Part IThe Story
One unusually warm evening in April 2025, a fifty-two-year-old man in a delivery jacket revved a scooter through Beijing's traffic-snarled streets, picked up boxed lunches alongside the other riders, and handed them to customers who did not, at first, recognize him. Later that night, over spicy hotpot and ice-cold beer, he welcomed two riders from rival delivery companies to his own. The cameras were rolling. The videos went viral. By the time anyone at Meituan or Alibaba's Ele.me understood what was happening, JD.com was handling 25 million takeout orders a day across 350 cities, more than half the volume of the market's runner-up. According to Bloomberg's account, neither company saw Richard Liu coming.
It was a publicity stunt, and an effective one. It was also something stranger: a return. The man on the scooter had spent most of the previous seven years out of public view. He had been arrested in Minneapolis in 2018 and was never charged. In 2022 he gave up the chief executive's title, during the same season Beijing's regulatory campaign pushed a whole generation of tech founders into quieter roles. He had watched from behind a curtain of governance documents while the company he founded in a twelve-square-meter stall lost its growth story, its share price, and much of its sense of itself. Two months after the scooter ride, at a rare news conference at JD's Beijing headquarters on June 17, 2025, he said what the market had already concluded:
— Richard Liu, JD.com news conference, Beijing, June 17, 2025For JD, it's a lost five years, to put it bluntly. No innovation, no growth, no progress. It should be considered the most unremarkable and least valuable five years in my entrepreneurial history.
The scooter was not new. For years Liu had spent one day annually as a JD courier, an act Fortune found charming enough in 2015 to caption: Yeah, that's him — on the tricycle scooter. The gesture is easy to read as theater, and some of it plainly is. Still, the delivery vehicle runs through Liu's life the way a river runs through a valley. It is his origin, his moat, his method, his alibi, and his argument. He has said, in a line quoted often enough that it has become a kind of epitaph written in advance, that his "whole life has revolved around the supply chain." Taken literally, the sentence is close to true. The things Liu has moved — a village's 76 eggs, a fat cut of pork, a stack of honest invoices, a tablet of magneto-optical disks, a refrigerator to a farmhouse in Sichuan, a 1.68-yuan bubble tea to a Beijing apartment — make up a map of what he believes, and of what he owes.
By the Numbers
From a Counter in Zhongguancun
¥12,000Capital Liu used to open his first Zhongguancun stall in 1998
¥8M+Losses suffered during the 2003 SARS outbreak, which pushed Jingdong online
$1.8BRaised in JD.com's May 22, 2014 Nasdaq IPO
¥1.309TJD.com revenue in 2025, past the ¥1 trillion goal Liu set for employees a decade earlier
776,682Employees at year-end 2025, up from seven in 1999
15.8%Liu's equity stake, which has carried roughly 77–80% of the votes
76Eggs his home village contributed, with 500 yuan, to send him to university in 1992
Seventy-Six Eggs and the Fattest Piece of Pork
Even the birthday is disputed. The Chinese identification number on a JD.com memorandum filed with the SEC points to March 10, 1973. Baidu's encyclopedia, and a good deal of Chinese-language reporting, gives February 14, 1974. English-language Wikipedia has gone back and forth. For a man whose company built its reputation on the premise that every receipt should match every transaction, it is a fitting irony that the ledger of his own life opens with a discrepancy.
The place is not in dispute. Guangming Village, Lailong Town, Suqian, in the north of Jiangsu province, a flat agricultural district far from the coastal boom that would later make Jiangsu rich. Accounts of the family's fortunes don't line up neatly. Some biographers say his grandparents were prosperous traders in the early twentieth century, and that by the time Liu was born his parents were rice farmers. Others describe parents who shipped coal from the north of China to the south on river barges. Ancestral records trace the line to Xiangtan in Hunan. In January 2018, at his father's request, one of the richest men in China posted a public appeal on social media asking other Lius from Xiangtan to help him reconstruct his family tree. Commenters predicted, with some justification, that a great many long-lost relatives were about to materialize.
All of the accounts agree that the household was poor. Liu has said his maternal grandmother shaped his character more than anyone, and that his father shaped his entrepreneurship. The grandmother kept him while his parents worked. Their daily meals were sweet potato and corn. On special occasions she took him to the nearest farmers' cooperative to buy the fattest piece of pork she could find. Lean meat would be eaten once. Fat could be rendered, and the rendered lard would flavor the family's cooking for the rest of the year. There is no better short course in supply-chain economics: buy the input with the most downstream utility, and make one transaction stretch across twelve months.
The boy decided he wanted to be a village head, so that every family in Guangming could afford to eat pork. He kept that ambition for a long time. Decades later, in a memoir dictated to the internet commentator Fang Xingdong, Liu titled one chapter 当官彻底没戏了: roughly, "Becoming an official was completely hopeless." He went to Beijing anyway, admitted in 1992 to Renmin University, the People's University, an institution known for its connection to China's political elite. Baidu records that he was the top scorer in Suqian that year. He enrolled in sociology.
The villagers paid his way. According to accounts Liu has repeated, and that his hometown confirms, the people of Guangming collected 500 yuan and 76 eggs to send him off. The eggs presumably did not survive the trip. The obligation did.
Two Yuan for Bean Sprouts
Sociology, Liu soon concluded, would not get him a job. He taught himself to program, which was a rare skill on a Beijing campus in the early 1990s, and took freelance coding work. He saved enough to buy his own computer and mobile phone, both expensive status objects at the time. Then, in 1994, he bought a restaurant near the university gate.
He knew he lacked experience, so he did what seemed modest and sensible: he stayed out of management and let the staff run the place. Within months they had emptied it. Years later he described the method to an Associated Press reporter with a precision that suggests he had gone over the arithmetic many times since.
— Richard Liu, interview with the Associated Press, 2015An employee would pay 2 yuan for bean sprouts but tell me he paid 4 yuan. The cashier and other employees would sneak money into their own pockets.
