density, refund guarantees, and city-by-city sales coverage, the things that brought a user back the second time, and let others pay for the first.
団
The War of a Thousand Groupons
How Meituan's choices diverged from the field in 2010–2012.
| Field consensus | Meituan's choice | Type |
|---|---|---|
| Pack the site with many deals daily | One deal per day, so merchants feel the volume | Focus |
| Spend heavily on brand advertising | Underspend; let rivals educate the category | Capital |
| Keep the money from expired coupons | Automatic, unconditional refunds (March 2011) | Trust |
| Sprint in good months, retrench in bad | "Thirty kilometers a day," every day | Discipline |
| Chase the take rate and the margin | Accept a single-digit take; reach monthly breakeven by 2012 | Survival |
Tactic: When your category is still being explained to customers, cut your share of awareness spending and redirect it to whatever drives the second purchase.
Principle 5
Spend trust before you spend reach
The unconditional refund policy announced on March 4, 2011, cost Meituan real money: the revenue from coupons customers bought and never used, which in the Groupon model is a quiet profit center. Wang gave it up publicly, onstage, as the centerpiece of his first-anniversary event. In 2015 he credited it with building the brand and winning the war.
The deeper logic is that in a market of near-identical clones, the only durable differentiator is a promise that costs you something. Anyone can match a discount. A competitor matching a refund guarantee has to give up the same breakage revenue, and most won't, because the war was already eating their margins. The refund became a moat precisely because it hurt.
The 500-yuan credit codes on the anniversary invitations are the same idea in miniature. Wang didn't ask the journalists to write about Meituan; he gave them a reason to use it and told them, in effect, to judge for themselves.
Tactic: Find the revenue line customers resent most, give it up publicly, and make the guarantee the centerpiece of your positioning.
Principle 6
March thirty kilometers a day
At the 2012 annual meeting in Peking University's Centennial Hall, Wang told the story of the race to the South Pole and set a cadence: thirty kilometers a day. Meituan then "frantically opened stations, frantically listed deals, frantically made visits," and by year's end its lead in group buying was established.
This principle is often misread as "work very hard." The South Pole parable is about consistency under variance: you go the same distance in a blizzard as on a clear day, which means you never exhaust yourself on good days and never stall on bad ones. In a subsidy war where rivals were raising and burning nine-figure sums in bursts, Meituan's steady, compounding expansion city by city outlasted competitors who sprinted and then ran out of money.
The same temperament shows in Wang's line about patience: the more confident you are about the future, the more patient you can be with the present. The thirty-kilometer rule is how that patience becomes an operating cadence.
Tactic: Set one fixed daily or weekly output target your team hits regardless of market conditions, and don't raise it in good weeks.
Principle 7
Win the highest-frequency transaction
Meituan Waimai launched in late 2013, and food delivery became the company's engine, "a key driver of user engagement" that pushed traffic to hotels, travel, and everything else. The insight is about frequency. A hotel booking happens a few times a year; a meal happens twice a day. Whoever owns the daily habit gets to cross-sell the occasional purchase almost for free.
Food delivery is operationally brutal and, for years, barely profitable. At the 2018 IPO Wang could only say it was "very close to break even." But it was the transaction that made Meituan the app people opened most, and that made the Dianping merger, the travel business, and later instant retail possible. Wang's 2017 call to go "into the ground," deep into the industry supply chain rather than skimming connection, was the strategic rationale. Delivery was the way down.
Tactic: Map every transaction in your category by how often a customer makes it, and build your acquisition strategy around winning the most frequent one, even at low margin.
Principle 8
Merge with your mirror when the war stops paying
In March 2015 Wang declared it "the year of the decisive battle for O2O." In October 2015 he merged with Dianping, the very company he would have been fighting. This is James Carse's Finite and Infinite Games applied to strategy. A finite player fights to win; an infinite player asks whether winning this particular war keeps the larger game going. Dianping held what Meituan lacked, a decade of reviews and photographs, the deepest store of opinion about where to eat in China. Fighting for it would have burned capital on both sides. Combining with it created the merged company worth $18 billion within months.
The merger also realigned Meituan's patronage, with Tencent becoming the main backer while Alibaba withdrew and funded Ele.me. Wang accepted a permanent rival in exchange for ending a ruinous symmetrical one. He told the press that neither side should expect to dominate soon. That was the infinite player's posture, and also, as the 65 percent share would later show, a bit of strategic modesty.
Tactic: When you and your closest rival hold complementary assets and are burning the same capital, run the merger math before you run the war math.
Principle 9
Treat the macro as temperature
In December 2018 Wang told his managers that a company of 35,000 people pays "extreme attention to macro," and illustrated why with temperature: a variable so constant that nobody notices it until a few degrees of change become lethal. He had already called the end of the mobile internet's free growth in April 2017, citing a 30 percent rise in promotion costs and flat smartphone shipments. He framed Meituan's tenth anniversary entirely in terms of urbanization, the service economy, and mobile penetration crossing fifty percent.
The irony is that the macro shift that most hurt Meituan, the 2021 regulatory turn, was exactly what his own framework should have flagged. He raised nearly $10 billion that April and committed to the community group-buying subsidy war just as the temperature dropped. Youxuan, by his own later admission, was one of his two great strategic errors. Even the founder who preaches macro awareness can misjudge it when he is in the middle of a land grab.
℃
Reading the Weather
Wang's macro calls, and the one that went against him.
| Year | Call | Outcome |
|---|---|---|
| 2011 | Build mobile apps early; follow users to phones | Right 60% mobile by 2013; 90%+ by 2014 |
| 2017 | Mobile dividend over; costs up 30%; next wave is to-B | Right Shift to supply-chain depth |
| 2017 | "Up to the sky, into the ground, globalization" | Delayed Global push slow until Keeta (2023) |
| 2020–21 | Community group buying; ~$10B raise | Wrong Regulatory winter; Youxuan closed 2025 |
Tactic: Before any major capital commitment, list the three slowest-moving assumptions it depends on, whether policy, demographics, or cost of capital, and stress-test each against a reversal.