The loss exceeded 200,000 yuan. Liu borrowed from his father to cover it, roughly $31,000 at the time, a crushing sum for a family from Lailong Town. A Chinese-language profile captured his shame: he had wanted the business to make his father see him differently. "I was extremely disappointed, and I started to think about it from the perspective of human nature," he recalled. "Are people good or evil? I treated my employees so well. Why did they do this to me?"
The story gets told as a parable of betrayal, but the conclusion Liu eventually reached was more interesting than the one he started with. After graduating in 1996, he took a job at Japan Life, a Japanese health-products company, and spent two years there paying down the debt. He rose quickly: director for computers, then director for business, then logistics supervisor. That last title would matter more than anyone could have guessed. At Japan Life he also revised his verdict on the restaurant. "Over those two years, I gradually understood that the restaurant's failure was my own fault," he said, "because I had not established a management structure, had not supervised, and had not built financial systems and procedures."
The distance between those two statements, why did they do this to me and it was my own fault, is where a temperament formed. Liu came out of the bean-sprout affair with two convictions held at once, and they have pulled against each other ever since. The first is that people will cheat if the system lets them. The second is that the system is the founder's responsibility. Everything JD would later become, from the in-house logistics to the anti-corruption campaigns that former colleagues describe as bordering on paranoid, and to the governance clause that let the board meet only in his presence, can be read as a long reply to a cashier in 1994.
In an essay for Women's Wear Daily two decades later, Liu reduced the restaurant to two lessons: "First, make smart choices about whom you entrust to run your business. Second — and more importantly — take pride in what you create and respect the work other people put into creating something of quality and value." A Chinese headline on an earlier profile put it more bluntly. It said he had learned to 将偏执控制在健康的程度之内, to keep paranoia within a healthy range. The headline does not say how healthy.
A Price Tag in Zhongguancun
On June 18, 1998, Liu rented a twelve-square-meter counter in Zhongguancun, Beijing's electronics district. He had 12,000 yuan, which was everything he had, and he sold magneto-optical drives and disks. He named the business Jingdong, 京东, by combining a character from the name of his university girlfriend, Gong Xiaojing, with the last character of his own. According to accounts of the relationship, they separated because they wanted different things from the entrepreneurial life, and she went on to become a government employee. The name stayed. Twenty-seven years later it is on 776,682 payroll records, on red delivery vans across China, and on a 90,000-square-meter fulfillment center in Oberhausen, Germany. Few brands carry the faint ghost of a breakup this far.
Zhongguancun in 1998 was a bazaar. Counterfeit goods were everywhere. Prices were negotiated, then renegotiated. Buyers routinely had electronics unboxed and plugged in at the counter, because a sealed box guaranteed nothing. Merchants padded bills, evaded taxes, and paid kickbacks to institutional buyers. Liu had none of the relationships that made the bazaar work, and he had just learned at great expense what happens when no one keeps an honest ledger.
So he made honesty his business model. He sold only authorized goods. He later claimed his stall was the first in Zhongguancun to put fixed price tags on merchandise, which ended the haggling. Employees remembered him issuing invoices to customers who had not asked for them. In a market where an invoice was a tax liability, this was close to eccentric. It also meant JD paid its taxes, which a later chronicler noted was "quite unusual in China at the time." Competitors, as Liu tells it, thought he was mad: How can you do business in Zhongguancun without relationships or kickbacks? He was doing it, in part, because he had no choice. "I felt this was an opportunity," he later said, "to establish a new kind of business."
It worked. Customers who did not want to be cheated came back and told their friends. By 2003 Jingdong had twelve stores. A video of the company's 1999 annual meeting, attended by seven employees including Liu, was still playable sixteen years later. Liu screened it for a ballroom of employees in January 2016 and treated its survival as evidence: "Every time we sell a CD, record a video, we have preserved it for sixteen years without problem — this type of product quality is extremely hard to achieve." Even in his own nostalgia, the archival footage counted as proof of authenticity.
The comparison he would use for the rest of his career was already there in 1998. His rival was the counterfeiter, not some other honest retailer. In the second of his WWD essays he went after the platforms that let fakes circulate, without naming them, though everyone knew which one he meant: "These practices are bad for brands. They are bad for consumers. But I want to be absolutely clear: These are not Chinese values." It was a patriotic argument for honesty, made by a man who had become rich by being the only honest stall in a dishonest market.
The Year the Streets Emptied
In the spring of 2003, SARS emptied Beijing. Zhongguancun went silent. Customers stayed home, staff stayed home, and twelve stores that had been earning money began to bleed. Before the virus passed, Jingdong lost more than eight million yuan. Each day of rent on a store with no customers was a measurable step toward bankruptcy.
The improvisation that followed has hardened into legend. Employees stuck at home started posting products on online bulletin boards and forums. In January 2004 Liu launched the company's first retail website, at a domain that dates the moment exactly: jdlaser.com. That much was triage, and Jeffrey Towson, a longtime JD analyst, makes the fair point that going online during a pandemic "wasn't a particularly bold move." Any shopkeeper with a modem would have tried it.
The decision that followed did take nerve. Once the virus had passed and customers could walk through the doors again, Liu closed all twelve stores in 2005. The physical business had been healthy. It had survived the worst year in its history. He shut it down because the online business was better, and he did not want to divide his attention or his capital between the two. Staff threatened to quit; one chapter of his memoir is titled 关店上线，员工要离职, "Closing the stores and going online, employees want to leave." He closed them anyway.
That same year someone offered to buy JD for 18 million yuan. For a man who had been in debt seven years earlier and had just lost eight million yuan to a virus, the offer was large, and it would have cleared every obligation he had. He turned it down.
China's internet in 2004 was still mostly urban. In 2003, fewer than 0.2 percent of rural Chinese had ever been online. The customers who found jdlaser.com were young, male, educated, and based in big cities, exactly the people Liu later described as "young people familiar with the Internet and open to online commerce." The company changed its name and domain to 360buy.com in June 2007, became Jingdong Mall, and in 2013 settled on JD.com. Under every name, the business sold the same thing it had sold from the twelve-square-meter counter: a guarantee that the thing in the box was the thing on the label.
Please Shop Elsewhere
In 2008, JD posted a notice on its website asking customers not to order. It went further and suggested they buy from someone else.
Retail rarely produces a gesture like that. The warehouse had passed capacity. Goods could neither come in nor go out. Liu reasoned that a customer who waited two weeks for a laptop would be lost for good, and that a customer told honestly to go elsewhere might come back. The company blog later framed this as a matter of principle. It was also another receipt, and the restaurant owner who had learned to distrust padded figures was now issuing the receipt against himself.
The warehouse crisis came out of the most consequential bet Liu ever made. In 2007, against the objections of investors and many of his own employees, he began building JD's own logistics network from scratch. At that point China had no national logistics system worth the name. There were thousands of small regional carriers, careless handling, lost parcels, and broken goods. JD's repair rate for hard disks was very high, and the company traced the damage to third-party couriers loading and unloading roughly. Most early customer complaints concerned delivery rather than products. Liu concluded that he could not guarantee the authenticity of a product he could not physically protect between his warehouse and the customer.
The bet was large enough to sink the company if it failed. One summary of the period puts the commitment at a billion dollars over the years that followed. Liu, who had been a logistics supervisor at Japan Life, appears to have seen in the fragmentation of Chinese shipping the same disorder he had seen in Zhongguancun pricing: a broken system that customers had stopped expecting anything from, and therefore an opportunity. He later said it outright, in a speech at Oxford's Saïd Business School: "the more problems an industry has, the more opportunities there are for success, and the more one should courageously invest in it."
Out of the warehouses came a promise with a numerical name. JD called it 211. An order placed before 11 a.m. arrived the same day. An order placed before 11 p.m. arrived by 3 p.m. the next day. Where JD's own network reached, it guaranteed 100 percent next-day delivery and same-day delivery on 40 to 50 percent of orders. No other Chinese retailer of comparable size promised anything close. When Liu sat down at Fortune's Brainstorm Tech conference in Aspen in July 2018, he described free delivery in under 24 hours on almost any item over roughly $10, and said that competing in Chinese retail without it would be "a disaster."
The year 2007 brought one more strategic decision. Liu abandoned the company's identity as an electronics retailer and adopted what he called a full-category strategy: books, apparel, groceries, furniture, everything. Each new category demanded more warehouse space, and more warehouse space demanded more categories to fill it. The logistics network and the catalog grew together, each one justifying the other.
The book war showed his willingness to bleed. On December 10, 2010, Liu announced on Weibo that every book on JD would be priced 20 percent below competitors. Dangdang, then China's leading online bookseller, answered four days later with a 40-million-yuan discount fund. JD cut again. Orders surged so violently in November 2011 that the network jammed and deliveries slipped. Publishers complained. JD began issuing coupons instead of cutting list prices, to protect them. Chinese headlines said both companies were "going mad" and competing to be the "prodigal son." Fortune's 40 Under 40 entry in 2012 called it an "all-out price war" and recorded the trade-off without sentiment: 200 percent annual revenue growth, $3.3 billion in 2011 sales, $1.3 billion in cash, 10,000 new hires that year and 10,000 more the next, with profits and an American IPO pushed further away. In 2023, Dangdang opened a flagship store on JD.com.
Through all of this Liu kept doing his annual delivery shift, and he kept paying his couriers more than the market rate. Delivery work in China was among the most precarious and badly treated jobs in the economy. Liu understood, earlier than his rivals, that the person ringing the doorbell was the only JD employee most customers would ever see. A photograph by Sim Chi Yin for The New York Times shows him bending to kiss the hand of one of his delivery men. By 2015 the company had more than 20,000 couriers guaranteeing same-day service in 130 districts and counties. By June 2017 Liu told CNBC the number was over 70,000.
He made that last remark during a conversation about getting rid of them.
A Quorum of One
JD.com listed on Nasdaq on May 22, 2014, raising $1.8 billion. The stock rose about 15 percent on its first day and closed with a market value Baidu put at $29.7 billion. It was the largest initial public offering of a Chinese internet company in New York to that date, and the largest Chinese company listed only in the United States. Alibaba's record-setting offering would follow in September, but JD was first.
Two months earlier, Tencent had taken a 15 percent stake in JD. It paid partly in cash and partly by handing over its own struggling e-commerce operations, Paipai and QQ Wanggou, along with a stake in Yixun. The logic was openly geopolitical in the commercial sense: Tencent owned China's social graph, and Alibaba owned its commerce. Combining Tencent's traffic with JD's warehouses would build the strongest challenger available. In 2015 the two companies launched the "Jingteng Plan," a portmanteau of their names that linked JD's purchase data with Tencent's social data. The Chinese press began calling the rivalry the "cat and dog war," after Tmall's cat and JD's grinning mascot dog, Joy. The cat held 59 percent of the business-to-consumer market. The dog held 23.
In 2016 Walmart sold JD its Chinese e-commerce business, Yihaodian, in exchange for a 5.9 percent stake worth $1.5 billion. It later nearly doubled its holding and by February 2017 owned 12.1 percent. The world's largest physical retailer had looked at China and concluded that its best option was to become a minority shareholder in a former Zhongguancun electronics stall. Walmart sold its stake for about $3.7 billion in August 2024.
The investors got a company in which they would never have control. The prospectus disclosed weighted voting rights that gave Liu, with roughly 16 percent of the equity, nearly 80 percent of the votes. The articles of association went further. As long as Liu remained a director, the board could not make binding decisions unless he was present, in person or by teleconference. If he was absent, the board could act only with his permission or if he was ill. The clause explicitly excluded "any confinement against his will" from the exceptions, which suggested he would keep control even from a jail cell. Variety, profiling him as a media executive, summarized it in a parenthesis: "JD.com's board cannot form a quorum without his presence."