Principle 10
Keep the founding table intact for decades
At Tsinghua in 2017, Wang paused his own origin story to say that classmate relationships are "very important" and that students should value them. The record bears him out. The November 2003 email went to Wang Huiwen, who became one of his most important lieutenants; two decades later Wang backed Wang Huiwen's AI startup as an angel, joined its board, and then had Meituan buy it for $281 million. Mu Rongjun, another classmate, remains among Meituan's key people. The Fanfou team largely waited through more than a year of shutdown rather than leave, and many came along to Meituan.
Chinese commentators note that Meituan's senior team has been unusually stable for a company of its scale and volatility. For a founder described as introverted and not naturally charismatic, this stability is less a matter of inspiration than of loyalty compounded over twenty years. It goes both ways: the Light Year acquisition was, among other things, an act of reciprocity toward the first person who said yes to the 2003 email.
Tactic: Identify the three to five people who were with you before anything worked, and build explicit structures, whether equity, board seats, or capital for their own ventures, that keep them in your orbit for decades.
Principle 11
Put the frontier on the founder's desk
The February 2, 2024 reorganization handed Meituan's enormous core businesses to two senior vice presidents and kept exactly two units reporting directly to Wang: drones and overseas. In a company of more than 100,000 people, the founder assigned himself the parts with the least revenue and the most uncertainty.
The logic connects to Wang's second admitted regret, not going international fast enough after the IPO. Exploratory businesses inside large companies die when they report through executives whose bonuses depend on the core. By taking them himself, Wang gave Keeta the founder's patience and appetite for risk. The results came quickly: Hong Kong coverage completed early, then Riyadh, Qatar, Dubai, and Brazil within roughly two years. On the technology side, a 1.6-trillion-parameter model was trained on domestic chips.
Tactic: Once the core business can run without you, delegate it and personally own the one or two bets most likely to define the company in ten years.
Principle 12
At scale, the founder's voice is a market event
Wang kept Fanfou as a private notebook: Murakami, classical verse, fragments. In May 2021 a single Tang dynasty poem posted there coincided with a 9.8 percent single-day share drop, roughly $16 billion in value lost in a session, and a summons from officials asking him to lie low. The investigation was already open. The poem made the founder personally a source of risk.
This is Principle 3 applied to a person rather than a product. At a certain size, a founder takes on the attributes of an institution, and every public utterance gets read as policy, signal, or provocation. Wang's literary habit was once part of his charm, the "poet entrepreneur," and it may also reflect something real in a family whose grandfather died in the Cultural Revolution. But the market and the state do not read for nuance. After 2021 Wang largely went silent, and Meituan's subsequent public posture has been all operations: refunds, rider programs, models, launches.
Tactic: Once your company becomes systemically important, assume every personal post will be read as a corporate statement, and route anything ambiguous through the question of whether you would say it on an earnings call.
Free playbook
Get The Business Model Playbook
58 business models, one visual page each: how the money flows, the metrics that matter, and who runs it. Free when you join the Faster Than Normal email.
Free. No spam. Unsubscribe anytime.
Part IIIQuotes and Maxims
In their words
— Wang Xing, on the lessons of cloning, in a 2012 account of his entrepreneurial historyThe difficulty of cloning lies in the feeling that you can differentiate yourself slightly; when your understanding of the product is not deep enough, you don't know why it does what it does, nor do you know which differences are critical. Subtle differences can have significant implications.
— Wang Xing, China Business Leaders Annual Conference, Beijing, December 8, 2013Mobile internet is only in the pocket-watch era. It hasn't reached the wristwatch era yet.
— Wang Xing, China Business Leaders Annual Conference, December 2014Entrepreneurship is a pursuit — the pursuit of opportunity. And the most important characteristic of pursuing opportunity is that, for the moment, you don't consider the resources you currently control. That is the best definition of an entrepreneur I've seen.
— Wang Xing, address to Tsinghua students, September 27, 2017Don't take yourself too lightly. Don't take yourself too seriously. And don't take other people too seriously — but treat the customer as the most important thing of all.
— Wang Xing, Bloomberg TV, on the day of Meituan Dianping's IPO, September 20, 2018We have our largest core business, food delivery, that's growing super fast and it's also very close to break even. We are going to become profitable in our more mature business and at the same time invest in our new initiatives.
Maxims
- Copy to understand, not to imitate. A clone that doesn't know why the original makes its choices will change the wrong things.
- Raise for the giant. Undercapitalizing a winning product hands it to whoever writes the bigger check, as Xiaonei's $2 million sale and Renren's $740 million IPO demonstrated.
- Size changes species. At a certain scale a website becomes a newspaper, a marketplace becomes an employer, and a founder becomes an institution.
- Let rivals pay for the lesson. In a new category, awareness is a public good; spend on the second purchase instead.
- The promise that costs you is the only one that differentiates. Unconditional refunds beat three thousand clones because they hurt.
- Thirty kilometers, every day. Steady cadence outlasts sprints in any war of attrition.
- Own the meal, and the rest follows. The highest-frequency transaction is the front door to every other one.
- Play the infinite game. End the war that only burns capital, even if it means merging with the enemy you just declared a decisive battle against.
- Read the temperature, especially when you're winning. The macro shift that hurts most is the one you underweight in the middle of a land grab.
- Name your mistakes aloud. Public accounting of strategic errors, like Youxuan and the delayed global push, is how an infinite player updates the rules.