The clause would draw attention later. At the time it looked like the logical end point of the bean-sprout lesson. A founder who had learned that unsupervised systems get looted had written a corporate constitution in which nothing could happen without him in the room.
The corporate blog records a related refusal. Sometime in 2008 or 2009, when JD was private and short of money, someone advised Liu to push the company's valuation down in a new round. That would have handed him more shares and squeezed out early backers. He declined, on the grounds that those investors had helped the company when it needed help. After the IPO, the blog adds with some satisfaction, they made more money. Liu took control through voting structures, openly disclosed, and not by quietly diluting the people who had trusted him. The distinction mattered to him. Whether it mattered to minority shareholders who could never outvote him is a separate question.
The personal story turned public too. Early in 2014, Liu took a course at Columbia University. Zhang Zetian, an exchange student at Barnard, had become briefly famous in China as "Sister Milk Tea" after a photograph of her holding a cup of milk tea went viral. She was nineteen years younger than he was. The relationship became tabloid material. Chinese gossip sites also reported, falsely, that Liu was secretly in New York for cancer treatment, a rumor he later gave its own chapter in his memoir. They married in Beijing in August 2015 and held a grand ceremony at the Great Barrier Reef on October 1. That Spring Festival they went back to Guangming together and handed red envelopes to more than 650 elderly villagers. The founder who had once named his company after one woman would later appear on rich lists jointly with another. Hurun's 2026 Global Rich List ranked "Richard Liu and his wife Zhang Zetian" 508th, at 55.5 billion yuan.
The honors came as expected. In January 2016 he told employees that JD now had nearly 110,000 full-time staff, planned to add 40,000 more that year, had over 500,000 part-time workers in crowdsourced logistics, and had village representatives in more than 170,000 villages. It served more than 100 million households, he noted, "equivalent to the total number of families in several European countries together." He reminded them that a year earlier he had set a target of one trillion yuan in net revenue. He compared the climb to the Himalayas and asked how far up they were. In 2015, Fortune named him one of the World's 50 Greatest Leaders and praised his "mighty small ego." In 2016 JD became the first Chinese internet company on the Fortune 500. In January 2018 Liu was appointed to the National Committee of the Chinese People's Political Consultative Conference. Late in 2017 he had become honorary village head of Pingshitou, a hamlet of 655 people in Hebei's mountains who grew apples, mushrooms, and walnuts, and he promised to raise their incomes tenfold in five years. The boy who had once written off officialdom as hopeless had become, in a sense, both a village head and a member of the national political advisory body.
Minneapolis
On August 31, 2018, Richard Liu was arrested in Minneapolis on suspicion of rape. He was held a little over sixteen hours and released without bail.
He was in the city for a residency program run by the University of Minnesota's Carlson School of Management for Chinese executives. The night before, there had been a dinner, which The Wall Street Journal later described as a night of wine and Japanese food. A Chinese student at the university, Liu Jingyao, who is unrelated to him, had attended as a volunteer. She said he followed her back to her apartment afterward and raped her. He denied it, through his lawyers, and maintained the sex had been consensual. JD's communications staff told the press that police had "quickly determined" there was "no substance" to the claim, even though the investigation was still underway. It took the company nearly two days to acknowledge he had spent a night in custody.
In the first two days of trading after the arrest, JD lost $7.2 billion, 16 percent of its market value. The governance documents that had looked like a founder's prerogative now looked like a hostage situation. Reuters reported the confinement clause. Jamie Allen, secretary general of the Asian Corporate Governance Association, a man who has spent a career reading the fine print of Asian corporate charters, said what investors were thinking: "We can't think of any other company that has such articles. I find it baffling." James Robinson of APCO Worldwide described the problem in plain operational terms: "If your top person is in a jail in Minnesota, then it could be a question of a lack of decision-making authority." Three law firms filed class actions against JD, alleging it had failed to disclose pertinent information.
On December 21, 2018, the Hennepin County Attorney, Mike Freeman, declined to bring criminal charges. He called it a "complicated situation" and said his office had found evidentiary problems that made it highly unlikely a charge could be proven beyond a reasonable doubt. JD said Liu had been falsely accused.
Liu Jingyao did not drop the case. On April 16, 2019, she filed a civil suit in Minnesota, where the burden of proof is lower, seeking at least $50,000 in damages. In China, she became the target of sustained online attack. The New York Times reported in December 2019 that the Chinese internet had turned against her. A judge rejected a motion to remove JD's corporate entity as a defendant. Jury selection began in Minneapolis on Thursday, September 29, 2022, in what The Times said would be one of the first times a Chinese figure this prominent had faced a jury on such an accusation. Xiaowen Liang, a Chinese feminist activist and lawyer in New York, told the paper that the story felt familiar thousands of miles from Minneapolis: "Many women in China have had similar experiences with this workplace culture of drinking, dinner and sexual harassment."
The case was settled within days, before a verdict. Both sides issued a joint statement. The terms were not disclosed in the sources reviewed here. Liu was never charged with a crime and was never found liable in a civil court. Liu Jingyao never got the jury she had spent four years trying to reach.
That is the whole of what the record establishes, and a profile cannot add to it. What the episode exposed was structural and is not in dispute. The company had been built so that nothing happened without one man in the room, and the room was suddenly in Hennepin County. Months later Liu stepped back from day-to-day management and said he would concentrate on new businesses. He largely left public view. The quorum of one now had an empty chair.
The Founder Who Could Not Stay Away
The formal handover took four years, and the regulatory climate set the timing. In late 2020 and 2021, Beijing pursued the country's largest internet companies on antitrust, data protection, and consumer protection. Alibaba was fined $2.8 billion. Founders began to step aside one after another. Colin Huang, the former Google engineer who built Pinduoduo out of group buying and farmers' produce, resigned as chairman. Zhang Yiming of ByteDance stepped down. Su Hua of Kuaishou resigned. In February 2022, Liu pledged JD shares then worth $2.3 billion to charity, at the height of what CNN called a "billionaire charity rush" in response to Xi Jinping's "common prosperity" campaign.
On April 7, 2022, JD announced that Xu Lei would replace Liu as chief executive, effective immediately. Xu had spent more than a decade at JD, most recently running its retail business, and had been made president the previous September. Citi analysts credited him with improving user acquisition and retention. Liu stayed on as chairman with about 77 percent of the votes. The company said his focus would be "long-term strategies, mentoring younger management, and contributing to the revitalization of rural areas." The last phrase repeated almost word for word a stated priority of the national leadership. JD shares fell 3 percent in Hong Kong that day.
The retirement held for about seven months.
In early October 2022, the Minneapolis case was settled. In November and December, Liu, speaking by video conference from Hong Kong, ran two meetings with JD's senior executives that participants described to the South China Morning Post in terms that could be called a scolding. He called some executives "liars," without naming them. One attendee recalled him saying that "only one and a half vice-presidents in the retail business unit" told him the truth. The unit had about forty vice presidents. He announced that he would fire 10 percent of vice presidents and 10 percent of business directors by year-end. About 2,000 senior managers had their salaries cut by up to 20 percent. The money was redirected toward housing and education benefits for frontline workers. Liu personally gave 100 million yuan, matched by the company, to a fund supporting the families of employees who died or were injured. The company's housing guarantee fund was raised to 10 billion yuan, and 77 percent of applicants since 2023 have been frontline workers. JD Digits came in for criticism too. It was run by Li Yayun, JD's former chief compliance officer, who had taken over the fintech unit in 2021 after its IPO was cancelled under tighter rules on online lending. In a leaked speech he insisted the company return to basics: low prices and good service.
The structure that had looked like vulnerability in Minneapolis now looked like an operating principle. Analysts told the Post that the episode showed Liu remained firmly in control of the company despite leaving the frontline roles. In May 2023, after barely a year as CEO, Xu Lei stepped down. He was replaced by Sandy Xu Ran, then chief financial officer, the second CEO in thirteen months.
Pinduoduo, meanwhile, now operating as PDD Holdings and the owner of Temu, was growing faster than either JD or Alibaba by selling cheap goods to price-conscious consumers in a weakening economy. Colin Huang, who had stepped down under the same regulatory weather as Liu, had become the third-richest person in China. In 2023, JD's shares fell more than 55 percent. PDD's rose more than 60 percent. In December 2023, an employee posted a list of JD's failings on the company's internal forum: inadequate support for third-party sellers, too few low-priced products. Liu answered in the thread, in public, in front of the whole company.
— Richard Liu, JD.com internal discussion board, December 2023So many issues have emerged, certainly because I mismanaged.
"I feel our brother has said it all too well," he wrote. "We need to change, or there's no way out for our company." He told employees not to tang ping, or "lie flat," the slang for doing the minimum. A few weeks earlier, Jack Ma had posted a similar call for reform on Alibaba's forum. The two founders who had defined Chinese e-commerce were, in the same month, responding to anonymous employee posts like anxious alumni writing in to a class newsletter.
The self-criticism was striking. It was also familiar. It was my own fault, because I had not established a management structure. The sentence from Japan Life in 1997 and the sentence from the internal forum in 2023 are the same sentence. Twenty-six years apart, Liu had found that someone was again telling him two yuan was four, and he again concluded that the failure was in the system and that the system was his. In April 2024, JD put an AI-generated version of him into its livestream shopping channels, nicknamed "Procurement and Sales Manager Brother Dong." It drew more than 20 million views in its first hour and sold 50 million yuan of goods. The founder had made himself available in more places at once than any human could manage.
Robots, Riders, and the Ledger
The automation question has followed Liu for a decade, and his answers have shifted in ways worth recording precisely.
In June 2017, he told CNBC that drones would cut rural delivery costs by "at least 70 percent." He described the system with a farmer's concreteness: "In each village, we have a delivery man; he lives there and works there, and he's got a backyard. He's like a small distributor of us in that village. So we can use our drone to fly to the delivery man's backyard directly." He also said: "Today we have over 70,000 delivery men working on the street. It's high cost, you know. If you can use robotics to deliver a parcel, the cost will be very low." JD was flying drones in Beijing, Sichuan, Shaanxi, and Jiangsu, testing an aircraft that could carry a ton, and planning drone airports across the southwest. By 2018 it ran a daily fleet of about 500 drones.
Speaking in English at the World Retail Congress in Madrid in 2018, he went further. "I hope my company would be 100% automation someday," he told Forbes, "no human beings anymore, 100% operated by AI and robots." He described people-free warehouses and thirty-minute drone deliveries. Around the same period, Reuters reported him criticizing "slackers" as the company made cuts, and KrASIA reported him citing "huge losses" when he scrapped couriers' base salaries.
Then the courier came back to the center of the story. JD's chief economist told a Fortune conference in September 2024 that the company had grown from 180,000 to 620,000 employees in six years while raising wages. In February 2025, JD became the first platform in China to pay full social insurance for its food-delivery riders, and Meituan followed within hours. In an internal speech that circulated in April 2025, which was widely reported but whose authenticity was never formally confirmed, Liu attacked the average 25 percent commissions charged by China's food-delivery platforms, blamed them for restaurants cutting corners on hygiene, and promised that JD's delivery business would never make more than 5 percent net profit. Then he told the room something they already knew: "I've also run a restaurant, and to be honest, it doesn't make much money." The bean-sprout restaurant, three decades later, had become a credential.
In the same speech, he reportedly said: "The technologies and knowledge we've acquired should never be used to exploit restaurant owners or delivery workers, nor to deceive or endanger consumers with unsafe food." In June 2026 he said he hoped to train employees to service and maintain robots, so that they would not lose their jobs to artificial intelligence. WWD summed up the position in a headline: robots would eventually replace all 700,000 couriers, and JD would retrain them. Taken together, Liu's statements on automation form a ledger. Read in order, they show a man trying to reconcile two debts that cannot both be paid in full: to the customer, who wants the cheapest and fastest delivery, and to the courier, who is the customer's only human contact with JD and also its largest cost.
The scooter ride of April 2025 belongs in that ledger. So does its aftermath. JD offered restaurants zero commission, hiring bonuses for 100,000 new full-time riders, and a 10-billion-yuan discount campaign, and during its June festival it sold coffee and bubble tea for as little as 1.68 yuan. The festival is called 618, and it falls on the anniversary of the day Liu opened the Zhongguancun counter. JD's Hong Kong secondary listing in 2020 began trading on that date too. For the full year 2025 the company reported revenue of 1.309 trillion yuan. It had passed the trillion-yuan summit Liu had shown his employees a decade earlier, but operating income was only 2.8 billion yuan, a margin of about two-tenths of one percent. In the fourth quarter, with the food-delivery war absorbing cash, JD posted its first loss in nearly four years. The price war had come back exactly as it had in 2010, run by the same man and by the same logic.
The company also went abroad, under his chairmanship and with the same appetite. In July 2025 JD agreed to buy Ceconomy, the German parent of MediaMarkt and Saturn, in a deal valuing it at $2.5 billion, and held more than 70 percent by November. Talks to buy Argos from Sainsbury's collapsed within a day in September 2025. In January 2026 France blocked a JD bid for Fnac Darty, explicitly to keep it from becoming a direct sales channel for Chinese goods. Austria resisted the Ceconomy deal, and JD resubmitted with new commitments on data protection and local jobs. In March 2026 JD launched Joybuy in six European countries. It began as a first-party retailer that bought and held its own inventory, the 1998 model carried to Milton Keynes, Luton, and Oberhausen and backed by more than sixty warehouses and depots. It offered same-day delivery on orders placed before 11 a.m. and next-day delivery on orders placed before 11 p.m., which is 211 under a European name. In May 2026, reports said JD was preparing a £2 billion bid for The Very Group. The honest stall in the dishonest bazaar was now asking European regulators to accept a familiar argument: look at our records, not our passports.
In January 2025, Guangming Village prepared for the Spring Festival. Officials asked every household to submit its registration book and identity papers, and the village counted 1,400 households. Liu arrived on January 8. Each villager aged sixty or older received 10,000 yuan in cash. Every household received food, clothing, and appliances. The year before, it had been down jackets, more than a thousand of them, sent to more than 1,300 households. A farmer named Xu told a reporter that his elderly parents were eligible. A seventy-one-year-old villager said that even if the gifts stopped, the village would remember. Some skeptics online asked whether it was publicity. It probably was in part, the same way the scooter ride was.
Liu said what he says every year: "I am eternally grateful to the people of my village." The papers were checked, the names verified, the envelopes counted against the register, because a founder who once lost everything to a falsified receipt will not hand out money without one. Thirty-three years after the fact, the account was still open. Five hundred yuan has been repaid many thousands of times over, with interest no bank would set. The eggs have no price, so they cannot be repaid. In 1992 someone in Guangming, a household no reporter has named, counted them out by hand, wrapped them against the road to Beijing, and gave them to a boy who said he wanted every family in the village to be able to afford pork. There were seventy-six.
Part IIThe Playbook
The principles below come from twenty-seven years of decisions, some brilliant, some costly, and a few that should serve as warnings. Liu's career is a record of a founder turning a personal humiliation into an operating system. That system made JD China's largest retailer by revenue, and it also concentrated so much authority in one man that his absence nearly paralyzed the company. Operators should take the method and leave the hazards.
The principles
- 1.Make honesty the product, not the policy.
- 2.Own the last mile, because the last mile owns the customer.
- 3.Turn away demand you cannot serve.
- 4.Close the profitable door behind you.
- 5.Treat the courier as the brand.
- 6.Convert betrayal into controls, not cynicism.
- 7.Hunt in the most broken industries.
- 8.Use price wars as acquisition, and cap the upside on purpose.
- 9.Take control in the open; never dilute the people who trusted you.
- 10.Design for the founder's absence before it happens.
- 11.Don't hand over the title unless you hand over the truth.
- 12.Keep your origin as a reference customer.
Principle 1
Make honesty the product, not the policy
In 1998 Zhongguancun, honesty was not an ethical extra on top of a business. It was the business. Price tags, authorized goods only, and invoices issued without being asked all solved a problem every customer had: they did not trust the seller. Liu had no relationships and no kickback budget. What he could offer was something nobody else in the bazaar was offering, which was a match between what was promised and what was delivered.
That matters because in low-trust markets, trust is scarce, and scarce things command a premium. Liu's later essays put this as a moral argument: "There should be no second chance for deceiving your customers." Underneath the morality is a pricing argument. Every counterfeit sold on a competitor's platform made JD's guarantee more valuable. Twenty years on, the "zero tolerance" policy on fakes was still JD's main differentiation from Alibaba's open marketplaces.
The lesson is not that founders should be nice. It is that in some markets the widely tolerated dishonesty is the opening. Find the place where customers have learned to expect to be cheated, and stop cheating them conspicuously.
Tactic: List the three things your customers have learned to expect from your industry that they hate, and make the absence of each one a written, visible, verifiable guarantee.
Principle 2
Own the last mile, because the last mile owns the customer
The 2007 decision to build logistics in-house started with a narrow diagnosis: hard drives were arriving broken because third-party couriers handled them carelessly. The data showed that customer complaints were mostly about delivery, not products. Liu understood that you cannot guarantee authenticity for a product you cannot physically protect. A promise that ends at your warehouse door does not hold.
The investment was large enough to bankrupt JD if it failed, and investors and employees both doubted it. It became the company's lasting moat: a nationwide network capable of 211 delivery, later spun out as JD Logistics in a HK$24.6 billion IPO in 2021, and eventually rebuilt in Europe for Joybuy. Asset-light competitors could match JD's prices, but they could not match its delivery promise without building what JD had built.
The general point is to look for where your customer's experience actually breaks. It is often in the step you outsourced because it seemed like a commodity. If that step is where the customer judges you, it is not a commodity.
Tactic: Map every complaint from the last quarter to the step in the value chain where it originated, and treat any step causing more than a third of them as a candidate for ownership.
Principle 3
Turn away demand you cannot serve
In 2008 JD's website asked customers not to order and suggested they shop elsewhere until the warehouses caught up. Most founders would do nearly anything to avoid that. Liu's reasoning was that a lost sale is a one-time cost, while a customer who waits two weeks and decides you are unreliable is a permanent loss.
The 2008 notice was a cash-flow sacrifice made to protect a reputation that took ten years to build. It fits the 2025 decision to cap food-delivery margins at 5 percent and the 2010 choice to defer profits during the book war. Each time, Liu treated short-term revenue as the variable he could give up and the customer relationship as the one he could not.
Capacity limits are honest information. Hiding them borrows against your future credibility.
Tactic: Define the service level below which you would rather refuse an order than accept it, and build the "we're full" message before you need it.
Principle 4
Close the profitable door behind you
Moving online during SARS was survival, and any shopkeeper would have done it. The decision with strategic weight came in 2005, when the virus was gone, the twelve stores could have reopened at full trade, and Liu shut them anyway. Employees threatened to quit. A buyer offered 18 million yuan. He refused both the buyer and the comfort of a hybrid business.
Jeffrey Towson called it burning the boats. The more exact description is that Liu recognized the old business would keep taking capital, attention, and identity from the new one, simply by being there and working. Keeping a profitable legacy line running is rarely neutral. It pulls the organization back toward what it already knows.
J
The Decision Timeline
Each pivotal choice traded a present certainty for a future position.
1994
Restaurant near Renmin University collapses after staff embezzle funds; loss exceeds ¥200,000.
1998
Opens a 12-square-meter counter in Zhongguancun with ¥12,000; fixed prices, authorized goods only.
2003
SARS costs Jingdong more than ¥8 million across twelve stores.
2004
Launches jdlaser.com, the first online store.
2005
Closes all physical stores; rejects an ¥18 million acquisition offer.
2007
Begins building in-house logistics; adopts full-category strategy.
2010
Launches book price war against Dangdang via Weibo.
2014
Tencent takes 15% stake; $1.8B Nasdaq IPO on May 22.
2018
Arrested in Minneapolis; never charged; steps back from daily management.
2022
Hands CEO title to Xu Lei; civil suit settled in October; re-intervenes by year-end.
2025
Personally relaunches the food-delivery war; acquires control of Ceconomy.
2026
Launches Joybuy across six European markets.
Tactic: When a new model is clearly better, set a public date for retiring the old one, even if the old one is profitable, and tell your team before you lose your nerve.
Principle 5
Treat the courier as the brand
Liu saw before most of his rivals that the delivery worker was JD's main human interface with its customers. Hundreds of millions of people would never visit a JD office, read its prospectus, or meet its executives. They would meet the person at the door. So he paid couriers above market, did a delivery shift himself every year, and, in 2025, became the first platform in China to pay full social insurance to food-delivery riders.
His record here is not clean. He also talked about a "100% automation" company with "no human beings anymore," scrapped some couriers' base salaries, and lectured "slackers." The tension is real and unresolved: the courier is both JD's most valuable brand asset and its largest controllable cost. Liu's most recent answer, retraining couriers to maintain the machines that replace them, is a sincere attempt to reconcile the two. It has not yet been tested.
Tactic: Identify the lowest-paid role that has the most customer contact in your business, and make its compensation and dignity a board-level metric, not an HR line item.
Principle 6
Convert betrayal into controls, not cynicism
After the restaurant failed, Liu first asked whether people are good or evil. Two years later, at Japan Life, he gave a better answer: the failure was his own, because he had built no management structure, no supervision, and no financial systems. The first answer leads to bitterness. The second leads to architecture.
Many of JD's defining features follow from the second answer: an obsession with integrity that former colleagues called close to paranoid, an anti-corruption coalition JD set up with other internet companies in 2017 to blacklist offenders across the sector, and tight control over every step from supplier to doorstep. The danger is that controls can grow into control, a point Principle 10 takes up. The core insight still holds. When someone cheats you, the useful question is what in your system let it happen.
Tactic: After any internal fraud or failure, write the postmortem with a single required section titled "What I failed to build," and fill it in yourself.
Principle 7
Hunt in the most broken industries
"The more problems an industry has, the more opportunities there are for success," Liu told Oxford students, "and the more one should courageously invest in it." His career follows that rule closely. He entered electronics retail because it was full of fakes, built logistics because logistics was fragmented and careless, and went into food delivery in 2025 because, in his telling, 25 to 30 percent commissions were pushing restaurants into unsafe kitchens.
Dysfunction is a moat that incumbents built for you. When an industry has trained customers to expect bad outcomes, even modest competence stands out, and real excellence becomes a category of its own. The difficulty is that broken industries usually stay broken for structural reasons, so the fix generally requires capital and patience on the scale of the 2007 logistics bet.
Tactic: Rank the markets adjacent to yours by how much their customers complain, then ask which of those complaints your existing infrastructure could solve at marginal cost.
Principle 8
Use price wars as acquisition, and cap the upside on purpose
Liu has fought three major price wars: books against Dangdang in 2010, a broad assault on Alibaba in 2012 that Fortune called "all-out," and food delivery in 2025. Each time he gave up profit for share, and each time he set the terms publicly. The clearest version is the 2025 pledge, reported from an internal speech, that JD's delivery business would never earn more than a 5 percent net margin.
Announcing a ceiling on your own profits is unusual. It does two things. It tells partners, restaurants in this case, that you will not extract from them the way incumbents do, which gives them a reason to switch. It also gives competitors notice that the war will be long. JD's 2025 numbers, ¥1.309 trillion in revenue on ¥2.8 billion in operating income, show what the strategy costs. Dangdang opening a store on JD in 2023 shows what it can win.
¥
Conventional Playbook vs. Liu's Playbook
Where JD diverged from the dominant model in Chinese e-commerce.
| Conventional approach | Liu's approach | Outcome |
|---|---|---|
| Asset-light marketplace; outsource delivery | First-party inventory; build own logistics | Moat |
| Tolerate fakes, monetize takedowns | Zero tolerance; no "three-strike" rule | Trust premium |
| Maximize commission take rate | Zero commission; cap margin at 5% | Share gain, margin cost |
| Keep profitable legacy channels | Close every physical store in 2005 | Focus |
| Board operates independently of founder | No binding decisions without founder present | Key-man risk |
| Founder exits cleanly after succession | Chairman re-intervenes within seven months | CEO churn |
Tactic: Before starting a price war, write down and share the maximum margin you will accept at victory, because it is the most credible signal you can send to partners and rivals.
Principle 9
Take control in the open; never dilute the people who trusted you
Around 2008 and 2009, someone advised Liu to depress JD's valuation in a financing round so he could increase his own stake at early investors' expense. He refused, because those investors had backed him when it counted. He secured control instead through weighted voting rights disclosed in the prospectus, which any investor could read before buying.
The distinction is between control taken by stealth and control taken openly. Investors who bought JD shares knew they were buying into a founder's company. That is a legitimate bargain, and many investors made a lot of money on it. Diluting early backers through a manipulated round would have been a betrayal of the kind Liu spent his career guarding against. The JD corporate blog's remark that those investors "earned more after the company was listed" was the payoff for that choice.
Tactic: If you want control, structure it explicitly in your governing documents rather than engineering it through valuation games, and let anyone who disagrees decline to invest.
Principle 10
Design for the founder's absence before it happens
This principle is a warning. The clause that stopped JD's board from making binding decisions without Liu present, explicitly including periods of "confinement against his will," made sense as a defense against what he feared most: other people running the company badly while he wasn't looking. In September 2018 it turned into the company's largest liability. JD lost $7.2 billion in market value in two days, partly because no one could say who was in charge.
Concentrated authority makes a company fast while the founder is present. It also means the company has one point of failure, and the cause of a founder's absence can be illness, scandal, regulation, or a sixteen-hour detention in Minneapolis. Governance experts, Jamie Allen among them, found the arrangement "baffling" because it ignored a basic fact: every founder eventually leaves the room.
Tactic: Run a tabletop exercise in which the founder is unreachable for thirty days, and rewrite any governance provision that leaves the company unable to act.
Principle 11
Don't hand over the title unless you hand over the truth
Liu gave up the CEO title in April 2022. By November he was berating executives over video, saying only "one and a half" of roughly forty retail vice presidents told him the truth. Xu Lei left within thirteen months of taking over. Liu then went to the internal forum to take the blame himself: "So many issues have emerged, certainly because I mismanaged."
The pattern points to a specific failure in succession. The title moved, but information did not reach the person who still held 77 percent of the votes. If the people reporting to a new CEO believe the real authority is elsewhere, they will manage upward to that authority and give the CEO a filtered version of events. Liu's diagnosis that people were lying to him was probably correct. The structure gave them every reason to.
Tactic: If you step back but keep control, put the incoming CEO's direct reports on a single shared reporting channel that you both read, so that no one can tell either of you a different story.
Principle 12
Keep your origin as a reference customer
Liu's rural childhood is not just a source of philanthropy, though the red envelopes, down jackets, and 10,000-yuan gifts are real. It works as a product requirement. The drone network was designed around a village courier's backyard. The Pingshitou project set out to raise a hamlet's income tenfold by branding its apples and walnuts. One of Liu's more lasting claims is that e-commerce let rural Chinese see prices they had been cheated on for generations.
A founder who keeps a vivid picture of a customer the market ignores has a standing advantage in finding the next market. The village showed Liu what logistics costs looked like in third- through sixth-tier cities long before analysts produced the slides. Guangming was his first market, even though for most of his career it was not one he sold to.
Tactic: Pick one specific customer from your own past, the person least served by your industry, and test every major product decision against whether it reaches them.
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
— Richard Liu, as quoted in Li Zhigang, The JD.com StoryIf the founder of a company is always right and never wrong, then the company is doomed. I am not God. It's impossible for me to be right all the time on everything. I have to subject myself to the collective intelligence.
— Richard Liu, Women's Wear DailyDoing the right thing is always simple, but never easy. Commitments to quality, to protecting consumers and to being an honest and trustworthy partner are long-term efforts. They are not practices of convenience.
— Richard Liu, Women's Wear DailyDo not listen to the promises anyone makes about the future — including me — but rather look at our records.
— Richard Liu, speech at Saïd Business School, University of OxfordOpportunities for success come from observing the industry; the more problems an industry has, the more opportunities there are for success, and the more one should courageously invest in it.
— Richard Liu, on gifts to Guangming Village, January 2025I am eternally grateful to the people of my village. Their support was the foundation that allowed me to pursue my education and ultimately build JD.com.
For the fullest account of the company's first two decades, written with extensive access to Liu and his lieutenants, see Li Zhigang's The JD.com Story.
Maxims
- Sell the receipt. In a market where everyone expects to be cheated, a verifiable promise is worth more than a discount.
- The failure is in the system, and the system is yours. When someone steals from you, the useful question is what you failed to build.
- Own the step where the customer judges you. If delivery is where trust breaks, delivery is not a commodity.
- Say "we're full" before you say "we're sorry." Turning away demand you cannot serve protects the customer you want to keep.
- Close the door that still works. A profitable legacy business can still drain focus from the better one.
- The person at the door is the brand. Pay and protect the employee your customer actually meets.
- Go where the industry is most broken. Incumbent dysfunction is a moat someone else dug for you.
- Announce your margin ceiling. Committing in public to how little you will take is the strongest message you can send partners and rivals.
- Every founder eventually leaves the room. Build a company that can still decide things when you cannot be reached.
- Repay what cannot be priced. Some debts, like 76 eggs, are never settled, and keeping them on the books is what keeps a founder honest.
